Can one really figure value?

[Capper Al]

The answer is probably not, at least not precisely. While reading Precision by CXWONG, I'm taking a trip back through stats class. In his book, Mr. Wong talks of Normal Distributions, Standard Deviations and Fair Odds. He uses an example between horses where it is given that horse 'A' is one length faster than horse 'B'. He compares this to having two dices. Let's say that die 'A' has numbers 2 through 7 and die 'B' has numbers 1 through 6. The mean or average roll for die 'A' should be 4.5 while die 'B' will be 3.5. Correlate our two mentioned horses to these two dice. The 4.5 of 'A' being one length greater than the 3.5 of 'B'. Now we know that most of the time we threw these dice 'A' will have the higher number over the long run, but not always. The chart below shows the possible out comes:

On average for 6 rolls of each number of Die 'A'

----Roll on Die 'A' ---- 7-6-5-4-3-2: 36 rolls. ( 6 for each number/side )

Die 'A' is higher ------ 6-5-4-3-2-1: 21 'A' wins
Ties --------------------0-1-1-1-1-1: 5 ties
Die 'A' is lower --------0-0-1-2-3-4: 10 'B' wins

So with our correlations here, our one length advantage translates to 2.1/1 advantage for horse/die 'A'. Are we there yet? Don't think so. Should we figure our advantage higher, lower or level it at 2.1/1?

[Capper Al]

I forgot to clarify the correlation between the dice and the horses. Using 1/5 second per length in the example in post #1, let's say horse A's 6 furlong time is 109 then horse B's would be 109.2. The idea is that horses don't run their average time each race. Their speeds vary from race to race. In post #1 we are assuming both horses vary from 1 to 6 lengths, only horse A's average is 1 length faster than horse B's. The numbers on the dice for the example are set up for the correlation. So, in other words, the slower horse can win sometimes.

[HUSKER55]

that is why there is confetti on the floor. slower horse won.:D

[Overlay]

I think the fact that the slower horse sometimes beats the faster horse can be taken into account, or compensated for, by the following three considerations:

1) using winning probabilities associated with relative speed rankings for the entire field, rather than looking at winning as an all-or-nothing proposition for one horse;

2) taking the effect of other relevant handicapping factors into consideration, rather than having speed (as important as it may be) be the sole criterion on which a horse's chance of winning is gauged; and

3) after doing 1) and 2) above, also factoring odds into the equation to see whether any given horse (even the fastest horse in the field) is an acceptable risk in light of its overall chance of winning.

[TrifectaMike]

that is why there is confetti on the floor. slower horse won.:D

There is another room, which has so much confetti that one would suffocate upon entry. Confetti deposited as a result of the fastest horse winning.

Mike (Dr Beav)

[TrifectaMike]

I forgot to clarify the correlation between the dice and the horses. Using 1/5 second per length in the example in post #1, let's say horse A's 6 furlong time is 109 then horse B's would be 109.2. The idea is that horses don't run their average time each race. Their speeds vary from race to race. In post #1 we are assuming both horses vary from 1 to 6 lengths, only horse A's average is 1 length faster than horse B's. The numbers on the dice for the example are set up for the correlation. So, in other words, the slower horse can win sometimes.

What exactly in your own words are you trying to say?

Put the dice aside, what are you looking for?

State the problem, and define the solution you are seeking.

Mike (Dr Beav)

[Spiderman]

Dice:horses 10 hardway:2-1 horse in 6,7 horse field

[TrifectaMike]

I think the fact that the slower horse sometimes beats the faster horse can be taken into account, or compensated for, by the following three considerations:

1) using winning probabilities associated with relative speed rankings for the entire field, rather than looking at winning as an all-or-nothing proposition for one horse; I agree generally, but disagree with the use of speed rankings.

2) taking the effect of other relevant handicapping factors into consideration, rather than having speed (as important as it may be) be the sole criterion on which a horse's chance of winning is gauged; I agree. Although you don't need too many factors to achieve this step... the less the better and I believe six is too many (that is if you have the proper combination). and

3) after doing 1) and 2) above, also factoring odds into the equation to see whether any given horse (even the fastest horse in the field) is an acceptable risk in light of its overall chance of winning. How do you determine acceptable risk? Individual overlays are nearly impossible to determine on a series of bets? There has to be a better way. There is a better way. A way that is less susceptible to noise.


Mike (Dr Beav)

[Capper Al]

What exactly in your own words are you trying to say?

Put the dice aside, what are you looking for?

State the problem, and define the solution you are seeking.

Mike (Dr Beav)

After post #1 I realized that the point of the discussion was standard deviation. With regards with speed, I was trying to make a point that if somehow we knew the horses' real abilities, because of standard deviation, the fastest horse still might not win. I was correlating lengths off as one length to one number on a die and randomly playing out the possibilities. Sorry about how it came out. Hope it's cleared up a bit now. Tough topic to discuss.

[fmolf]

After post #1 I realized that the point of the discussion was standard deviation. With regards with speed, I was trying to make a point that if somehow we knew the horses' real abilities, because of standard deviation, the fastest horse still might not win. I was correlating lengths off as one length to one number on a die and randomly playing out the possibilities. Sorry about how it came out. Hope it's cleared up a bit now. Tough topic to discuss.
i think that the fastest horse wins maybe 35% of the time if you use last race beyer as a criteria .....could use best speed at distance or average of last three races...or plain old raw time.This is a very difficult question to answer i am now starting to lean more towards running styles as a determining which horses offer the best value...I may think horse a offers value at 2/1 you may think he's underlaid...very subjective

[TrifectaMike]

After post #1 I realized that the point of the discussion was standard deviation. With regards with speed, I was trying to make a point that if somehow we knew the horses' real abilities, because of standard deviation, the fastest horse still might not win. I was correlating lengths off as one length to one number on a die and randomly playing out the possibilities. Sorry about how it came out. Hope it's cleared up a bit now. Tough topic to discuss.

What you are trying to determine is not too easy nor is it too difficult. It is an approach Benter described in his lecture (There's a link to it on PA, I don't recall the thread).

He defines an actual horse performance( running time) as, U = V + e and assumes e ( e represents unknown influences) is normally distributed, which results in a performance distribution that is normally distributed around a certain mean. So, each horse has a normal distribution. From the individual horse distributions, you determine a joint performance distribution for the race.

Let's put Benter aside (what I'll describe is not how he solves the problem), and for the purpose of understanding the process do the following:

Let,

X1....f1(X1),
X2....f2(X2),
X2....f3(X2),
..............,
Xn....fn(Xn)

X1,X2,....,Xn represents the performance of Horse X1,Horse X2,...Horse Xn
and
f1(X1),f2(x2),....,fn(xn) represents the probability density of Horse X1,Horse X2,...Horse Xn

Let us describe how we would determine the probability that Horse X1 wins the race. So we want

P(X1-X2 <0) (recall that we are using running time)
P(X1-X3 <0)
................
P(X1-Xn <0)

and then multiple the probabilities together to arrive at HorseX1 wins the race.

P(HorseX1 Wins Race) = P(X1-X2 <0)P(X1-X3 <0)...P(X1-Xn <0)

From your posts, this is what I believe you are seeking.

Mike (Dr Beav)

[Broad Brush]

I forgot to clarify the correlation between the dice and the horses. Using 1/5 second per length in the example in post #1, let's say horse A's 6 furlong time is 109 then horse B's would be 109.2. The idea is that horses don't run their average time each race. Their speeds vary from race to race. In post #1 we are assuming both horses vary from 1 to 6 lengths, only horse A's average is 1 length faster than horse B's. The numbers on the dice for the example are set up for the correlation. So, in other words, the slower horse can win sometimes.

I always respect those who try to think "outside the box" and look for ways
to improve their game. In this case though, I don't think it is "apples to apples".
The problem is that the correlation of the "roll of the dice" has only two variables: the numerical value on each. This will be the same no matter how you
throw the dice. It will not change if you throw the dice hard or soft. In a race, the correlation between two horses depends on many
factors: the start, the pace, the trip, the performance of others and so on.
These and other factors changes the difference between the two.

To me the only way to establish the relative chances of each horse is to
choose a standard that you can compare each horse to each other, such as,
Beyer fig, final time, or another speed figure. Then you say to yourself:
what is each horse capable of today. The ones who can't meet what you have set as the winning fig or time get a 0% and are eliminated. For the remaining ones
you use all handicapping tools to figure if the race was run ten times with
various pace setups how often would each horse be capable of a winning performance. It should add up to 100%. Sometimes--a lot of times, I find races
that add up to more than 100%. This is a race to pass because I don't have a good enough read on the race to set fair odds versus chances.

Good Luck with your efforts though.

[Dave Schwartz]

i think that the fastest horse wins maybe 35% of the time if you use last race beyer as a criteria

That horse wins about 25.5% of the time.

[Capper Al]

What you are trying to determine is not too easy nor is it too difficult. It is an approach Benter described in his lecture (There's a link to it on PA, I don't recall the thread).

He defines an actual horse performance( running time) as, U = V + e and assumes e ( e represents unknown influences) is normally distributed, which results in a performance distribution that is normally distributed around a certain mean. So, each horse has a normal distribution. From the individual horse distributions, you determine a joint performance distribution for the race.

Let's put Benter aside (what I'll describe is not how he solves the problem), and for the purpose of understanding the process do the following:

Let,

X1....f1(X1),
X2....f2(X2),
X2....f3(X2),
..............,
Xn....fn(Xn)

X1,X2,....,Xn represents the performance of Horse X1,Horse X2,...Horse Xn
and
f1(X1),f2(x2),....,fn(xn) represents the probability density of Horse X1,Horse X2,...Horse Xn

Let us describe how we would determine the probability that Horse X1 wins the race. So we want

P(X1-X2 <0) (recall that we are using running time)
P(X1-X3 <0)
................
P(X1-Xn <0)

and then multiple the probabilities together to arrive at HorseX1 wins the race.

P(HorseX1 Wins Race) = P(X1-X2 <0)P(X1-X3 <0)...P(X1-Xn <0)

From your posts, this is what I believe you are seeking.

Mike (Dr Beav)

Believe or not, I was trying to simplify this with a lot of givens and the dice example.

[Capper Al]

OK, last try.


We are dealing with an AVERAGE time that varies hence the word "average".
We determine on AVERAGE, horse A is one length faster than horse B. For example, horse A's average six furlong time is 109 then horse B's six furlong time is assumed to be 109.2.
We have two magical dice that represent the equally distributive standard deviations for each horse on their numbers. The first die A represents horse A with numbers 2 thru 7. Die A will have an average roll of 4.5. The second die B represents horse B with numbers 1 thru 6. Die B will have an average roll of 3.5.
Our expectations of one greater between the averages of both dice correlates to horse A being one length faster than horse B. We randomly throw the dice to represent what could be the possible outcome between these two horses in our artificially made up contest.
We get the chart in post #1 showing the die representing horse A beats the die representing horse B and visa versa.
On average for 6 rolls of each number of Die 'A'

----Roll on Die 'A' ---- 7-6-5-4-3-2: 36 rolls. ( 6 for each number/side )

Die 'A' is higher ------ 6-5-4-3-2-1: 21 'A' wins
Ties --------------------0-1-1-1-1-1: 5 ties
Die 'A' is lower --------0-0-1-2-3-4: 10 'B' wins

[Robert Goren]

I have looked into predicting a horse's SR for a race based on past SRs and then using those predictions to predict the winner a little. The thing will jump out very quickly is that the SD of different horses vary a fair amount. For instance a horse returning from a layoff of 6 mo has a large SD than a horse with a layoff of 3 weeks. I found it very tough to get the SD down to a number that is very useful. I am not saying it is impossible but I was not able to do it with the somewhat limited resources I had available at the time.

[Capper Al]

I have looked into predicting a horse's SR for a race based on past SRs and then using those predictions to predict the winner a little. The thing will jump out very quickly is that the SD of different horses vary a fair amount. For instance a horse returning from a layoff of 6 mo has a large SD than a horse with a layoff of 3 weeks. I found it very tough to get the SD down to a number that is very useful. I am not saying it is impossible but I was not able to do it with the somewhat limited resources I had available at the time.

The problem for the lone capper, not enough time or resources to figure this all out. I'm with you on this one.

[pondman]

In a race, the correlation between two horses depends on many
factors: the start, the pace, the trip, the performance of others and so on.
These and other factors changes the difference between the two.

Bingo...

There are to many unknowns-- including my favorite-- How much money will the trainer make?

Standard deviation? Correlation? These are term for describing social behavior; They're really not suitable for predicting the downfall of Roman Society.

[Nitro]

From time to time I see threads like this and this time for the fun of it I asked one of my good friends (who’s a professional Actuary) to check it out. He told me in a nutshell that if any insurance company relied on all of that subjective mumbo-jumbo they’d be “rolling the dice” and out of business in no time. He explained how any statistical analysis when operating in one dimension is useless especially when there are many accountable variables involved in causing a potential outcome.

He related a simple correlation between the odds of having a car accident when looking at various types of drivers. They are typically defined by their age group and sex, the type of cars they own and drive, where they live, how much they drive, their driving records, etc., etc. Would it be fair then to certain groups of drivers to determine the value of their insurance premiums based on the potential odds of having an accident derived from a single variable? Would it be a smart move for any insurance company to cover a driver with a policy using only single variable? Of course not!

Apparently all of these variables are first examined individually for validity and then weighed according to their potential to produce the odds of having an accident. Then they are comingled in various permutations to see if any combination raises or lowers these potential odds.

Of course there are unknown situations that are always present and unaccountable that can also cause an accident. When driving a car a sudden distraction or even a back-seat driver can cause an accident. Is this starting to sound familiar?

As our discussion progressed, we got into the intangibles involved in horseracing that seem to override all of the many performance variables involved. In many situations the outcome of a horserace is more often then not determined by the races where the horses are knowingly entered and the intentions of those who are constantly monitoring the current condition of their animal. On race day the final intentions for each entry are rationalized by examining the competition, the conditions of the race and of course knowing the current condition of the horse. This is all done in order to determine if it has the ability to run its race and potentially do well. And finally there are the instructions given to the jockey/driver that could very well impact the outcome.

The questions that the handicappers should be asking is: How do you quantify these ever present intangibles and how much does this calculation impact the real odds value for playability?

[Robert Goren]

The actuary science approach has been tried many times by people who are not actuaries. The result of this a lot of torn up tickets. A guy I went to college with is an actuary. I asked him about horse racing and the use of the actuary tool box. His first response was "Why?" We talk a little more and basically said the actuary end of the insurance business is measuring a non competing events and horse races are competing events between similar contestants. You have to measure trait compares to some other trait in a head to head competition to get a usable result. The actuary tool box does not do that although he did say that he thought there were statistical tools that might do it. He also said that he did not think most gamblers were going to put in the years of to sort through the numbers but a group of very bright people with the right training and somebody to pay the bills might get to workable method fairly quickly. He emphasized the words "very bright" and "right training".

[pondman]

"Why?" We talk a little more and basically said the actuary end of the insurance business is measuring a non competing events and horse races are competing events between similar contestants.

Statistics is suitable for recording human behavior.

It will never be able to quantify:

...a 6 month layoff
...being shipped 5oo miles
...Chantel Southerland
...Tyler Baze dropping his whip

All of these could be a + or -.

To be successfull you need to think as a 4th level quant trader waiting for an unknown disaster, and being willing to loose 4 bets in order to have that 20-1 shot win.

Linear style regression will combine a star jockey, a high percentage trainer, and a high speed figure. This will result in a best bet. And these usually turn into trainwrecks.

Do horses with 6 month layoffs, being shipped 500, ridden by girls ever win at 40-1? Yes they do. It would be best to wait for these unknowns to show up
.

[Dave Schwartz]

Statistics is suitable for recording human behavior.

It will never be able to quantify:

...a 6 month layoff
...being shipped 5oo miles
...Chantel Southerland
...Tyler Baze dropping his whip

Are you suggesting that the human mind can quantify this accurately?

[TrifectaMike]

Statistics is suitable for recording human behavior.

It will never be able to quantify:

...a 6 month layoff
...being shipped 5oo miles
...Chantel Southerland
...Tyler Baze dropping his whip
Any of the above can be quantified. Maybe you can't, but there are those who can.
All of these could be a + or -.

To be successfull you need to think as a 4th level quant trader waiting for an unknown disaster, and being willing to loose 4 bets in order to have that 20-1 shot win. Bullshit!

Linear style regression will combine a star jockey, a high percentage trainer, and a high speed figure. This will result in a best bet. And these usually turn into trainwrecks. More bullshit!

Do horses with 6 month layoffs, being shipped 500, ridden by girls ever win at 40-1? Yes they do. It would be best to wait for these unknowns to show up And more bullshit!!
.

I suggest you learn the difference between probability theory and statistical analysis.

Mike (Dr Beav)

[TrifectaMike]

From time to time I see threads like this and this time for the fun of it I asked one of my good friends (who’s a professional Actuary) to check it out. He told me in a nutshell that if any insurance company relied on all of that subjective mumbo-jumbo they’d be “rolling the dice” and out of business in no time. He explained how any statistical analysis when operating in one dimension is useless especially when there are many accountable variables involved in causing a potential outcome.

He related a simple correlation between the odds of having a car accident when looking at various types of drivers. They are typically defined by their age group and sex, the type of cars they own and drive, where they live, how much they drive, their driving records, etc., etc. Would it be fair then to certain groups of drivers to determine the value of their insurance premiums based on the potential odds of having an accident derived from a single variable? Would it be a smart move for any insurance company to cover a driver with a policy using only single variable? Of course not!

Apparently all of these variables are first examined individually for validity and then weighed according to their potential to produce the odds of having an accident. Then they are comingled in various permutations to see if any combination raises or lowers these potential odds.

Of course there are unknown situations that are always present and unaccountable that can also cause an accident. When driving a car a sudden distraction or even a back-seat driver can cause an accident. Is this starting to sound familiar?

As our discussion progressed, we got into the intangibles involved in horseracing that seem to override all of the many performance variables involved. In many situations the outcome of a horserace is more often then not determined by the races where the horses are knowingly entered and the intentions of those who are constantly monitoring the current condition of their animal. On race day the final intentions for each entry are rationalized by examining the competition, the conditions of the race and of course knowing the current condition of the horse. This is all done in order to determine if it has the ability to run its race and potentially do well. And finally there are the instructions given to the jockey/driver that could very well impact the outcome.

The questions that the handicappers should be asking is: How do you quantify these ever present intangibles and how much does this calculation impact the real odds value for playability?

Your professional actuary needs to rethink his position. He doesn't understand the problem nor the solution.

Group stats and frequentist concepts are worth nada in racing.

Mike (Dr Beav)

[thaskalos]

IMO...the word "BULLSHIT" should NEVER be used when responding to another player's opinion on how to play this game.

There is enough room in this game for more than one way to win...and I seriously doubt that ANY of us have the game all figured out...

[TrifectaMike]

IMO...the word "BULLSHIT" should NEVER be used when responding to another player's opinion on how to play this game.

There is enough room in this game for more than one way to win...and I seriously doubt that ANY of us have the game all figured out...

The statements are pure BULLSHIT! They are not opinions. They are factually incorrect and nonsensical or BULLSHIT!

Mike (Dr Beav)

[thaskalos]

Do horses with 6 month layoffs, being shipped 500, ridden by girls ever win at 40-1? Yes they do. It would be best to wait for these unknowns to show up
.
This looks like an opinion to me...

And I would have to know a little more about the bettor who made it, before I categorize it as..."BULLSHIT"!

Remember...advanced math degrees are not a prerequisite to being an expert horseplayer.

ALL winning players can claim this title...

And Pondman professes to be a winning player...

[GameTheory]

This looks like an opinion to me...

And I would have to know a little more about the bettor who made it, before I categorize it as..."BULLSHIT"!

Remember...advanced math degrees are not a prerequisite to being an expert horseplayer.

ALL winning players can claim this title...

And Pondman professes to be a winning player...
Pondman's approach sounds good to me -- I have been successful with similar strategies. The only thing "BS" about what he stated was to be successful you *need* to do this, and also pointing out how other approaches *must* fail. In other words, he may be right that it can work (and works for him), but he is wrong to think that others must also do it that way. And if Mike is suggesting that Pondman's approach cannot work, that is also BS.

So, I agree with Thaskalos...

[TrifectaMike]

This looks like an opinion to me...

And I would have to know a little more about the bettor who made it, before I categorize it as..."BULLSHIT"!

Remember...advanced math degrees are not a prerequisite to being an expert horseplayer. It sure helps!

ALL winning players can claim this title...

And Pondman professes to be a winning player...

Ok, Then you explain to me how an "unknown" shows itself when it is "unknown". (Whatever that means)

Winning horse player or not the statements are BULLSHIT!

Mike (Dr Beav)

[TrifectaMike]

Pondman's approach sounds good to me -- I have been successful with similar strategies. The only thing "BS" about what he stated was to be successful you *need* to do this, and also pointing out how other approaches *must* fail. In other words, he may be right that it can work (and works for him), but he is wrong to think that others must also do it that way. And if Mike is suggesting that Pondman's approach cannot work, (I didn't suggest anything more than his statements were Bullshit.) as that is also BS.

So, I agree with Thaskalos...

Mike (Dr Beav)

[GameTheory]

Remember...advanced math degrees are not a prerequisite to being an expert horseplayer. It sure helps!I think it is a hindrance. Most with advanced training in statistics and math that come to the game bring lots of assumptions with them about how the problem should be approached -- assumptions that sink most of them. The vast majority of "math guys" are good at academic problems and things with known solutions, and often clueless about what to do with messy real-world problems, of which horse-racing is one of the toughest.

If you took 100 math guys and a 100 random other people of average intelligence, none of whom knew a thing about horse racing, and told them all to go be successful at it, I'd bet on the average joe group to have most success stories a year or two down the line...

[thaskalos]

I think it is a hindrance. Most with advanced training in statistics and math that come to the game bring lots of assumptions with them about how the problem should be approached -- assumptions that sink most of them. The vast majority of "math guys" are good at academic problems and things with known solutions, and often clueless about what to do with messy real-world problems, of which horse-racing is one of the toughest.

If you took 100 math guys and a 100 random other people of average intelligence, none of whom knew a thing about horse racing, and told them all to go be successful at it, I'd bet on the average joe group to have most success stories a year or two down the line...
I couldn't have said it better myself...:ThmbUp:

[TrifectaMike]

I think it is a hindrance. Most with advanced training in statistics and math that come to the game bring lots of assumptions with them about how the problem should be approached -- assumptions that sink most of them. The vast majority of "math guys" are good at academic problems and things with known solutions, and often clueless about what to do with messy real-world problems, of which horse-racing is one of the toughest.

If you took 100 math guys and a 100 random other people of average intelligence, none of whom knew a thing about horse racing, and told them all to go be successful at it, I'd bet on the average joe group to have most success stories a year or two down the line...

A hindrance? I doubt it. I was handicapping at 16 years of age. Safe to say way before any degrees. And at the present age of 61 I'm still a handicapper first.

Although a theoretical mathematician, I've worked on many real-world problems. Yeah, the average joe does real well in this game. You must be kidding.

Mike (Dr Beav)

[Jeff P]

The title of this thread is "Can one really figure value?"

From reading the posts in this (and previous) threads there appears to be support for the idea that quantifying value is a futile pursuit and/or something that is simply not doable. I think nothing could be further from the truth.

Not only do I think that value can be quantified.... I think quantifying value is one of the most important skills a player can develop. I also think the ability to quantify value is far more important to the player's bottom line than his or her ability to make selections.


-jp

.

[GameTheory]

A hindrance? I doubt it. I was handicapping at 16 years of age. Safe to say way before any degrees. And at the present age of 61 I'm still a handicapper first.And you probably would have done ok without your degrees too. You were a handicapper first, and still are -- you're making my point.

Remember, I didn't say "all", I said "most". In general, its a hindrance because of the unlearning that needs to be done. Most people aren't going to deal well with the rude awakening that they can't just throw some historical data through their favorite statistical procedure and be able to make money, which is what most of them seem to assume when they have their freshly minted advanced degree and think that they can now quantify the world. (Most quit after realizing they really don't know nuthin' bout nuthin' when it comes to horse racing. I've seen it many times.) And on top of that, most people learn those idiotic ad-hoc frequentist methods in school -- that stuff is still gospel most places. It infects the brain with ideas that everything must be shown to be "statistically significant" in a frequentist sense, but most of them never figure out that disproving the null hypothesis does not answer the question they really want to know. (And is in fact rather useless.) Look at how badly so-called experts in statistics misapply their methods to social problems and other real-world phenomena. Math guys do real well in this game? Yeah you must be kidding.

Now, you may have been able to avoid that if you knew how to handicap before you were an advanced mathematician. You may not have had many or any false starts in figuring out how to use the tools in your toolbox appropriately (and you have better tools than most being a Bayesian). BUT THAT'S A RARE SKILL.

In any case, you can be successful at this game without using any math or figures of any kind. (Beyond the very basics, that is -- obviously you'll need to know that 3/1 pays $8, etc.) Would you deny that?

[thaskalos]

This game is not a math problem, waiting to be solved...

[Cratos]

I think it is a hindrance. Most with advanced training in statistics and math that come to the game bring lots of assumptions with them about how the problem should be approached -- assumptions that sink most of them. The vast majority of "math guys" are good at academic problems and things with known solutions, and often clueless about what to do with messy real-world problems, of which horse-racing is one of the toughest.

If you took 100 math guys and a 100 random other people of average intelligence, none of whom knew a thing about horse racing, and told them all to go be successful at it, I'd bet on the average joe group to have most success stories a year or two down the line...

I disagree because knowledge is never a hindrance; it is the lack of knowledge that is the hindrance and if you are inferring that one with good math and statistics skill, but limited horseracing knowledge lacks the ability to be a good horse player/gambler; I agree.

However the opposite is not true, a person with good math and statistics skills plus a good knowledge of horseracing handicapping is far above the pack and the late Phil Bull a mathematician would be proof of that assertion.

[GameTheory]

I disagree because knowledge is never a hindrance; it is the lack of knowledge that is the hindrance and if you are inferring that one with good math and statistics skill, but limited horseracing knowledge lacks the ability to be a good horse player/gambler; I agree.I think that knowledge is very often a hindrance -- very often. Because knowledge is not understanding, but those with knowledge usually think it is. That is a hindrance they have to get over -- to realize that they don't "have it" immediately without bothering to learn it through hard experience.

However the opposite is not true, a person with good math and statistics skills plus a good knowledge of horseracing handicapping is far above the pack and the late Phil Bull a mathematician would be proof of that assertion.Most people with math backgrounds -- when they come to horse-racing -- assume that their math tools are going to show them the way and they don't really try to learn horse racing properly because they've got it into their heads they don't need to ("just show me the data"). That's what I mean by a hindrance -- their math baggage is hindering them from learning the game. Obviously if they have skills and experience in both areas that will be helpful, but that can't happen until they gain that experience in both areas (duh), so that's not really an argument. They were either not hindered or got over it (a hindrance is just a speed bump, after all, not a brick wall). But often the one skill is PREVENTING the learning of the other (or slowing it down) because of the arrogance that usually comes with it. Like I said, seen it many times -- most just quit after they build their first system (no doubt horribly overfit) and don't understand why they make profits on paper and not at the track.

[thaskalos]

However the opposite is not true, a person with good math and statistics skills plus a good knowledge of horseracing handicapping is far above the pack and the late Phil Bull a mathematician would be proof of that assertion.
IMO...for every ONE mathematician like Phil Bull, there are ONE THOUSAND mathematicians like WILLIAM QUIRIN...who, in spite of his math skills and handicapping knowledge, proved that - although he could TALK a good game - he ultimately lacked the flexible thinking needed to beat this game.

[Capper Al]

Hello Everybody,

I was trying to point out author CXWONG's simple explanation of the effects of variance in predicting the outcome. All the unknowns were coupled within the given and a correlation to a number on a die was made. His suggestion is a simply, straight forward, and countable visual example of what happens because of deviation. In our example, with its givens, we could actually count 21 wins for the fastest horse and 10 wins for the slower horse. Why or how these variances occur was beyond the scope of the example. I just thought it was neat (like cool man.)

[TrifectaMike]

IMO...for every ONE mathematician like Phil Bull, there are ONE THOUSAND mathematicians like WILLIAM QUIRIN...who, in spite of his math skills and handicapping knowledge, proved that - although he could TALK a good game - he ultimately lacked the flexible thinking needed to beat this game.

Not doubting Quirin's math skills, but there is little math skills on display in his books...basically just data. That said. Do you know one thousand mathematicians like William Quirin? Did he claim to be a gambler?

I personally know at least 9 mathematician/horse players that have made a substantial living playing the ponies. I also know players with minimal math skills that are consistent winners. I can also tell you that they have learned from their association with the mathematicians.

Why are you so hung up on the math thing.

"Remember...advanced math degrees are not a prerequisite to being an expert horseplayer"

What have I forgotten that I need to remember?

And what is your definition of an expert horse player? Can't an expert horse player posses math skills beyond your own? Or is that unacceptable to you? Should the mathematician/horse player shun his skills, because you say that horse racing is not a math problem?

Well, it is a math problem to me. It might not be the type of math problem you have been exposed to. I can assure you that my math skills coupled with my handicapping knowledge allows me to view racing in a way that is foreign to you. As I have said often that I function in a probabilistic framework. Unfortunately there is a misconception in the meaning of probability theory and statistical analysis.

As a Bayesian I use data in ways that are unimaginable to you. My Bayesian Belief Networks describe causality and not correlations. Correlations in my world are a baby step in the overall process.

Knowledge, in spite of what has been said, is power and the right knowledge is even more powerful. It would be much easier for me to teach a a skilled mathematician about handicapping/horse racing than to teach a horse player/handicapper some advanced math.

For example, you'd probably scratch your head and call me a liar if I said to you that the probability of a continuous random variable equal to a constant is zero. Or that we can discuss causality in Bayesian terms. That's right causality!

Isn't a wonder that I can take your insights and incorporate them into my model effortlessly. Wouldn't you like to take a longshot angle and determine at what level does it have value with respect to the field? Or as GT inferred in his post instead of disproving the null hypothesis, one can instead assign a probability to numerous hypothesis's.

Should I ignore what Jeff Platt's interest in the beta transformed linear opinion pool, and how he intended to calibrate probabilities and assess sharpness, because as you say horse racing is not a math problem. Or ignore Benter's use of ordered statistics and solving a complex problem using Markov Chain Monte Carlo techniques.

Should I not wonder whenever someone mentions Confidence Intervals, do they understand that they are not what they think they are?

As Cratos said knowledge is a good thing, not an affliction.

Mike (Dr Beav)

[Tom]

:sleeping:

Wake me up when its over.

So far, one guy has a thought and all we have had is an attack on it with nothing to back it up but blather.

Give me a post from CJ about how he came up with a variant any day.
Or one of Suff's classic race analyses.

[CincyHorseplayer]

Not doubting Quirin's math skills, but there is little math skills on display in his books...basically just data. That said. Do you know one thousand mathematicians like William Quirin? Did he claim to be a gambler?

I personally know at least 9 mathematician/horse players that have made a substantial living playing the ponies. I also know players with minimal math skills that are consistent winners. I can also tell you that they have learned from their association with the mathematicians.

Why are you so hung up on the math thing.

"Remember...advanced math degrees are not a prerequisite to being an expert horseplayer"

What have I forgotten that I need to remember?

And what is your definition of an expert horse player? Can't an expert horse player posses math skills beyond your own? Or is that unacceptable to you? Should the mathematician/horse player shun his skills, because you say that horse racing is not a math problem?

Well, it is a math problem to me. It might not be the type of math problem you have been exposed to. I can assure you that my math skills coupled with my handicapping knowledge allows me to view racing in a way that is foreign to you. As I have said often that I function in a probabilistic framework. Unfortunately there is a misconception in the meaning of probability theory and statistical analysis.

As a Bayesian I use data in ways that are unimaginable to you. My Bayesian Belief Networks describe causality and not correlations. Correlations in my world are a baby step in the overall process.

Knowledge, in spite of what has been said, is power and the right knowledge is even more powerful. It would be much easier for me to teach a a skilled mathematician about handicapping/horse racing than to teach a horse player/handicapper some advanced math.

For example, you'd probably scratch your head and call me a liar if I said to you that the probability of a continuous random variable equal to a constant is zero. Or that we can discuss causality in Bayesian terms. That's right causality!

Isn't a wonder that I can take your insights and incorporate them into my model effortlessly. Wouldn't you like to take a longshot angle and determine at what level does it have value with respect to the field? Or as GT inferred in his post instead of disproving the null hypothesis, one can instead assign a probability to numerous hypothesis's.

Should I ignore what Jeff Platt's interest in the beta transformed linear opinion pool, and how he intended to calibrate probabilities and assess sharpness, because as you say horse racing is not a math problem. Or ignore Benter's use of ordered statistics and solving a complex problem using Markov Chain Monte Carlo techniques.

Should I not wonder whenever someone mentions Confidence Intervals, do they understand that they are not what they think they are?

As Cratos said knowledge is a good thing, not an affliction.

Mike (Dr Beav)

All you've proved in this post is that you are long winded and can make obscure references to math related problems that have no bearing on horse racing whatsoever.

And you backed Thask into not being able to say no as a solution.

This is exactly the crap people have a problem with regarding math minded individuals regarding racing.You think you have the monopoly on the opinion about the sport and the validity of intellectual concepts regarding it.You assume this presumptuous status as if it's owed you.And that my arrogant friend will incite some hatred.

Math is relevant for coming up with an accurate speed figure.An accurate pace figure.How we look at odds and bet.But it has absolutely zero to do with interpreting probable pace.Understanding a horse's inherent abilities within the range of form and conditions they raced against.Understanding the relationships of objective and subjective factors in relation to one another.

Yet there you sit.Thinking this domain is at your whimsy.That headstrong self appointed ability is.

I don't know what's going on with this board.This place is normally great.There are too many people that have all the answers and want to tell us how awesome they are!!!

[PaceAdvantage]

All you've proved in this post is that you are long winded and can make obscure references to math related problems that have no bearing on horse racing whatsoever.And you know this how?

This is exactly the crap people have a problem with regarding math minded individuals regarding racing.You think you have the monopoly on the opinion about the sport and the validity of intellectual concepts regarding it.You assume this presumptuous status as if it's owed you.And that my arrogant friend will incite some hatred.No he didn't. He never stated he has a monopoly on the validity of the intellectual concepts regarding racing. In fact, he stated quite plainly, and I quote:

"I also know players with minimal math skills that are consistent winners."

There are too many people that have all the answers and want to tell us how awesome they are!!!Look in the mirror my friend. Although I will grant you that you didn't tell us how awesome you are...

[thaskalos]

No he didn't. He never stated he has a monopoly on the validity of the intellectual concepts regarding racing.

You are right...he didn't.

But he DID say that the horse racing related opinions of another poster were BULLSHIT!

This thread is not about "higher mathematics"...it is about assessing VALUE in horse races.

Do we have to possess advanced mathematics degrees in order to participate in discussions such as this...and should we have our opinions categorized as BULLSHIT because we don't possess Dr. Beav's academic credentials?

[CincyHorseplayer]

And you know this how?

No he didn't. He never stated he has a monopoly on the validity of the intellectual concepts regarding racing. In fact, he stated quite plainly, and I quote:

"I also know players with minimal math skills that are consistent winners."

Look in the mirror my friend. Although I will grant you that you didn't tell us how awesome you are...

Pace.There are inferences scattered throughout his post.And I can read through what it's saying.I thought it was overbearing and elitist.That's my opinion.And at the length of the post I thought it was irrelevant adding on.The guy is obviously a smart individual,but it doesn't mean he can throw people under the bus.And I don't expect that from you either.I don't think my reaction was too much,do you?Guy came across overbearing and I gave him an honest answer.That's what it's about isn't it?

I had somebody saying I was too cheesy the other day for embracing my fellow horseplayer on your tip thread.This after meeting up with 3 board members on Saturday and having a great time.I feel like I can't win here man.I get critical,you chastise me.I get loving somebody calls me a dork.What gives?

[PaceAdvantage]

But he DID say that the horse racing related opinions of another poster were BULLSHIT!His first three bullshits were directed at said poster's statistical and mathematical opinions. His last bullshit was indeed aimed at the poster's horse racing related opinion.

Bullshit is just a stronger word for "I don't agree."

I never said this board has to be angelic.

[PaceAdvantage]

What gives?Just go with the flo man... :lol:

[thaskalos]

His first three bullshits were directed at said poster's statistical and mathematical opinions. His last bullshit was indeed aimed at the poster's horse racing related opinion.

Bullshit is just a stronger word for "I don't agree."

I never said this board has to be angelic.
Have you determined what his argument with ME is? Because it's a mystery to me...

Read my posts in this thread, and tell me where I have erred.

[CincyHorseplayer]

Just go with the flo man... :lol:

Hey Ive been on here a little bit.I just want you to know I'm not Joe trouble.But can unleash on somebody on occasion!!And I'm allowed!:cool:

[Nitro]

Yes the game we all seem to enjoy is about assimilating information, but not necessarily from the perspective of Academia. Horses are not mechanical machines!

Anyone that believes that an “exact” science like Math can magically conjure up consistent value based decisions when applied to inexact, ever changing, completely subjective GAME like horse racing is in for some big surprises when it comes to actually tallying up the net gain. This simple but true premise exists because all of the intangibles related to the effects of handling, caring for and managing the animal are absolutely unquantifiable.

Of course those in the Math corner will conveniently ignore these ambiguities by implying that their figures, statistical models and probability theories (that are based solely on the past) can somehow provide the answers for determining the future. Maybe its possible with only part the picture to achieve a modest positive ROI. But I thought this discussion was about “Value”.

How do the mathematicians rationalize their betting strategies when they’re all arriving at the same conclusions for playing? Or are they? Maybe these exacting results are open for interpretation? Possibly like all of the other information that’s available to those who might enjoy the game without a Harvard degree.

[TrifectaMike]

Yes the game we all seem to enjoy is about assimilating information, but not necessarily from the perspective of Academia. What does this mean? Horses are not mechanical machines! I don't believe anyone has implied they are machines.

Anyone that believes that an “exact” science like Math can magically conjure up consistent value based decisions when applied to inexact, ever changing, completely subjective GAME like horse racing is in for some big surprises when it comes to actually tallying up the net gain. I agree. It is a subjective game. And I approach racing exactly from that point of view. This simple but true premise exists because all of the intangibles related to the effects of handling, caring for and managing the animal are absolutely unquantifiable. True, but there exist evidence to guide you.

Of course those in the Math corner will conveniently ignore these ambiguities by implying that their figures, statistical models and probability theories (that are based solely on the past) Bayesian probability models are based on evidence...not the past. can somehow provide the answers for determining the future. Maybe its possible with only part the picture to achieve a modest positive ROI. And you know this how? But I thought this discussion was about “Value”. It is all about value and value can be quantified.

How do the mathematicians rationalize their betting strategies when they’re all arriving at the same conclusions for playing? Or are they? Maybe these exacting results are open for interpretation? There are no exacting results. Possibly like all of the other information that’s available to those who might enjoy the game without a Harvard degree. Enjoy the game. I have not implied that any degree(s) are required.

I find it amusing that you introduce your professional actuary to lecture on how to employ the 'proper' techniques as used by the insurance companies...group this and group that...run it like an insurance company...otherwise you are doomed to failure. It is laughable.

When did understanding math become such a bad thing?

Mike (Dr Beav)

[Robert Goren]

There are people out there using mathematical models betting lots of races in a day turn a relative small positive ROI into some pretty good money on a yearly basis. Some never even look at a horse race. There are also some people out there who go the the track look over the horse in the paddock and look at the DRF and pass a bunch of races, but still turn a pretty good profit at the end of the year. It can be done either way, but both require a lot of work and a certain amounts of smarts.

[Tom]

For all the great depth ok knowledge represented here, this thread has offered me no value whatsoever. Yet again.

[TrifectaMike]

Math is relevant for coming up with an accurate speed figure.An accurate pace figure.How we look at odds and bet.But it has absolutely zero to do with interpreting probable pace.Understanding a horse's inherent abilities within the range of form and conditions they raced against.Understanding the relationships of objective and subjective factors in relation to one another.


You are not serious or are you?

Mike (Dr Beav)

[thaskalos]

Not doubting Quirin's math skills, but there is little math skills on display in his books...basically just data. That said. Do you know one thousand mathematicians like William Quirin? No...and I don't know one like Phil Bull either. Did he claim to be a gambler? Well...he wrote books on how to beat the races. When someone writes books on how to beat the stock market...shoudn't we assume that he/she is a player?

I personally know at least 9 mathematician/horse players that have made a substantial living playing the ponies. I also know players with minimal math skills that are consistent winners. I can also tell you that they have learned from their association with the mathematicians. I don't doubt it. And I would also guess that MATHEMATICIANS could learn plenty from some non-credentialed handicappers. So it works BOTH ways.

Why are you so hung up on the math thing. I am not hung up on math at all! In fact...my approach to this game has a very distinct mathematical tilt.

"Remember...advanced math degrees are not a prerequisite to being an expert horseplayer" Yes...I said this. And I believe it to be very true.

What have I forgotten that I need to remember? That you don't have ALL the answers to this game...and that you should not be so quick to dismiss the opinions of others. We are all STUDENTS of this game. Our game has no masters...

And what is your definition of an expert horse player? Any regular player who shows a long-term profit in this game is a genuine EXPERT in MY book...and his opinion deserves respect. Can't an expert horse player posses math skills beyond your own? Point out to me where I stated that. Or is that unacceptable to you? I accept it cheerfully. Should the mathematician/horse player shun his skills, because you say that horse racing is not a math problem? Not at all! All of us try to use our particular skills in order to better understand this game. It seems perfectly logical to me...

Well, it is a math problem to me. It might not be the type of math problem you have been exposed to. You don't know what I have been exposed to. This is just your smug superiority complex talking...I can assure you that my math skills coupled with my handicapping knowledge allows me to view racing in a way that is foreign to you. And I, in turn, can assure YOU that MY skills allow ME to view this game in a way that is foreign to you. SO WHAT? As I have said often that I function in a probabilistic framework. Unfortunately there is a misconception in the meaning of probability theory and statistical analysis.

As a Bayesian I use data in ways that are unimaginable to you. My Bayesian Belief Networks describe causality and not correlations. Correlations in my world are a baby step in the overall process.

Knowledge, in spite of what has been said, is power and the right knowledge is even more powerful. It would be much easier for me to teach a a skilled mathematician about handicapping/horse racing than to teach a horse player/handicapper some advanced math. The player/handicapper really has NO need for some advanced math...so I don't know what you are trying to say here. For your information...I happen to also be an artist, whose paintings have received much critical acclaim. And I would bet my bottom dollar that you could teach ME advanced math A LOT quicker than I could teach YOU how to paint and sell your works. Again...SO WHAT!

For example, you'd probably scratch your head and call me a liar if I said to you that the probability of a continuous random variable equal to a constant is zero. Or that we can discuss causality in Bayesian terms. That's right causality!

Isn't a wonder that I can take your insights and incorporate them into my model effortlessly. Yes, it is. Wouldn't you like to take a longshot angle and determine at what level does it have value with respect to the field? I already do that. Or as GT inferred in his post instead of disproving the null hypothesis, one can instead assign a probability to numerous hypothesis's.

Should I ignore what Jeff Platt's interest in the beta transformed linear opinion pool, and how he intended to calibrate probabilities and assess sharpness, because as you say horse racing is not a math problem. I never asked you to ignore ANYTHING. Or ignore Benter's use of ordered statistics and solving a complex problem using Markov Chain Monte Carlo techniques.

Should I not wonder whenever someone mentions Confidence Intervals, do they understand that they are not what they think they are?

As Cratos said knowledge is a good thing, not an affliction. Knowledge is good thing...NOT an affliction. But "true" knowledge comes in many different forms. It is not displayed only by our mathematical capabilities.

Mike (Dr Beav)
Regards,

Thaskalos

[Cratos]

Regards,

Thaskalos
I don't understand the argument against the use of math and statistics in horserace handicapping. I might add another useful science, physics.

[Cratos]

IMO...for every ONE mathematician like Phil Bull, there are ONE THOUSAND mathematicians like WILLIAM QUIRIN...who, in spite of his math skills and handicapping knowledge, proved that - although he could TALK a good game - he ultimately lacked the flexible thinking needed to beat this game.

And for every Pittsburgh Phil who was a winner and not a mathematician, there are thousands of other non-mathematicians who go home more often than not with empty pockets.

[GameTheory]

I don't understand the argument against the use of math and statistics in horserace handicapping. I might add another useful science, physics.No one has argued that.

And for every Pittsburgh Phil who was a winner and not a mathematician, there are thousands of other non-mathematicians who go home more often than not with empty pockets.Just as there are thousands of mathematicians who go home with empty pockets. Do you have an evidence that percentage-wise the math guys are doing any better?

[thaskalos]

I don't understand the argument against the use of math and statistics in horserace handicapping. I might add another useful science, physics.
I never made such an argument.

THIS is my argument, in a nutshell:

NOBODY has this game completely figured out...and we ALL have a lot yet to learn...

The poster who goes by the name of "PONDMAN" has stated that he is a winning player...and I have no reason to disbelieve him. So, when Pondman made some comments concerning his opinions about the game, and the way he plays it...he deserved better than to have TrifectaMike declare that these comments were BULLSHIT...MORE BULLSHIT...AND STILL MORE BULLSHIT!

If this was a thread about "the function of higher mathematics in the handicapping and betting process"...then I would never argue with the likes of TrifectaMike -- just like I haven't tangled with him in the past, in the various mathematically oriented threads which pop-up here from time to time.

But this thread is about VALUE in horse racing; and value is not ascertained only through "advanced" mathematical means.

Just because we have different ways of approaching this game -- or determining the value therein -- than other winning players...does that make their racing opinions "BULLSHIT"?

[Elliott Sidewater]

Thaskalos, as usual I agree with you 100%. Horse racing is an endeavor of continual learning, discovery, and trial and error. This is not a universally precise figure, but in my own record of over 600 hand recorded bets, in 13% of the races the results (win only) were unexplainable even in retrospect. There are some undefinable, unknowable, and unpredictable factors that render our best efforts as handicappers impotent some of the time. I think that any reasonable approach to trying to predict or project "value" before a race is run has to set aside some percentage of the bets as lost in advance. However, that should not be taken to mean that long term profit is impossible, that's not what I'm implying at all.

As far as the BS remark goes, if I read something ridiculous on this forum I might laugh to myself but I would never impugn the reputation of someone I don't know, especially in writing. Why not just give a better explanation than the OP did, without all the vitriol? After this year's Triple Crown results, 3 winners with double digit odds, I thought we might see a bit more humility from the posters. Yeah right, dream on!;)

[Robert Fischer]

"Value" is "Risk", masquerading as an elusive horseplaying concept.

various types of Analysis(mathematical, intuitive, bullshit? ;), etc...) require insight, and also virtue

[Elliott Sidewater]

Agreed, risk or risk management is probably a better term.

[Robert Goren]

Agreed, risk or risk management is probably a better term.Isn't that what got Lehman in trouble?

[Cratos]

I never made such an argument.

THIS is my argument, in a nutshell:

NOBODY has this game completely figured out...and we ALL have a lot yet to learn...

The poster who goes by the name of "PONDMAN" has stated that he is a winning player...and I have no reason to disbelieve him. So, when Pondman made some comments concerning his opinions about the game, and the way he plays it...he deserved better than to have TrifectaMike declare that these comments were BULLSHIT...MORE BULLSHIT...AND STILL MORE BULLSHIT!

If this was a thread about "the function of higher mathematics in the handicapping and betting process"...then I would never argue with the likes of TrifectaMike -- just like I haven't tangled with him in the past, in the various mathematically oriented threads which pop-up here from time to time.

But this thread is about VALUE in horse racing; and value is not ascertained only through "advanced" mathematical means.

Just because we have different ways of approaching this game -- or determining the value therein -- than other winning players...does that make their racing opinions "BULLSHIT"?


I worded my post incorrectly to you because it inferred that you were making an argument against math and statistics use in the handicapping of horseracing. However there is an argument being put forth and if not this entire thread is illogical.

Therefore my question becomes how one use math and statistics to handicap racehorses and is it a beneficial endeavor to use such disciplines?

Horseracing along with baseball is probably the two most statistical documented sports in America and if you have the knowledge of math and statistics and want to analyze either of those sports you will go with your strengths.

But this will not preclude the requisite to have a good fundamental knowledge of either horseracing or baseball.

What is a speed figure but a statistical metric and that metric appears to be the Holy Grail in horseracing today. There are allegedly “good” figures and “poor” figures and I would surmise that the quality of the figure has something to do with the knowledge of horseracing and the knowledge of math and statistics.

[Capper Al]

:sleeping:

Wake me up when its over.

So far, one guy has a thought and all we have had is an attack on it with nothing to back it up but blather.

Give me a post from CJ about how he came up with a variant any day.
Or one of Suff's classic race analyses.

Yeah CJ, how do you come up with a variant?

[Capper Al]

Hello,

Wake up out there. This thread was started with the idea that we could explain mathematically why the slower horse will win at times. The example works. It doesn't explain how we are to figure out the variants or what the slope of the curve is or what one standard deviation is. That was within the given for the purpose of the example. Figuring out the givens is the horse sense that one must have. Once you have it, you can use it. Math helps like being able to realize that the top speed figure last race wins 25.5%. But one still has to appreciate the value of speed with all the other factors involved. I still lean toward intuition for this, although the math does help.

[Robert Fischer]

any reasonable approach to trying to predict or project "value" before a race is run has to set aside some percentage of the bets as lost in advance.



:eek:


:ThmbUp:

:jump::jump::jump:

[Overlay]

This is not a universally precise figure, but in my own record of over 600 hand recorded bets, in 13% of the races the results (win only) were unexplainable even in retrospect. There are some undefinable, unknowable, and unpredictable factors that render our best efforts as handicappers impotent some of the time. I think that any reasonable approach to trying to predict or project "value" before a race is run has to set aside some percentage of the bets as lost in advance.
Your quote reminds me of an observation by the late Dick Mitchell in connection with formulating an oddsline:

"The trick to making a betting line is to first separate the contenders from the non-contenders. Typically, we give the group of non-contenders a 20% probability of winning. In other words, we'll accept the notion that 'stuff happens' [as he euphemistically referred to it] at a racetrack. That is, 20% of the time, we can't explain the results of a thoroughbred race. The size of the field determines the size of the 'SH' pile. For typical races [at the time he wrote] of eight to twelve competitors, the pile will be 20%. As the size of the field gets smaller, we'll reduce the size of the 'SH' pile by 4% for each horse under the threshold total of eight."

[Robert Fischer]

Your quote reminds me of an observation by the late Dick Mitchell in connection with formulating an oddsline:

"The trick to making a betting line is to first separate the contenders from the non-contenders. Typically, we give the group of non-contenders a 20% probability of winning. In other words, we'll accept the notion that 'stuff happens' [as he euphemistically referred to it] at a racetrack. That is, 20% of the time, we can't explain the results of a thoroughbred race. The size of the field determines the size of the 'SH' pile. For typical races [at the time he wrote] of eight to twelve competitors, the pile will be 20%. As the size of the field gets smaller, we'll reduce the size of the 'SH' pile by 4% for each horse under the threshold total of eight."

good stuff

[Overlay]

good stuff
My main difficulty was that, when a longshot lit up the board, it was generally one of the horses from my "SH" pile. That's why I started assigning fair odds to every horse in a field (even though winning probabilities tend to get less precise the higher up you go on the odds spectrum). I was trying to find a happy medium between just betting longshots randomly to make sure that I didn't miss a boxcar payoff on the one hand, and completely disregarding the chances of any of my "non-contenders" (as Mitchell suggested) on the other. Basing my fair odds on multi-factor performance statistics rather than on subjective opinion, or on a single angle or overriding variable (such as speed) (which usually ends up getting overbet), has helped me with both selectivity and accuracy.

[TrifectaMike]

Statistics is suitable for recording human behavior.

It will never be able to quantify:

...a 6 month layoff
...being shipped 5oo miles
...Chantel Southerland
...Tyler Baze dropping his whip

All of these could be a + or -.
You know this to be an absolute or just Blowing smoke.
To be successfull you need to think as a 4th level quant trader waiting for an unknown disaster, and being willing to loose 4 bets in order to have that 20-1 shot win. You know this to be an absolute or just more smoke.

Linear style regression will combine a star jockey, a high percentage trainer, and a high speed figure. This will result in a best bet. And these usually turn into trainwrecks. You know this to be an absolute or just more smoke.

Do horses with 6 month layoffs, being shipped 500, ridden by girls ever win at 40-1? Yes they do. It would be best to wait for these unknowns to show up They also win at 2-1 through 100+ to 1...more smoke
.

You can also replace Bullshit with Chicken Shit, if that makes Thask happy.

Mike (Dr Beav)

[TrifectaMike]

This looks like an opinion to me...

And I would have to know a little more about the bettor who made it, before I categorize it as..."BULLSHIT"!

"Remember...advanced math degrees are not a prerequisite to being an expert horseplayer." Categorizing something as BULLSHIT is NOT a mathematical term. This sorta like a lib using the race card when they have nothing else to say.

ALL winning players can claim this title...

And Pondman professes to be a winning player...

Thask, what is REALLY on your mind?

Mike (Dr Beav)

[TrifectaMike]

OK, last try.



We are dealing with an AVERAGE time that varies hence the word "average".
We determine on AVERAGE, horse A is one length faster than horse B. For example, horse A's average six furlong time is 109 then horse B's six furlong time is assumed to be 109.2. Determine the standard deviation for horse A and B from the same data. Then look at Post 11. The dice thingy and Central Limit Theorem -> Normal distribution.
We have two magical dice that represent the equally distributive standard deviations for each horse on their numbers. The first die A represents horse A with numbers 2 thru 7. Die A will have an average roll of 4.5. The second die B represents horse B with numbers 1 thru 6. Die B will have an average roll of 3.5.
Our expectations of one greater between the averages of both dice correlates to horse A being one length faster than horse B. We randomly throw the dice to represent what could be the possible outcome between these two horses in our artificially made up contest.
We get the chart in post #1 showing the die representing horse A beats the die representing horse B and visa versa.
On average for 6 rolls of each number of Die 'A'

----Roll on Die 'A' ---- 7-6-5-4-3-2: 36 rolls. ( 6 for each number/side )

Die 'A' is higher ------ 6-5-4-3-2-1: 21 'A' wins
Ties --------------------0-1-1-1-1-1: 5 ties
Die 'A' is lower --------0-0-1-2-3-4: 10 'B' wins


Mike (Dr Beav)

[TrifectaMike]

Thaskalos, as usual I agree with you 100%. Horse racing is an endeavor of continual learning, discovery, and trial and error. This is not a universally precise figure, but in my own record of over 600 hand recorded bets, in 13% of the races the results (win only) were unexplainable even in retrospect. There are some undefinable, unknowable, and unpredictable factors that render our best efforts as handicappers impotent some of the time. I think that any reasonable approach to trying to predict or project "value" before a race is run has to set aside some percentage of the bets as lost in advance. However, that should not be taken to mean that long term profit is impossible, that's not what I'm implying at all. That's the reason probability distributions are used.

One has to love this thread. Posters using basic math to infer that math is not an essential part of the game.

Mike (Dr Beav)

[TrifectaMike]

I worded my post incorrectly to you because it inferred that you were making an argument against math and statistics use in the handicapping of horseracing. However there is an argument being put forth and if not this entire thread is illogical.

Therefore my question becomes how one use math and statistics to handicap racehorses and is it a beneficial endeavor to use such disciplines?

Horseracing along with baseball is probably the two most statistical documented sports in America and if you have the knowledge of math and statistics and want to analyze either of those sports you will go with your strengths.

But this will not preclude the requisite to have a good fundamental knowledge of either horseracing or baseball.

What is a speed figure but a statistical metric and that metric appears to be the Holy Grail in horseracing today. There are allegedly “good” figures and “poor” figures and I would surmise that the quality of the figure has something to do with the knowledge of horseracing and the knowledge of math and statistics.To many this is a fait accompli.

Mike (Dr Beav)

[Tom]

Less talk, more walk.
Show us.

[Capper Al]

One has to love this thread. Posters using basic math to infer that math is not an essential part of the game.

Mike (Dr Beav)

I'm not saying that math is not an essential part of the game. There's just too many variables to be precise about a prediction or value. The best one can have after all the math is a good feel about the race.

[classhandicapper]

For those that lack sophisticated mathematical skills or believe that the complexities of the game are such that experience and intuition do a better job of handling many of the very unique situations that come up each day, I use a somewhat simple method for determining value.

If you start with the assumption that the public does a pretty good job with the odds/values (which it does), I think if you simply rank the horses (or just contenders) in order of probability of winning and compare that ranking to the ranking on the odds board you have some very useful information.

Any horse that is mis-ranked (assuming you are doing a good job of ranking them) is probably going to be break even or better depending on how badly mis-ranked it is.

If you want you can take it a step further and rank them as 1, 1a, 2, 3 ,4 5, 5a, 5b where 1 and 1a are very similar etc...

The other simple approach is to identify situations that you KNOW the public often misunderstands. For instance, if you KNOW trainer "X" is terrible with layoffs and has one that's getting bet down to 8-5 favoritism, you can simply structure a bet around betting against that horse. If you have a portfolio of situations like that, you'll have a way determining value often enough.

[TrifectaMike]

For those that lack sophisticated mathematical skills or believe that the complexities of the game are such that experience and intuition do a better job of handling many of the very unique situations that come up each day, I use a somewhat simple method for determining value.

If you start with the assumption that the public does a pretty good job with the odds/values (which it does), I think if you simply rank the horses (or just contenders) in order of probability of winning and compare that ranking to the ranking on the odds board you have some very useful information.

Any horse that is mis-ranked (assuming you are doing a good job of ranking them) is probably going to be break even or better depending on how badly mis-ranked it is.

If you want you can take it a step further and rank them as 1, 1a, 2, 3 ,4 5, 5a, 5b where 1 and 1a are very similar etc...

The other simple approach is to identify situations that you KNOW the public often misunderstands. For instance, if you KNOW trainer "X" is terrible with layoffs and has one that's getting bet down to 8-5 favoritism, you can simply structure a bet around betting against that horse. If you have a portfolio of situations like that, you'll have a way determining value often enough.

I have no argument with the above. Actually it is clever. And if it works within your domain, more power to you. As I said in an earlier thread, I know many winning players that have a minimal understanding of math. (That is not too imply you have a minimal understanding of math}

Mike (Dr Beav)

[TrifectaMike]

I'm not saying that math is not an essential part of the game. There's just too many variables to be precise about a prediction or value. The best one can have after all the math is a good feel about the race.

The comment was not directed at you or anyone in particular. However, you did start the thread and asked about determining value, and you described an approach (The dice thingy...I'm picking up the thingy word from grandson). and asked how it worked. In post 11 I described in probability terms what you were simulating with the dice thingy.

Your asked how to get the standard deviation, and I told you how. If the question is:

Is it sufficient to derive an accurate line...my answer would be it depends on how one arrives at the expected running time of the race (and variance) and the expected running time of ALL the horses and their variances.

Mike (Dr Beav)

[Jeff P]

I'm not saying that math is not an essential part of the game. There's just too many variables to be precise about a prediction or value. The best one can have after all the math is a good feel about the race.

The above statement might lead someone looking to improve their game to the mistaken conclusion that value can't be quantified. I can't let that happen without responding to it because nothing could be further from the truth.

My experience is that not only can value can be quantified... but that value can quantified accurately enough that it certainly can be put to good use in real world live play.

In the PAIHL contest held right here at Paceadvantage.com this past spring, over the course of the 15 week regular season, the JCapper team I was on scored highest in total money won among the 50 teams entered. Anyone remember that?

How do you think that happened?

I'll tell you.

It happened because we focused on quantifying value instead of making selections.

Now here's the important part. The quantifying of value that we did wasn't done through feel or intuition. Over the entire course of the PAIHL contest I never once looked at past performances and hope my teammates never once looked at them either. (I'm not kidding about that last part.)

We were able to accomplish what we did through number crunching... using math and algorithms exclusively.

Now before anyone accuses me otherwise, here's what I'm not saying...

I'm not saying my way (with math and algorithms) is the only way.

I'm also not saying the other way (using feel and intuition) can't possibly work.

There are a lot of ways to beat this game. If intuition and feel is working for you, and you are happy with your results - I say ENJOY and have at it!

Here's what I am saying:

When someone tries to tell others on a public board that value can't be quantified I have to call them on it... because nothing could be further from the truth.

-jp

.

[thaskalos]

The above statement might lead someone looking to improve their game to the mistaken conclusion that value can't be quantified. I can't let that happen without responding to it because nothing could be further from the truth.

My experience is that not only can value can be quantified... but that value can quantified accurately enough that it certainly can be put to good use in real world live play.

In the PAIHL contest held right here at Paceadvantage.com this past spring, over the course of the 15 week regular season, the JCapper team I was on scored highest in total money won among the 50 teams entered. Anyone remember that?

How do you think that happened?

I'll tell you.

It happened because we focused on quantifying value instead of making selections.

Now here's the important part. The quantifying of value that we did wasn't done through feel or intuition. Over the entire course of the PAIHL contest I never once looked at past performances and hope my teammates never once looked at them either. (I'm not kidding about that last part.)

We were able to accomplish what we did through number crunching... using math and algorithms exclusively.

Now before anyone accuses me otherwise, here's what I'm not saying...

I'm not saying my way (with math and algorithms) is the only way.

I'm also not saying the other way (using feel and intuition) can't possibly work.

There are a lot of ways to beat this game. If intuition and feel is working for you, and you are happy with your results - I say ENJOY and have at it!

Here's what I am saying:

When someone tries to tell others on a public board that value can't be quantified I have to call them on it... because nothing could be further from the truth.

-jp

.
I agree completely with what you say here Jeff...and I applaud the manner with which you said it.

Here is a point I would like to make...and I ask for forgiveness if it takes me outside the parameters set by this thread.

I have been a member here for a few years now and, during that time, have come in contact - through the private message function of this site - with more than a few intelligent, WINNING players who, while applauding ME for offering MY thoughts on the game, confess to me that they too would like to share THEIR substantial knowledge with this board...but they don't want to have their good intentions derailed by the attacks and the mud-slinging that sometimes go on around here.

Now, I don't profess to be the quintessential "nice guy"...as the not too distant Zenyatta discussions here clearly indicate...

But it bothers me that other very knowledgeable players are dissuaded from sharing this knowledge with others here who so badly need it...all as a result of the vitriolic atmosphere which sometimes surrounds our discussions.

There are some brilliant minds here...but why not use that brilliance to explain and educate...rather than to attack and impugn?

[Capper Al]

The comment was not directed at you or anyone in particular. However, you did start the thread and asked about determining value, and you described an approach (The dice thingy...I'm picking up the thingy word from grandson). and asked how it worked. In post 11 I described in probability terms what you were simulating with the dice thingy.

Your asked how to get the standard deviation, and I told you how. If the question is:

Is it sufficient to derive an accurate line...my answer would be it depends on how one arrives at the expected running time of the race (and variance) and the expected running time of ALL the horses and their variances.

Mike (Dr Beav)

TrifectaMike,

I don't recall asking for help. The only thing I intended to do was through out an example where if we could break it down to one length deviation between only two horses that the slower one would still win a third of the time. Of course, there was a lot of givens to simplify the example. It would be more complex in the real world. I tip my hat to anyone who can do it over the long run.

[JeremyJet]

RAHYSTRADA 9/2
SIERRA ALPHA 99
MORYBA 6
SMART BID 3
SILVER NOUNTAIN 99
GUYS REWARD 11
DARK COVE 7
COURT VISION 9/2
We'll have to wait for the actual odds, but based on the M/L, RAHYSTRADA and DARK COVE look like the potential overlays. This is just my humble opinion, and someone looking at the same data (Thoro-Graph) might come up with something different.

I know very little about math and don't use a computer to handicap. Let's see what happens.

Regards,

JeremyJet

[Capper Al]

JCapper was impressive in the contest. Algorithms can work and can be tested using statistical methods. It's still difficult to know for sure the value of a horse in most races.

[cj]

There are some brilliant minds here...but why not use that brilliance to explain and educate...rather than to attack and impugn?

While I'm not trying to lump myself in as a brilliant mind, I'm often torn when it comes to contributing and helping others. I generally try to do just that, but I often wonder why. This sport, after all, is a competition for each other's dollars.

[GameTheory]

While I'm not trying to lump myself in as a brilliant mind, I'm often torn when it comes to contributing and helping others. I generally try to do just that, but I often wonder why. This sport, after all, is a competition for each other's dollars.Because you get satisfaction from both helping others and also from seeing yourself "being right" at the betting windows. Which one is more lasting?

Others seem to get satisfaction from making fun of and insulting others (we all do from time to time, but some seem to really revel in it). How lasting is that satisfaction?

[Capper Al]

For those that lack sophisticated mathematical skills or believe that the complexities of the game are such that experience and intuition do a better job of handling many of the very unique situations that come up each day, I use a somewhat simple method for determining value.

If you start with the assumption that the public does a pretty good job with the odds/values (which it does), I think if you simply rank the horses (or just contenders) in order of probability of winning and compare that ranking to the ranking on the odds board you have some very useful information.

Any horse that is mis-ranked (assuming you are doing a good job of ranking them) is probably going to be break even or better depending on how badly mis-ranked it is.

If you want you can take it a step further and rank them as 1, 1a, 2, 3 ,4 5, 5a, 5b where 1 and 1a are very similar etc...

The other simple approach is to identify situations that you KNOW the public often misunderstands. For instance, if you KNOW trainer "X" is terrible with layoffs and has one that's getting bet down to 8-5 favoritism, you can simply structure a bet around betting against that horse. If you have a portfolio of situations like that, you'll have a way determining value often enough.

I agree. The public does a good job as do the Public handicappers also.

[Cratos]

But it bothers me that other very knowledgeable players are dissuaded from sharing this knowledge with others here who so badly need it...all as a result of the vitriolic atmosphere which sometimes surrounds our discussions.

There are some brilliant minds here...but why not use that brilliance to explain and educate...rather than to attack and impugn?

My background academically is Mechanical Engineering and Econometrics at the graduate level and I am not revealing this to indicate a “brilliant” mind, but to say those disciplines has allowed to me to develop algorithms and a predictive model for handicapping thoroughbred race horses.

It has taken many hours and many lost wagers to get the model to a reliable point of prediction to maximize my profits in this game.

Would I give out the algorithms for public use and the answer is an astounding no because this game is about bettor against bettor and any bettor who has an edge regardless how slight is ahead of the other bettors.

However I will and do talk to concepts (energy expenditure is one and turn impact is another) that might help or inform other handicappers.

[Jeff P]

Because you get satisfaction from both helping others and also from seeing yourself "being right" at the betting windows. Which one is more lasting?...I'd argue BOTH provide lasting enjoyment. I know that both are a big part of why I enjoy this board as much as I do.


I agree completely with what you say here Jeff...and I applaud the manner with which you said it.Thank you. I applaud the way you've conducted yourself here as well.


-jp

.

[Capper Al]

RAHYSTRADA 9/2
SIERRA ALPHA 99
MORYBA 6
SMART BID 3
SILVER NOUNTAIN 99
GUYS REWARD 11
DARK COVE 7
COURT VISION 9/2
We'll have to wait for the actual odds, but based on the M/L, RAHYSTRADA and DARK COVE look like the potential overlays. This is just my humble opinion, and someone looking at the same data (Thoro-Graph) might come up with something different.

I know very little about math and don't use a computer to handicap. Let's see what happens.

Regards,

JeremyJet



I'm interested. Good Luck.

[Capper Al]

It's good to discuss the concepts. That's why I shared CXWONG's idea. Like I said, I thought it was neat. It's good that some of you took the time to complement each other here. I really do respect you guys. That's why I shared the example.

[classhandicapper]

The one problem I have with purely mathematical solutions is that when I handicap, I constantly feel like am faced with specific situations I have either never seen before or that are at least rare enough to remain a problem. It's not that there isn't some underlying mathematical reality to it all. It's just that if I spent the rest of my life doing probability studies and trying to convert it all into formulas for creating odds lines, I'd probably die of old age before I reached my goal.

I think at the core you should probably understand the mathematics of each of the major factors and how they all interrelate, but in the end some level of intuition in the analysis will help.

It's kind of like poker.

Many of the best poker players know the math inside and out. Some play almost exclusively by feel/intuition. However, I think most of the super great players use a combination of both.

[Overlay]

I think at the core you should probably understand the mathematics of each of the major factors and how they all interrelate, but in the end I think there will almost always be some intuition involved in the analysis.
To me it boils down to the marginal value of time. I don't criticize anyone who is adept at the intuitive/qualitative/subjective side of handicapping, but I think that it's possible to get close enough to optimal with quantitative approaches (especially if adopting a bulk, probability-oriented style of selection/play, as opposed to a more selective method) to not have to engage in the second-guessing and "What am I overlooking?" thoughts that eat up unjustified (to me at least) amounts of time, and that can adversely affect the mental/psychological component of one's game.

[classhandicapper]

To me it boils down to the marginal value of time. I don't criticize anyone who is adept at the intuitive/qualitative/subjective side of handicapping, but I think that it's possible to get close enough to optimal with quantitative approaches (especially if adopting a bulk, probability-oriented style of selection/play, as opposed to a more selective method) to not have to engage in the second-guessing and "What am I overlooking?" thoughts that eat up unjustified (to me at least) amounts of time, and that can adversely affect the mental/psychological component of one's game.

I totally understand your point.

I made two wagers today (both lost) and just spent 30 minutes reviewing both races, rethinking some of the issues, second guessing my original thinking etc... It's an ongoing use of time for me. But I think it's also part of the learning process for someone like me.

I'm actually mathematically inclined, but I didn't pursue that line of study beyond basic statistics and probability despite my aptitude. So I wound up as more of an intuitive/experience/feel handicapper. However, I am always open to stats, formulas etc.... when they are within my knowledge range. At this point, I just don't feel like educating myself on advanced math. I'm 52 and out of steam if you know what I mean. ;)

[JeremyJet]

RAHYSTRADA 9/2
SIERRA ALPHA 99
MORYBA 6
SMART BID 3
SILVER NOUNTAIN 99
GUYS REWARD 11
DARK COVE 7
COURT VISION 9/2
We'll have to wait for the actual odds, but based on the M/L, RAHYSTRADA and DARK COVE look like the potential overlays. This is just my humble opinion, and someone looking at the same data (Thoro-Graph) might come up with something different.

I know very little about math and don't use a computer to handicap. Let's see what happens.

Regards,

JeremyJet



Not much there in the win pool. The only play, based on the Kelly Criterion, was a $24 full-kelly on DARK COVE. I ranked them well, but the value just wasn't there.

Is it just me, or does anyone else think the exacta paid well? $32? That seems generous considering the horse to beat came in second.

Regards,

JeremyJet

[Robert Fischer]

While I'm not trying to lump myself in as a brilliant mind, I'm often torn when it comes to contributing and helping others. I generally try to do just that, but I often wonder why. This sport, after all, is a competition for each other's dollars.
Agreed.

this will probably hurt my bottom line - but mentorship is one of the best ways to learn from a winner and really get insight into a particular method.

while we shouldn't be giving away bread, this board has it's diamonds within the rough, and the discussion can be stimulating at times to provoke an idea or two :ThmbUp:

Math in bankroll management
back on topic - i use a rather extensive approach and when i am fortunate enough to recognize certain patterns I am able to then assign a rather conservative probability to the outcome and then the wager type. Then I enter that probability into my spreadsheet. The spreadsheet does a computation that uses that probability, and my bankroll amount, and variables, and tells me how much to bet. When the probabilities are lower, so goes the wager amount, and it is therefore restrictive on certain lower percentage exotics. For example, It tells me a number that is lower than the $2 minimum. As I become familiar with certain patterns, I don't always enter the probability, as I know the ballpark figure of the wager amount. At times when i am fortunate enough to see my bankroll increase, there are increments which "allow" me to wager on a wider variety of wager types including certain exotics.

Intuitive Value / Risk
From my perspective Value and Risk have a relationship in both semantics and intuition. The definitions are online for all to see, and the basics of those fields are as well.
The intuition of "Value" as I see it, in recognizing my patterns, relates to "Risk" because it "feels" much the same. If you've ever fumbled something while cooking into the flame (hot!), or how close to the edge it was safe to stand...
that's "risk", that's "value". Cash is a lot different on the psyche, than rescuing a ravioli, or getting a better view, but bear with the rough example

[Capper Al]

The one problem I have with purely mathematical solutions is that when I handicap, I constantly feel like am faced with specific situations I have either never seen before or that are at least rare enough to remain a problem. It's not that there isn't some underlying mathematical reality to it all. It's just that if I spent the rest of my life doing probability studies and trying to convert it all into formulas for creating odds lines, I'd probably die of old age before I reached my goal.

I think at the core you should probably understand the mathematics of each of the major factors and how they all interrelate, but in the end some level of intuition in the analysis will help.

It's kind of like poker.

Many of the best poker players know the math inside and out. Some play almost exclusively by feel/intuition. However, I think most of the super great players use a combination of both.

ClassHandicapper, I so agree with you on many of your postings. You hit the nail on the head with this one. I too am mathematically inclined but find myself limited by time. What isn't made clear here by the supporters of pure math is that the algorithm itself at first is an intuitive guess. It is then tested and if it pans out used in handicapping. And then the problem becomes using the newly discovered formula. In application it never becomes pure math again. A capper needs experience implementing the new formula. The formula that work so well in test now has other considerations to account for when applied. In the end, it isn't as simple as the dice example.

[Capper Al]

Not much there in the win pool. The only play, based on the Kelly Criterion, was a $24 full-kelly on DARK COVE. I ranked them well, but the value just wasn't there.

Is it just me, or does anyone else think the exacta paid well? $32? That seems generous considering the horse to beat came in second.

Regards,

JeremyJet

$32.00 with a favorite for an exacta sounds like a good price.

[Capper Al]

For those that lack sophisticated mathematical skills or believe that the complexities of the game are such that experience and intuition do a better job of handling many of the very unique situations that come up each day, I use a somewhat simple method for determining value.

If you start with the assumption that the public does a pretty good job with the odds/values (which it does), I think if you simply rank the horses (or just contenders) in order of probability of winning and compare that ranking to the ranking on the odds board you have some very useful information.

Any horse that is mis-ranked (assuming you are doing a good job of ranking them) is probably going to be break even or better depending on how badly mis-ranked it is.

If you want you can take it a step further and rank them as 1, 1a, 2, 3 ,4 5, 5a, 5b where 1 and 1a are very similar etc...

The other simple approach is to identify situations that you KNOW the public often misunderstands. For instance, if you KNOW trainer "X" is terrible with layoffs and has one that's getting bet down to 8-5 favoritism, you can simply structure a bet around betting against that horse. If you have a portfolio of situations like that, you'll have a way determining value often enough.

Class, I agree the public is good at setting value. Many times I find myself caught in two opposite directions. If my 8/1 morning line horse was ranked fifth by the trackman and the public has him list at 4/1 and third, I feel I'm viewing some smart money support. Then at other times, I'm trilled when the trackman's third selection at 4/1, and my pick also, is going off as the sixth choice at 12/1. The last scenario is just the opposite of the first. And may I add, that many times when the horse is an overlay the public is correct and the horse doesn't even try. Somehow the public knew something that I missed.

[TrifectaMike]

Class, I agree the public is good at setting value. Many times I find myself caught in two opposite directions. If my 8/1 morning line horse was ranked fifth by the trackman and the public has him list at 4/1 and third, I feel I'm viewing some smart money support. Then at other times, I'm trilled when the trackman's third selection at 4/1, and my pick also, is going off as the sixth choice at 12/1. The last scenario is just the opposite of the first. And may I add, that many times when the horse is an overlay the public is correct and the horse doesn't even try. Somehow the public knew something that I missed.

I disagree. I posted the following about a year ago:

Pareto (Power Law) Distribution
The distribution of the final odds in a horse race shows a power law
P(Odds) is directly proportional to 1/Odds

The interesting aspect of this is that it agrees with empirical data (real life odds). And it fits the data only if you assume that bettors are irrational (bettors do not try to maximize their expected rewards (winnings)). If the assumption is made that bettors are rational (bettors try to maximize their expected rewards), we would obtain a different power law,
P(Odds) is directly proportional to 1/(Odds times Odds),
which is different from empirical data.

I monitor this distribution on a 6 month basis by track and an aggregate of tracks. This distribution has barely changed from year to year.

The public doesn't bet value.

Mike (Dr Baav)

[HUSKER55]

mike, somehow I got lost. what does that number mean exactly?

[TrifectaMike]

mike, somehow I got lost. what does that number mean exactly?

Which number?

[HUSKER55]

(1/(odds*odds))

[classhandicapper]

I disagree. I posted the following about a year ago:

Pareto (Power Law) Distribution
The distribution of the final odds in a horse race shows a power law
P(Odds) is directly proportional to 1/Odds

The interesting aspect of this is that it agrees with empirical data (real life odds). And it fits the data only if you assume that bettors are irrational (bettors do not try to maximize their expected rewards (winnings)). If the assumption is made that bettors are rational (bettors try to maximize their expected rewards), we would obtain a different power law,
P(Odds) is directly proportional to 1/(Odds times Odds),
which is different from empirical data.

I monitor this distribution on a 6 month basis by track and an aggregate of tracks. This distribution has barely changed from year to year.

The public doesn't bet value.

Mike (Dr Baav)

I'm not sure I understand what you are saying, but I am 100% certain that the public is looking for value on the board and attempting to correct any inefficiencies.

To begin with, there are dozens of people on this forum that are doing that, but you can also see it on the board as horses whose odds are out of line are slowly corrected to more reasonable odds as post time approaches.

A great place to observe it is in the PLACE pools. If a solid favorite is way underbet to place with a minute or two to go, a FLOOD of money will come in late and correct it to at least some degree.

You can argue that it's possible to out handicap the smart money that's attempting to take away any inefficiencies, but that's what makes the game so tough. You don't have to out handicap the amateur that is not a good handicapper and that is not value oriented, you have to out handicap the other smart money.

[TrifectaMike]

I'm not sure I understand what you are saying, but I am 100% certain that the public is looking for value on the board and attempting to correct any inefficiencies.

To begin with, there are dozens of people on this forum that are doing that, but you can also see it on the board as horses whose odds are out of line are slowly corrected to more reasonable odds as post time approaches.

A great place to observe it is in the PLACE pools. If a solid favorite is way underbet to place with a minute or two to go, a FLOOD of money will come in late and correct it to at least some degree.

You can argue that it's possible to out handicap the smart money that's attempting to take away any inefficiencies, but that's what makes the game so tough. You don't have to out handicap the amateur that is not a good handicapper and that is not value oriented, you have to out handicap the other smart money.

okeydoky..if you say so.

I don't believe odds correction is the same as betting value.

Mike (Dr Beav)

[overthehill]

no disrespect intended but i wonder if you thought about spending your time investing/speculating as opposed to playing the horses. I cant help but feel that if you can make a living playing the horses you can make a killing in the stock market.

[overthehill]

what you say is true but also frequently not true. I have found that some oddsmakers give lightly raced horses in maiden races short shrift and list them with high odds even though there is a more than trivial likelihood that they will improve in their following race. the public frequently ignore horses that are 20 or 30-1 in the ml that dont deserve to be if you can apply some appropriate probability to their improving off their last race. even trainer statistically significant patterns get tossed out sometimes, and leading trainer/jockey combos end up paying better than 15-1!

[classhandicapper]

I don't believe odds correction is the same as betting value.

Mike (Dr Beav)

It is if the smartest players are specifically betting horses they think are going off at longer odds than their chances justify and avoiding the opposite.

Do a poll on this forum and I'd be willing to bet more than 50% of the sharpest people here play by looking for value on the board. Among winners it's probably closer to 90%.

What you can argue is that not everyone is evaluating value well (because it's so darn tough) or that the masses are not playing that way, but it's the biggest and best players that playing that way and correcting the odds. If that wasn't the case, this game would be comically easy to beat given that so many players are totally clueless.

[thaskalos]

What you can argue is that not everyone is evaluating value well (because it's so darn tough) or that the masses are not playing that way, but it's the biggest and best players that playing that way and correcting the odds. If that wasn't the case, this game would be comically easy to beat given that so many players are totally clueless.
Exactly!

The betting public is often underestimated because so few of the patrons at the racetracks or the OTBs appear "prosperous"...or because the vast majority of them buy the sketchy simulcast programs right at the door -- rather than burn the midnight oil studying the prestigious DRF...

Looks can be very deceiving...

[dansan]

I can figure value after the race no problem

[TrifectaMike]

The distribution of the odds is obtained from race results. Do the odds lie? You could say that my analysis of the data is wrong, but you can't contradict the data by telling me about a handful of people on PA and how do they bet.

Have you ever researched this area? Do you know what the distribution of the odds is? Have you applied some response surface methodology to determine the distribution? Have you correlated the data to a model to determine betting distributions, or are you also just blowing smoke.

Mike (Dr Beav)

P.S. I don't mind disagreements, but at minimum make a logical case.

[thaskalos]

The distribution of the odds is obtained from race results. Do the odds lie? You could say that my analysis of the data is wrong, but you can't contradict the data by telling me about a handful of people on PA and how do they bet.

Have you ever researched this area? Do you know what the distribution of the odds is? Have you applied some response surface methodology to determine the distribution? Have you correlated the data to a model to determine betting distributions, or are you also just blowing smoke.

Mike (Dr Beav)

P.S. I don't mind disagreements, but at minimum make a logical case.
Mike,

Have a little patience with us...we are not used to dealing with the complexities of this game from a purely mathematical viewpoint.

You say that the public doesn't bet value...

Why then is this game so damned hard to beat?

[TrifectaMike]

Mike,

Have a little patience with us...we are not used to dealing with the complexities of this game from a purely mathematical viewpoint.

You say that the public doesn't bet value...

Why then is this game so damned hard to beat?

Only a guess, possibly they don't know how to determine value, so the game for them is extremely hard to beat.


Mike (Dr Beav)

[Capper Al]

I disagree. I posted the following about a year ago:

Pareto (Power Law) Distribution
The distribution of the final odds in a horse race shows a power law
P(Odds) is directly proportional to 1/Odds

The interesting aspect of this is that it agrees with empirical data (real life odds). And it fits the data only if you assume that bettors are irrational (bettors do not try to maximize their expected rewards (winnings)). If the assumption is made that bettors are rational (bettors try to maximize their expected rewards), we would obtain a different power law,
P(Odds) is directly proportional to 1/(Odds times Odds),
which is different from empirical data.

I monitor this distribution on a 6 month basis by track and an aggregate of tracks. This distribution has barely changed from year to year.

The public doesn't bet value.

Mike (Dr Baav)

William Quirin in his book Winning At the Races in the Appendix correlates a horse's winning chances to public odds. The correlation has been examined in several other books since. The bottom line is that public has always done a good job overall. The number of winners always exceeds the number of expected winners for each level of odds. The ROI can't beat the 20% take out or the breakage.

[Capper Al]

no disrespect intended but i wonder if you thought about spending your time investing/speculating as opposed to playing the horses. I cant help but feel that if you can make a living playing the horses you can make a killing in the stock market.

It should be easier to make a buck in the markets with the same effort we put into the horses. The markets have a positive expectancy to grow while the horses have a negative expectancy with the 20% track take out. The only differences is that with the horses you can see the face of the jockey that stiffed you. Can't do the same with the markets.

[Cratos]

The subject of this thread is “Can one really figure value?” Value is a nebulous term in this context because each bettor sees value differently. I do believe that one can figure value, but value to one bettor might be a certain horse at 3-5 in closing odds, but the toteboard has the horse at 6-5 and for that bettor that is exceeded value. For another bettor anything less than 3-1 (that is me) is not value.

But what is being loss or hasn’t been clearly stated is what Mike alluded to with his explanation of Pareto (Power Law) Distribution; that each wager doesn’t contribute to the same amount of value.

Another point is that over the long haul you will want your probability of handicapping a winner to be better than the closing odds of that winner. What is being said is that if you can handicap winners at a 40% average rate (odds 1.50-1) from your historical wagering population you will want your average closing winning odds to be X where X>1.50-1. Your objective is to see what the betting public does not see and exploit them.

This is easier said than done when historically the betting public selects about 1/3 of races correctly and if you are a non-favorite bettor you are now selecting in the aggregate from 67% of the horses that run because approximate 33% of the population that run win as favorites.

[thaskalos]

The subject of this thread is “Can one really figure value?” Value is a nebulous term in this context because each bettor sees value differently. I do believe that one can figure value, but value to one bettor might be a certain horse at 3-5 in closing odds, but the toteboard has the horse at 6-5 and for that bettor that is exceeded value. For another bettor anything less than 3-1 (that is me) is not value.

But what is being loss or hasn’t been clearly stated is what Mike alluded to with his explanation of Pareto (Power Law) Distribution; that each wager doesn’t contribute to the same amount of value.

Another point is that over the long haul you will want your probability of handicapping a winner to be better than the closing odds of that winner. What is being said is that if you can handicap winners at a 40% average rate (odds 1.50-1) from your historical wagering population you will want your average closing winning odds to be X where X>1.50-1. Your objective is to see what the betting public does not see and exploit them.

This is easier said than done when historically the betting public selects about 1/3 of races correctly and if you are a non-favorite bettor you are now selecting in the aggregate from 67% of the horses that run because approximate 33% of the population that run win as favorites.
Here is the argument as I see it Cratos:

When it comes to accurately assessing the value present in a race...the winning players come in three types, IMO.

1.) THOSE USING A MATHEMATICAL APPROACH. These players obviously have some sort of mathematical background, and they have relied upon these math skills to create a process through which they can accurately - and precisely - determine the value in a race...and the best way of profiting from it.

2.) THE LARGELY INTUITIVE PLAYERS. These players also use sophisticated methods in determining the value offered by a given race...but they lack the math skills necessary to bring their handicapping opinions to the level of precision achieved by the first group of players.

These intuitive players, however, operate under the belief that their vast experience in the game goes a long way toward negating their mathematical deficiency. Eventhough they cannot precisely calculate the exact value offered by a given horse...they are confident that they will make the correct play anyway - because of the handicapping knowledge and experience they have accumulated playing this game over the span of many years.

3.) THOSE PROFICIENT IN BOTH THESE APPROACHES. These players are rare, IMO...and they are probably the game's biggest winners.


The argument rages between the first two groups of players, because they don't fully understand each other's approach...and they cannot believe that both these approaches can do an adequate job.

IMO...both the mathematical AND the intuitive players can have success playing this game...assuming that they have the internal fortitude necessary to deal with the game's wild swings.

Given a choice...I too would choose a more mathematical approach, but - as they say - "you can't teach an old dog new tricks" :)

[Dave Schwartz]

Thasky, that was a great post. I think you hit it really well.

The odd part is that there seems to be a natural tendency among losing players of one group to disrespect the players and methodologies of the other groups. I emphasize "losing players" because in my experience a true winning player usually has no need to convince other players that his way is best.


Most of you on this forum probably know Formula_2000. When you see his posts over the recent past he talks about his selections with an expectation that they will produce a profit.

Those of you who have been on this forum for awhile may recall when he constantly beat many of us up with why the game couldn't possibly be beaten. Nobody could sway him from that belief.

Then all of a sudden, he made some personal discoveries and his belief structure seemed completely changed.

His posts seem to indicate (to me) that he still believes his way is best, but he seems much more open-minded to other peoples' ideas now.


My point is that so many of us are trapped in what we know to be true. Amazingly, much of that, in the grand scheme of things, will turn out to be untrue, or at the least, not as true as we thought.

If we are fortunate we will only have to have this lesson presented to us a few times in order to learn it.


Regards,
Dave Schwartz

[thaskalos]

My point is that so many of us are trapped in what we know to be true. Amazingly, much of that, in the grand scheme of things, will turn out to be untrue, or at the least, not as true as we thought.

If we are fortunate we will only have to have this lesson presented to us a few times in order to learn it.


Regards,
Dave Schwartz
I could not have said it better myself! :ThmbUp:

[showbet]

I'm not sure I understand what you are saying, but I am 100% certain that the public is looking for value on the board and attempting to correct any inefficiencies.

To begin with, there are dozens of people on this forum that are doing that, but you can also see it on the board as horses whose odds are out of line are slowly corrected to more reasonable odds as post time approaches.

A great place to observe it is in the PLACE pools. If a solid favorite is way underbet to place with a minute or two to go, a FLOOD of money will come in late and correct it to at least some degree.
How well I know this to be true. I determine all my wagers based on the tote. I've basically given up trying to find any place/show pool "inefficiencies" at any track with a decent handle, which is to say all thoroughbred tracks.

It's not unusual to see a 4/5 favorite (which is ~45% of the win pool) have only 20% of the place pool, or 15% of the show pool, at the second to last flash of the tote, then end up with 35% of the place pool and 30% of the show pool at the off.

About the only time I can find a pool "inefficiency" at a thoroughbred track is when there is a "bridgejumper" horse; ie, a horse whose share of the place or show pool far exceeds his share of the win pool.

[Overlay]

Your objective is to see what the betting public does not see and exploit them.

This is easier said than done when historically the betting public selects about 1/3 of races correctly and if you are a non-favorite bettor you are now selecting in the aggregate from 67% of the horses that run because approximate 33% of the population that run win as favorites.
However, even favorites can still be overlays in the win pool, so it's not necessary for a value player to disregard or discard them out of hand as a group, with no consideration at all.

[Capper Al]

Determining value is difficult. Both math and intuition help, but neither is an exact science. The fun is in the uncertainty.

[Cratos]

However, even favorites can still be overlays in the win pool, so it's not necessary for a value player to disregard or discard them out of hand as a group, with no consideration at all.

I thought that was inferred with the example between the 3-5 odds and the 6-5 odds, but in the long run we all wager along our own risk curve.

[classhandicapper]

Here is the argument as I see it Cratos:

When it comes to accurately assessing the value present in a race...the winning players come in three types, IMO.

1.) THOSE USING A MATHEMATICAL APPROACH. These players obviously have some sort of mathematical background, and they have relied upon these math skills to create a process through which they can accurately - and precisely - determine the value in a race...and the best way of profiting from it.

2.) THE LARGELY INTUITIVE PLAYERS. These players also use sophisticated methods in determining the value offered by a given race...but they lack the math skills necessary to bring their handicapping opinions to the level of precision achieved by the first group of players.

These intuitive players, however, operate under the belief that their vast experience in the game goes a long way toward negating their mathematical deficiency. Eventhough they cannot precisely calculate the exact value offered by a given horse...they are confident that they will make the correct play anyway - because of the handicapping knowledge and experience they have accumulated playing this game over the span of many years.

3.) THOSE PROFICIENT IN BOTH THESE APPROACHES. These players are rare, IMO...and they are probably the game's biggest winners.


The argument rages between the first two groups of players, because they don't fully understand each other's approach...and they cannot believe that both these approaches can do an adequate job.

IMO...both the mathematical AND the intuitive players can have success playing this game...assuming that they have the internal fortitude necessary to deal with the game's wild swings.

Given a choice...I too would choose a more mathematical approach, but - as they say - "you can't teach an old dog new tricks" :)

I'm not sure this is exactly correct.

You make it sound like the mathematical players can value the horses with perfect precision and the intuitive players cannot.

That's simply not true.

If the mathematical formulas being used are not all encompassing and perfectly weighted, the results will not be great.

If a player has a huge amount of experience and knowledge, his ability to calculate the fair odds intuitively might produce very precise results.

That's where the real debate is.

The math guys are trying to cover as many factors and issues that come up as possible and weighing them as well as possible. The intuitive players are saying that there is so much that is unique about so many races and therefore not covered by the math, the formulas will fall short of calculating a good line in a lot of instances even though IN THEORY it is possible.

On the flip side the intuitive guys can feel their way through hundreds of thousands of possibilities that the math guys may not have coded for, but that does not mean all of them are doing it well in all instances.

That's why I agree that a combination is best.

The math is good for things with large samples that can be studied in depth. It's not as good for those rarer combinations of things that only come up from time to time.

[Capper Al]

Amen ClassHandicapper. Tell it like it is!

[Nitro]

Originally Posted by Nitro & Original Reply by TrifectaMike
Yes the game we all seem to enjoy is about assimilating information, but not necessarily from the perspective of Academia. What does this mean? It simply means that based on many years of experience in this great game and having met players of all descriptions and backgrounds, that I believe that you don’t have to rely on a formal education to comprehend the information that’s available in order to succeed.

Horses are not mechanical machines! I don't believe anyone has implied they are machines. And I believe that anyone who assumes that Math (at any level) can solve the relationship between an intricate apparatus and some predictable performance must be referring to some sort of machinery. Please advise us when a formula is developed that can evaluate and express the health, well-being, physical condition and competitive instincts of a living race horse.

Anyone that believes that an “exact” science like Math can magically conjure up consistent value based decisions when applied to inexact, ever changing, completely subjective GAME like horse racing is in for some big surprises when it comes to actually tallying up the net gain. I agree. It is a subjective game. And I approach racing exactly from that point of view. If this statement is true it might be interesting to hear how the obvious conflict between the subjective nature of the beast is handled by an objective science like Math.

This simple but true premise exists because all of the intangibles related to the effects of handling, caring for and managing the animal are absolutely unquantifiable. True, but there exist evidence to guide you. Perhaps you can expand on this comment and let us all in on where this evidence is and how we can properly evaluate it for our benifit.

Of course those in the Math corner will conveniently ignore these ambiguities by implying that their figures, statistical models and probability theories (that are based solely on the past) Bayesian probability models are based on evidence...not the past. can somehow provide the answers for determining the future. I have real problem with this rationale! Are you trying to tell us that “evidence” by its very definition is not based on something factual that already exists? I think common sense says otherwise!

Maybe It’s possible with only part the picture to achieve a modest positive ROI. And you know this how? A very simple observation really: If value is the real objective then why haven’t the mathematical wizards discovered the means to take advantage of having truly found it?

But I thought this discussion was about “Value”. It is all about value and value can be quantified. Yes “value” as a numerical component can be quantified, but trying to correlate the ambiguities in this game to some known entity is I believe a bit far-reaching.

How do the mathematicians rationalize their betting strategies when they’re all arriving at the same conclusions for playing? Or are they? Maybe these exacting results are open for interpretation? There are no exacting results. Exactly my point!

Possibly like all of the other information that’s available to those who might enjoy the game without a Harvard degree. Enjoy the game. I have not implied that any degree(s) are required. There are certainly no degrees required to enjoy this game, but there might a degree required to comprehend some of the Math that’s being thrown around as if its second nature to those viewing.

I find it amusing that you introduce your professional actuary to lecture on how to employ the 'proper' techniques as used by the insurance companies...group this and group that...run it like an insurance company...otherwise you are doomed to failure. It is laughable.

When did understanding math become such a bad thing?

Mike (Dr Beav) I apologize if you felt that my friend’s brief description was a lecture of some sort. I believe you missed the point of the entire correlation of weighted variables between two totally different disciplines both looking at potential outcomes. But that’s neither here nor there.

What’s really laughable is the mindset that some people have when trying to apply Mathematics to things like horseracing. The truth is that a numerical approach to handicapping Thoroughbreds only works up to a point. The horses have their own agenda when running and this agenda coincides only periodically with what we believe to be fact (evidence). This causes many distortions which math of any description can't explain. If you believe that the numbers, statistical models and probability theories will someday overcome or better yet explain nature, then I've got some land on the Moon you might be interested in.

[PaceAdvantage]

I will never, ever, understand the mindset of someone like Nitro, who insists on poo-pooing the thoughts of those who have embraced a mathematical solution to this game. And usually, the only defense you have is "they're animals...flesh and blood...you can't possibly incorporate all the intangibles into your equations."

News flash: HE DOESN'T HAVE TO IN ORDER TO PROFIT HANDSOMELY

[jasperson]

What’s really laughable is the mindset that some people have when trying to apply Mathematics to things like horseracing. The truth is that a numerical approach to handicapping Thoroughbreds only works up to a point. The horses have their own agenda when running and this agenda coincides only periodically with what we believe to be fact (evidence). This causes many distortions which math of any description can't explain. If you believe that the numbers, statistical models and probability theories will someday overcome or better yet explain nature, then I've got some land on the Moon you might be interested in.Then how are we suppose to handicap the races, with ESP or astrology? Right now my ESP is flat and my horoscope is bad.

[thaskalos]

I'm not sure this is exactly correct.

You make it sound like the mathematical players can value the horses with perfect precision and the intuitive players cannot.

That's simply not true.

If the mathematical formulas being used are not all encompassing and perfectly weighted, the results will not be great.

If a player has a huge amount of experience and knowledge, his ability to calculate the fair odds intuitively might produce very precise results.

That's where the real debate is.

The math guys are trying to cover as many factors and issues that come up as possible and weighing them as well as possible. The intuitive players are saying that there is so much that is unique about so many races and therefore not covered by the math, the formulas will fall short of calculating a good line in a lot of instances even though IN THEORY it is possible.

On the flip side the intuitive guys can feel their way through hundreds of thousands of possibilities that the math guys may not have coded for, but that does not mean all of them are doing it well in all instances.

That's why I agree that a combination is best.

The math is good for things with large samples that can be studied in depth. It's not as good for those rarer combinations of things that only come up from time to time.

NOTHING can be taken at face value in this game Classhandicapper...and you KNOW that.

Whether we are dealing with speed and pace figures, class distinctions or value determination...nothing is certain in handicapping, and everything is open to interpretation.

Neither the math nor the intuitive guys really think that a perfectly precise assessment of value is really possible (not in a game with as many unknown variables as horse racing)...but it behooves them to be as precise as they can be in their work nonetheless.

We are not trying to assess value PERFECTLY; we are just trying to be a little better at it than the crowd is...in enough races to make our effort worthwhile.

The sports bettor knows that he is unable to predict the exact score of a sporting event also...but that doesn't stop him from trying to do it as precisely as he can.

[thaskalos]

What’s really laughable is the mindset that some people have when trying to apply Mathematics to things like horseracing. The truth is that a numerical approach to handicapping Thoroughbreds only works up to a point. The horses have their own agenda when running and this agenda coincides only periodically with what we believe to be fact (evidence). This causes many distortions which math of any description can't explain. If you believe that the numbers, statistical models and probability theories will someday overcome or better yet explain nature, then I've got some land on the Moon you might be interested in.
Is is still "laughable", if these math guys are showing a good profit doing their thing?

[PaceAdvantage]

Thaskalos, you misread jasperson because he had his quote all messed up, which I have since fixed.

[Capper Al]

All this from a simple correlation to dice to clarify the effects of deviation. Who would of thought it possible? Seriously, I never knew that this thread was going to end up here. Don't get me wrong. I'm OK with it, but just surprised.

[Cratos]

What’s really laughable is the mindset that some people have when trying to apply Mathematics to things like horseracing. The truth is that a numerical approach to handicapping Thoroughbreds only works up to a point. The horses have their own agenda when running and this agenda coincides only periodically with what we believe to be fact (evidence). This causes many distortions which math of any description can't explain. If you believe that the numbers, statistical models and probability theories will someday overcome or better yet explain nature, then I've got some land on the Moon you might be interested in.

After reading your entire post it was the last paragraph that got to me and in particular when you say: “If you believe that the numbers, statistical models and probability theories will someday overcome or better yet explain nature, then I've got some land on the Moon you might be interested in.” That statement clearly tells me that you have a lack of understanding of mathematical or statistical modeling.

Modeling is about the ASSUMPTIONS you make and the mathematical or statistical formulas you use. Horseracing is not much different from any other dynamic endeavor when it comes to a predictive model. Physics is the “mother” science in these endeavors and math and statistics are the tools which are used to tighten the screws, but if your assumptions are incorrect you will strip the threads to use a mechanical analogy.

In horseracing we have the variables: velocity (it is speed when not directional), acceleration, distance, and load (weight). The trick is what assumptions you make about those variables with respect to environmental conditions, breeding, track surface, and track geometry.

Yes, there are the non-parametric variables like class and style that can cause the model to go awry. Also there is the variability caused by human intervention (e.g. trainer, owner, and jockey), but people like Mike with his knowledge of Bayesian statistics would be very helpful in weighing those variables.

In summary a carefully designed model with good ASSUMPTIONS is the best predictor of future events no matter if it is horseracing, the weather, the stock market, etc.

[classhandicapper]

In summary a carefully designed model with good ASSUMPTIONS is the best predictor of future events no matter if it is horseracing, the weather, the stock market, etc.

Assuming you are willing to code for a few hundred thousand combinations of conditions. ;)

[TrifectaMike]

I'm not sure I understand what you are saying, but I am 100% certain that the public is looking for value on the board and attempting to correct any inefficiencies market efficiency and rationality of the bettors is not the same thing.

To begin with, there are dozens of people on this forum that are doing that, but you can also see it on the board as horses whose odds are out of line are slowly corrected to more reasonable odds as post time approaches.

A great place to observe it is in the PLACE pools. If a solid favorite is way underbet to place with a minute or two to go, a FLOOD of money will come in late and correct it to at least some degree.

You can argue that it's possible to out handicap the smart money that's attempting to take away any inefficiencies, but that's what makes the game so tough. You don't have to out handicap the amateur that is not a good handicapper and that is not value oriented, you have to out handicap the other smart money.

What I am saying is that based on odds data, analysis of the data shows that horse players (obviously not all) are still betting on horses that have maximum probability of winning instead of betting on horses whose expected rewards are maximum (winning probability multiplied by odds). If players bet in a manner to maximize their rewards, it would have to be equivalent for every horse in every race that has been decided.

Mike (Dr Beav)

[Capper Al]

Let's bring this thread back to it's topic, "Can One Really Figure Value?". We've seemed to argue here intuition vs math with some siding with both. We are dealing with fuzzy logic at best. We can make our guesses with whatever method we choose. Our guesses, with validating experience, should be in the ballpark.

Barry Meadows in his book "Money Secrets at the Racetrack" suggest we figure if today's race was run a 100 times how many wins would you guess a particular horse will have. (Sounds intuitive to me, but not necessarily.) From our projections we set our morning line. Then Barry adds a 50% markup to cover error in our estimates plus be able to make a profit. This would work if the horse's value was certain. If odds vary from expected either higher or lower, we more than likely have other information to deal with. The Public, one of the best known odds estimator, is communicating there's something wrong if the odds are higher, or we're on to something if the odds are lower. We can't win by not playing our hot horse because the odds are low, or by throwing our money away on a horse with a problem that's not visible to us. It might come down to what chalk players say -- value is in cashing tickets only.

[aaron]

Let's bring this thread back to it's topic, "Can One Really Figure Value?". We've seemed to argue here intuition vs math with some siding with both. We are dealing with fuzzy logic at best. We can make our guesses with whatever method we choose. Our guesses, with validating experience, should be in the ballpark.

Barry Meadows in his book "Money Secrets at the Racetrack" suggest we figure if today's race was run a 100 times how many wins would you guess a particular horse will have. (Sounds intuitive to me, but not necessarily.) From our projections we set our morning line. Then Barry adds a 50% markup to cover error in our estimates plus be able to make a profit. This would work if the horse's value was certain. If odds vary from expected either higher or lower, we more than likely have other information to deal with. The Public, one of the best known odds estimator, is communicating there's something wrong if the odds are higher, or we're on to something if the odds are lower. We can't win by not playing our hot horse because the odds are low, or by throwing our money away on a horse with a problem that's not visible to us. It might come down to what chalk players say -- value is in cashing tickets only.
In my opinion,the really good players do not give away winners because of price.The best players pick enough big price horses,so that they don't have to throw out their short price horses because of lack of value.If you are a player whose best picks are only going to be in the even money,two to 1 range,you have no shot. If the only time you can get value is by going with your 4th or 5th choice you are in trouble.

[Cratos]

Assuming you are willing to code for a few hundred thousand combinations of conditions. ;)

It is clear that you don’t understand the function of the modeling assumption. The modeling assumption(s) can be written out in paragraph form or sketched out schematically. The software coding can or cannot come later; it depends on whether you want the model to be a computerized model or not. It is true that in the case of horserace handicapping a computer model would be the best, but if you understand logical constructs as they relate to horseracing the model will not take “a few hundred thousand combinations of conditions” for coding.

[Capper Al]

In my opinion,the really good players do not give away winners because of price.The best players pick enough big price horses,so that they don't have to throw out their short price horses because of lack of value.If you are a player whose best picks are only going to be in the even money,two to 1 range,you have no shot. If the only time you can get value is by going with your 4th or 5th choice you are in trouble.

This is my understanding also. Ainslie once suggested a win only strategy with one exception. If the odds were 4/1 or more then buy a place ticket. At first I thought this was a simple conservative method that he just recommended for his book. Now after playing with various exotic wagers methods, Ainslie strategy is making sense.

[aaron]

This is my understanding also. Ainslie once suggested a win only strategy with one exception. If the odds were 4/1 or more then buy a place ticket. At first I thought this was a simple conservative method that he just recommended for his book. Now after playing with various exotic wagers methods, Ainslie strategy is making sense.
I also feel there is nothing wrong with w/p bets. As a matter of fact long shots tend to run 2nd twice as much as they do to win. Betting w/p can help avoid some long losing streaks

[classhandicapper]

It is clear that you don’t understand the function of the modeling assumption. The modeling assumption(s) can be written out in paragraph form or sketched out schematically. The software coding can or cannot come later; it depends on whether you want the model to be a computerized model or not. It is true that in the case of horserace handicapping a computer model would be the best, but if you understand logical constructs as they relate to horseracing the model will not take “a few hundred thousand combinations of conditions” for coding.

I understand, but have to disagree.

1. I often have to weigh a dozen issues for each horse in order to come to some reasonable conclusion about how good it actually is, how suited it is to the conditions, and how well it should run today etc...

2. For each horse in the typical 8 horse field, those dozen issues are not even all the same ones. There are probably "at least" 100 things I consider from time to time in various combination.

3. For each of these races, I am going to weigh things differently depending whether it's a sprint, route, marathon, on turf, dirt, synthetic, or mud, there's an inside bias, outside bias, speed bias, closer bias, combination of biases, a lot of early speed, lack of early speed, various combinations of the above.

It goes on and on and on and on.

That doesn't even include all the subjective judgments I have to make about biases and conditions today, about previous biases, whether various trainer stats are short term random occurrences or indications of trainer intent, strength, weakness, sickness in barn, new drug use etc..

To me, the combinations seem totally endless. That's what makes it both a great and difficult game.

From a theoretical point of view it's clearly possible to either model or model and code for everything that comes up, but from a practical point of view it seems like a gargantuan task that will ultimately miss some of the nuances of how various combinations interplay and how unique most races actually are.

I'm not much of a chess player, but it seems similar. It took decades and whole teams of brilliant programmers and chess players using the fastest computers on earth to finally beat Karpov's combination of knowledge and natural ability. And he argues to this day that the only reason he lost was that there was also human intervention from other chess experts.

Now of course, the flip side is that weighing all these things intuitively is fraught with peril. Almost without question a good model would be vastly superior in the hands of a beginner or even moderate player than intuitive/feel handicapping just like even a cheap hand held computer game can beat me at chess. :lol:

But I think it's going to take a massive effort and great technology to model and code in a way that beats the very best knowledgeable intuitive/feel handicappers in the world. (to be clear, I'm not saying I am one of them)

[classhandicapper]

What I am saying is that based on odds data, analysis of the data shows that horse players (obviously not all) are still betting on horses that have maximum probability of winning instead of betting on horses whose expected rewards are maximum (winning probability multiplied by odds). If players bet in a manner to maximize their rewards, it would have to be equivalent for every horse in every race that has been decided.

Mike (Dr Beav)

If you are saying "most players" I agree with you, but in terms of money, IMO there is a concentration of "large sums" that comes in from "some players" towards horses with perceived good value. That's what makes it so hard to find good value despite the large number of players that are either totally clueless or just playing the most probable winners no matter how low the odds.

[Capper Al]

It is clear that you don’t understand the function of the modeling assumption. The modeling assumption(s) can be written out in paragraph form or sketched out schematically. The software coding can or cannot come later; it depends on whether you want the model to be a computerized model or not. It is true that in the case of horse race handicapping a computer model would be the best, but if you understand logical constructs as they relate to horse racing the model will not take “a few hundred thousand combinations of conditions” for coding.

Hey Cratos, ClassHandicapper is on the money with this one. I'm working on a C++ application myself. Narrowing on the bare essentials speed, class, workouts, days since lasted raced, form, trainer, jockey and so it goes. But what goes into the application is a finite set of variables. I will test the good to fit for my variables and that will tell me each variable's value which I can determine odds. These variables are at the core of the game, but every time a race is put together there is a question about these factors. For example, all of a sudden a good speed figure pops up. Was this a fluke? Was the daily variant over estimated. Will the horse bounce? And so on.

[Cratos]

Hey Cratos, ClassHandicapper is on the money with this one. I'm working on a C++ application myself. Narrowing on the bare essentials speed, class, workouts, days since lasted raced, form, trainer, jockey and so it goes. But what goes into the application is a finite set of variables. I will test the good to fit for my variables and that will tell me each variable's value which I can determine odds. These variables are at the core of the game, but every time a race is put together there is a question about these factors. For example, all of a sudden a good speed figure pops up. Was this a fluke? Was the daily variant over estimated. Will the horse bounce? And so on.

I still disagree, but to each his own. Assumptions on a model are individual and again I say it will not take a few hundred thousand lines of code. I have a model which is computerized and is based on the principles of physics as they relate to horseracing along with my knowledge of horseracing gained over 40 years and this model is very predictive.

[classhandicapper]

Hey Cratos, ClassHandicapper is on the money with this one. I'm working on a C++ application myself. Narrowing on the bare essentials speed, class, workouts, days since lasted raced, form, trainer, jockey and so it goes. But what goes into the application is a finite set of variables. I will test the good to fit for my variables and that will tell me each variable's value which I can determine odds. These variables are at the core of the game, but every time a race is put together there is a question about these factors. For example, all of a sudden a good speed figure pops up. Was this a fluke? Was the daily variant over estimated. Will the horse bounce? And so on.

Exactly.

Assume I have two horses.

One earned a speed of 95 loose on the lead and the other earned an 87 with a more average trip. That means one thing if studied on a aggregate basis for a large sample.

What if the 95 was earned on a speed favoring track?

What is the 95 was earned on a closer favoring track?

What if he was loose but it was a very fast pace?

What if he was loose, but it was a very hot pace, and it was a speed favoring track?

What if the horse never ran a speed figure higher than a 75 before?

What if the horse never ran a speed figure higher than a 75 before but was recently claimed by a move up trainer?

What if the trainer has historically never been a move up guy, but over the last 3 weeks a few have moved up?

What if the horse never ran a speed figure higher than a 75, was loose on the lead to earn that 95, it was a speed favoring track, but it was very hot pace, the horse was recently claimed by a move up trainer, but he's stretching an extra 2 furlongs, he's clearly not really bred to go that far and tired a little the one time he tried, but his new trainer is especially good at stretching horses out etc...

It's endless and this is just one horse.

Each of these are factors that can be statistically studied and weighed at a high level, but there are endless combinations that are very specific where the weights could be different and involve some very subjective judgments.

I'm all for what you, Cratos, and others are doing with models, statistics, computers etc... I wish I had the time and inclination to do more of that for myself or had the opportunity to review a lot of the work of others (like you guys), but in the end I can't imagine myself allowing a model to more or less dictate odds in a mechanical fashion.

[Cratos]

I understand, but have to disagree.

1. I often have to weigh a dozen issues for each horse in order to come to some reasonable conclusion about how good it actually is, how suited it is to the conditions, and how well it should run today etc...

2. For each horse in the typical 8 horse field, those dozen issues are not even all the same ones. There are probably "at least" 100 things I consider from time to time in various combination.

3. For each of these races, I am going to weigh things differently depending whether it's a sprint, route, marathon, on turf, dirt, synthetic, or mud, there's an inside bias, outside bias, speed bias, closer bias, combination of biases, a lot of early speed, lack of early speed, various combinations of the above.

It goes on and on and on and on.

That doesn't even include all the subjective judgments I have to make about biases and conditions today, about previous biases, whether various trainer stats are short term random occurrences or indications of trainer intent, strength, weakness, sickness in barn, new drug use etc..

To me, the combinations seem totally endless. That's what makes it both a great and difficult game.

From a theoretical point of view it's clearly possible to either model or model and code for everything that comes up, but from a practical point of view it seems like a gargantuan task that will ultimately miss some of the nuances of how various combinations interplay and how unique most races actually are.

I'm not much of a chess player, but it seems similar. It took decades and whole teams of brilliant programmers and chess players using the fastest computers on earth to finally beat Karpov's combination of knowledge and natural ability. And he argues to this day that the only reason he lost was that there was also human intervention from other chess experts.

Now of course, the flip side is that weighing all these things intuitively is fraught with peril. Almost without question a good model would be vastly superior in the hands of a beginner or even moderate player than intuitive/feel handicapping just like even a cheap hand held computer game can beat me at chess. :lol:

But I think it's going to take a massive effort and great technology to model and code in a way that beats the very best knowledgeable intuitive/feel handicappers in the world. (to be clear, I'm not saying I am one of them)

Class,modeling is about logical constructs and in my model I built it on Factor assumptions which are negative in the model and Angle assumptions which are positive in the model with the intent to answer the simple question: “What factors make a horse lose and what angles make a horse win?” The factors are finite and the angles are infinite.

[Cratos]

Exactly.

Assume I have two horses.

One earned a speed of 95 loose on the lead and the other earned an 87 with a more average trip. That means one thing if studied on a aggregate basis for a large sample.

What if the 95 was earned on a speed favoring track?

What is the 95 was earned on a closer favoring track?

What if he was loose but it was a very fast pace?

What if he was loose, but it was a very hot pace, and it was a speed favoring track?

What if the horse never ran a speed figure higher than a 75 before?

What if the horse never ran a speed figure higher than a 75 before but was recently claimed by a move up trainer?

What if the trainer has historically never been a move up guy, but over the last 3 weeks a few have moved up?

What if the horse never ran a speed figure higher than a 75, was loose on the lead to earn that 95, it was a speed favoring track, but it was very hot pace, the horse was recently claimed by a move up trainer, but he's stretching an extra 2 furlongs, he's clearly not really bred to go that far and tired a little the one time he tried, but his new trainer is especially good at stretching horses out etc...

It's endless and this is just one horse.

Each of these are factors that can be statistically studied and weighed at a high level, but there are endless combinations that are very specific where the weights could be different and involve some very subjective judgments.

I'm all for what you, Cratos, and others are doing with models, statistics, computers etc... I wish I had the time and inclination to do more of that for myself or had the opportunity to review a lot of the work of others (like you guys), but in the end I can't imagine myself allowing a model to more or less dictate odds in a mechanical fashion.


One difference between the two of us is I don’t use or consider speed figures at all and another difference is I am not “allowing a model to more or less dictate odds in a mechanical fashion.” My model dictates odds in a logical fashion

[classhandicapper]

Class,modeling is about logical constructs and in my model I built it on Factor assumptions which are negative in the model and Angle assumptions which are positive in the model with the intent to answer the simple question: “What factors make a horse lose and what angles make a horse win?” The factors are finite and the angles are infinite.

I understand what you are saying, but I can't see how a model that doesn't ask all the relevant questions can determine the values all the time.

To be super clear, I'm not saying it can't be profitable! Like a said, your model and others may be dynamite in the hands of a lot of people. It might even make me a lot better.

I am saying there are going to be instances where you "think" you have value or where you "think" the favorite is very over bet but your the model is missing some very relevant information. That causes problems for me, but it may be a personality issue on my part.

I used a very simple model to play basketball profitably last year.

I use a fairly simple method to find value in the stock market when I make investments (since 1986).

It's only horse racing that makes me a nut job. :rolleyes:

[thaskalos]

It's only horse racing that makes me a nut job. :rolleyes:
I agree CH...

Our game is very peculiar...and the assessment of "value" in this game becomes much trickier than in other gambling games...because of the fact that there is so much pertinent information out there that we are not aware of.

A sports bettor does his calculations and decides that he likes the Packers to defeat the Steelers by 4 points...and, to his surprise, discovers that the oddsmakers have declared the STEELERS as 3 point favorites.

What is this bettor's reaction?

He is THRILLED!

What he thought might be a solid play now has become an OUTSTANDING wager...and he contemplates increasing his bet.

Compare that to the horseplayer who has decided that the 8 horse in Philly Park's 8th race tomorrow should be a 2-1 favorite...and hopes to get 3-1 odds on the horse in order to justify making a win bet on it.

He goes to the OTB tomorrow and, to his surprise, discovers that the 8 horse is offered at 8-1 odds with 2 minutes to post...while a nondescript recent claim by an unknown trainer is being "hammered" down to 2-1 favoritism...from a 15-1 morning line!

What is our horseplayer's reaction?

He is paralyzed by fear and paranoia!

He knows that his "8" horse "looks too good" to be offered at 8-1 odds...and he also knows that no "dumb money" alone could make this mysterious recent claim the 2-1 favorite. So what does our "hero" do?

Instead of increasing his bet on this wonderful "overlay" he has uncovered (like his sports-betting counterpart wants to do)...our "terrified" horseplayer now decides to PASS the race altogether...lest he end up looking like a fool...

It's hard to determine "true" value in this game...:)

[Native Texan III]

I understand, but have to disagree.

1. I often have to weigh a dozen issues for each horse in order to come to some reasonable conclusion about how good it actually is, how suited it is to the conditions, and how well it should run today etc...

2. For each horse in the typical 8 horse field, those dozen issues are not even all the same ones. There are probably "at least" 100 things I consider from time to time in various combination.

3. For each of these races, I am going to weigh things differently depending whether it's a sprint, route, marathon, on turf, dirt, synthetic, or mud, there's an inside bias, outside bias, speed bias, closer bias, combination of biases, a lot of early speed, lack of early speed, various combinations of the above.

It goes on and on and on and on.

That doesn't even include all the subjective judgments I have to make about biases and conditions today, about previous biases, whether various trainer stats are short term random occurrences or indications of trainer intent, strength, weakness, sickness in barn, new drug use etc..

To me, the combinations seem totally endless. That's what makes it both a great and difficult game.

From a theoretical point of view it's clearly possible to either model or model and code for everything that comes up, but from a practical point of view it seems like a gargantuan task that will ultimately miss some of the nuances of how various combinations interplay and how unique most races actually are.

I'm not much of a chess player, but it seems similar. It took decades and whole teams of brilliant programmers and chess players using the fastest computers on earth to finally beat Karpov's combination of knowledge and natural ability. And he argues to this day that the only reason he lost was that there was also human intervention from other chess experts.

Now of course, the flip side is that weighing all these things intuitively is fraught with peril. Almost without question a good model would be vastly superior in the hands of a beginner or even moderate player than intuitive/feel handicapping just like even a cheap hand held computer game can beat me at chess. :lol:

But I think it's going to take a massive effort and great technology to model and code in a way that beats the very best knowledgeable intuitive/feel handicappers in the world. (to be clear, I'm not saying I am one of them)

The points you have raise are true if you are attempting to determine the exact speed rating the horse will achieve in the upcoming race. But for value what you are looking for is the long term (not the next race and the one after). The only question to answer is does MY model identify horses that are likely to be value and do those identified horses show a long term profit. If they do, YOUR model is fine, if not back to the drawing board. You are identifying long term value, not the winner of every race - not the "true odds" of every horse in a particular race - this is the vital point that the "value" books omit.

What you might consider as more achievable and practical is to say that with all things considered I am sure that the horse can achieve a rating in the range 70-95, but with the most likely value around 80-85. More like 85 if it is a speed favouring track for example, 80 if not. If you think in terms of likely range of performance you can compare one horse against another with a different likely range. The range is readily identifiable but the precise future single number figure is not. The more "known unknowns" the wider the range.

If you could identify the precise future race speed figure of each and every horse then the top rated horse would win every race and the betting market would collapse.

[Capper Al]

Have you ever made up a system for a particular race type, say claiming sprints on dirt, thinking you had every combination covered. Tested it on paper or against a database then went to actually play it and find new combinations that you never thought of? I have several times.

I'm hearing what you are saying about ranges and modeling but wouldn't you have to narrow down the factors to fit so that your model can go unchallenged?

[Native Texan III]

Have you ever made up a system for a particular race type, say claiming sprints on dirt, thinking you had every combination covered. Tested it on paper or against a database then went to actually play it and find new combinations that you never thought of? I have several times.

I'm hearing what you are saying about ranges and modelling but wouldn't you have to narrow down the factors to fit so that your model can go unchallenged?

No, systems only work if you have one or two particular angles that the market does not fully account for. Systems are long term - they do not depend on a single race but the overall profit/loss result from a long series of races. Systems do not even consider factors they do not account for - by definition. Looking for the unknowable issue in a single race is by definition not ever going to work out. Just allow for it in the price.

Thinking in terms of ranges applies to any betting or method under uncertainty. The price you want to take depends on the amount of uncertainty YOU have about the selection/s. That uncertainty is found by analysing how YOUR method works or does not work over the long term. You know then what extra you need in a price to cover the "unknown unknowns".

Long term you do not even have to think about the factors that just crop up in the occasional single race - just be "righter" on average more often, long term, about the factors that consistently crop up in every race. That consistency is often easiest to obtain from a model (computer or hand driven) - which does not have emotions nor hunches. Changing your model because something else has cropped up in a particular race leads inevitably to inconsistency and a failed model or method.

[aaron]

No, systems only work if you have one or two particular angles that the market does not fully account for. Systems are long term - they do not depend on a single race but the overall profit/loss result from a long series of races. Systems do not even consider factors they do not account for - by definition. Looking for the unknowable issue in a single race is by definition not ever going to work out. Just allow for it in the price.

Thinking in terms of ranges applies to any betting or method under uncertainty. The price you want to take depends on the amount of uncertainty YOU have about the selection/s. That uncertainty is found by analysing how YOUR method works or does not work over the long term. You know then what extra you need in a price to cover the "unknown unknowns".

Long term you do not even have to think about the factors that just crop up in the occasional single race - just be "righter" on average more often, long term, about the factors that consistently crop up in every race. That consistency is often easiest to obtain from a model (computer or hand driven) - which does not have emotions nor hunches. Changing your model because something else has cropped up in a particular race leads inevitably to inconsistency and a failed model or method.
I Think you are correct. In races that I feel are chaos races, I tend to spread much more than in races where I have a strong opinion. I only bet these races when I think a good outcome will result in a fairly big return. In these type of races,I will never use the favorite as a key.

[pondman]

Are you suggesting that the human mind can quantify this accurately?

Nope.

I'm saying the opposite.

Statistics is the wrong tool for making money at the track.

When you single out a horse based on a single variable, such as a potential repeat off a maiden win, you hope the horse has many unknowns...slow times, low percentage jockey.

None of the variable can be quantified. The math becomes gimmicky, and the human mind might believe: theses horses sometime win at $8,000.

If you examine this, the problem becomes an issue of conditions, which is another vague idea. After enough observations a person may decide to bet on this idea, because it returns a profit. However, you methodology wouldn't be quantitative...It becomes an expert method base on experience.

[pondman]

No, systems only work if you have one or two particular angles that the market does not fully account for.

Do you mean the crowd doesn't understand the angle (for example class)? And get's it wrong?

[Capper Al]

I understand what the modelers are trying to get at. If you set up a model correctly and have tested it with a large number of races then the same set of circumstances over the long run should produce a similar outcome and make a profit. I think, those of us in disbelief of modeling, are saying is that we don't buy that a repeatable model can be made because of too many changing variables and circumstances.

[Cratos]

No, systems only work if you have one or two particular angles that the market does not fully account for. Systems are long term - they do not depend on a single race but the overall profit/loss result from a long series of races. Systems do not even consider factors they do not account for - by definition. Looking for the unknowable issue in a single race is by definition not ever going to work out. Just allow for it in the price.

Thinking in terms of ranges applies to any betting or method under uncertainty. The price you want to take depends on the amount of uncertainty YOU have about the selection/s. That uncertainty is found by analysing how YOUR method works or does not work over the long term. You know then what extra you need in a price to cover the "unknown unknowns".

Long term you do not even have to think about the factors that just crop up in the occasional single race - just be "righter" on average more often, long term, about the factors that consistently crop up in every race. That consistency is often easiest to obtain from a model (computer or hand driven) - which does not have emotions nor hunches. Changing your model because something else has cropped up in a particular race leads inevitably to inconsistency and a failed model or method.

Well stated

[Cratos]

Nope.

I'm saying the opposite.

Statistics is the wrong tool for making money at the track.

When you single out a horse based on a single variable, such as a potential repeat off a maiden win, you hope the horse has many unknowns...slow times, low percentage jockey.

None of the variable can be quantified. The math becomes gimmicky, and the human mind might believe: theses horses sometime win at $8,000.

If you examine this, the problem becomes an issue of conditions, which is another vague idea. After enough observations a person may decide to bet on this idea, because it returns a profit. However, you methodology wouldn't be quantitative...It becomes an expert method base on experience.

If "Statistics is the wrong tool for making money at the track" then please tell what the correct tool is because a lot of players are using the wrong tool and you with your knowledge of the correct tool should be exploiting that ignorance big time.

[Native Texan III]

Do you mean the crowd doesn't understand the angle (for example class)? And get's it wrong?


How do you get from what I said " No, systems only work if you have one or two particular angles that the market does not fully account for" to "the crowd doesn't understand the angle and gets it wrong"?

The angles I was thinking of are more hidden than "class".
But you can use "class" as an example of where today it is under-appreciated in terms of say Beyers and pace, when in earlier times class and prizemoney were over-appreciated and time raters cleaned up.

In any market, things are mostly never so bad nor as good as they look. Those who do not not fully account for that lose to those who do.

[Native Texan III]

I understand what the modelers are trying to get at. If you set up a model correctly and have tested it with a large number of races then the same set of circumstances over the long run should produce a similar outcome and make a profit. I think, those of us in disbelief of modeling, are saying is that we don't buy that a repeatable model can be made because of too many changing variables and circumstances.


No wish to convince you against your will but you may be thinking what modelling entails in the wrong way.

Modellers should not discount any angle that is proven profitable long term. They might discount "angles" which are simply folk putting a cause to a transient effect which turns out to be simply plain wrong.

"Too many changing variables and circumstances".
A horse race is possibly the most controlled and repetitive event you can think of. After thousand upon thousand of races have been run there are almost no unanticipated surprises left. There are very few changing variables and circumstances that matter in racing over the long term.

How many of these minor "angles" are the sole cause of a horse winning or losing? How many of these angles might effect a rival horse to the benefit of your horse? If you can't put some figures to that then you cannot say one way or the other.

Some simplified examples:

Dayjur was coasting into winning a Breeders Cup race. It jumped a shadow cast across the track by something in the stands and lost a race by a small margin it could have won any day of the week. Is that an angle you would include in a model - no, because it and odd factors like it dilutes the weight of other factors. If the model had selected the chasing horse, the model would have scored a correct win prediction without including anything about shadow jumping. The minor angles even themselves out to zero long term.

A favorite has been off the track for 6 months:
It loses - the commentator says the horse lost because the trainer messed up and the the horse was unfit.
It wins - the commentator says the horse won because the trainer gave the horse a rest and got it really fit for its reappearance.
An "angle" is the sole cause of a defeat or a win? - what about the rest of the horses? Who really knows if the horse was actually fit and still lost or unfit and still won - only the commentator apparently, with a reason plucked from thin air.

What is the angle? Avoid horse with a lay-off or back them?
You can only answer that properly by looking at what happens in general and in particular for that trainer. A handicapper can't be like the commentator with the "right" wrong answer every time after the event.

[thaskalos]

No wish to convince you against your will but you may be thinking what modelling entails in the wrong way.

Modellers should not discount any angle that is proven profitable long term. They might discount "angles" which are simply folk putting a cause to a transient effect which turns out to be simply plain wrong.

"Too many changing variables and circumstances".
A horse race is possibly the most controlled and repetitive event you can think of. After thousand upon thousand of races have been run there are almost no unanticipated surprises left. There are very few changing variables and circumstances that matter in racing over the long term.

How many of these minor "angles" are the sole cause of a horse winning or losing? How many of these angles might effect a rival horse to the benefit of your horse? If you can't put some figures to that then you cannot say one way or the other.

Some simplified examples:

Dayjur was coasting into winning a Breeders Cup race. It jumped a shadow cast across the track by something in the stands and lost a race by a small margin it could have won any day of the week. Is that an angle you would include in a model - no, because it and odd factors like it dilutes the weight of other factors. If the model had selected the chasing horse, the model would have scored a correct win prediction without including anything about shadow jumping. The minor angles even themselves out to zero long term.

A favorite has been off the track for 6 months:
It loses - the commentator says the horse lost because the trainer messed up and the the horse was unfit.
It wins - the commentator says the horse won because the trainer gave the horse a rest and got it really fit for its reappearance.
An "angle" is the sole cause of a defeat or a win? - what about the rest of the horses? Who really knows if the horse was actually fit and still lost or unfit and still won - only the commentator apparently, with a reason plucked from thin air.

What is the angle? Avoid horse with a lay-off or back them?
You can only answer that properly by looking at what happens in general and in particular for that trainer. A handicapper can't be like the commentator with the "right" wrong answer every time after the event.
Allow me to present this argument in a slightly different way...

Some of us "intuitive" handicappers do not see this game as a mathematical puzzle...because we think that the INTERPRETATION of the figures is more important than the figures themselves.

In another thread, TrifectaMike suggested that a group of speed figures can be used to estimate the winning chances of the corresponding horses.

To some of us...speed ratings are only the byproduct of the DYNAMICS which govern the race. And these dynamics cannot be properly understood unless the handicapper has the knowledge and the experience to properly analyze the CIRCUMSTANCES under which these figures were earned.

[pondman]

If "Statistics is the wrong tool for making money at the track" then please tell what the correct tool is because a lot of players are using the wrong tool and you with your knowledge of the correct tool should be exploiting that ignorance big time.

Statistics is great for predicting your child's chances of smoking cigarettes, if you smoke and the child is born in Louville, Kentucky. It's great for predicting your child's chances of growing to over 6 feet tall if his parents are from Ireland.

Betting on horses is a game, it's a mess, and requires at most a sixth grade education and a decent memory or a recording device (beans or rocks will do.) Sixth grade is about the time I was taught set theories, which leads to simple algorithms, such as "if then", for manipulating the sets.

Observe enough races, and a person will stumble onto what works and what doesn't work at a unique track. I think it's more important to keep records of your observations, and develop a methodology.

I like the question, do horses repeat off of cheap, slow maiden claiming wins? Do they? How many people (at your favorite track) can answer this? You can utilize fourth grade math and add the columns on a paper bag. And you'd come up with a gimmicky method-- and potentially make a lot of money over time. Because the crowd is off on some quantitative, speed number. Or avoid the disaster, because cheap, slow maiden claimers don't repeat at some tracks. Is it a play, pass, or bad bet to avoid?

How is statistic going to handle this problem? And why waste the time (other than for your own amusement)?

[TrifectaMike]

In another thread, TrifectaMike suggested that a group of speed figures can be used to estimate the winning chances of the corresponding horses.


Let me correct the abopve statement. TrifectaMike is not suggesting that a group of speed are the sole factor utilized to estimate the winning chances of the corresponding horses. TrifectaMike is using speed ratings to determine a probability of winning based on a single variable, speed rating gaps as an example. That is not to say that speed ratings gaps are the only factors considered. Actually the truth is far from using only one variable.

Mike (Dr Beav)

[pondman]

How do you get from what I said " No, systems only work if you have one or two particular angles that the market does not fully account for" to "the crowd doesn't understand the angle and gets it wrong"?


I believe most people at the track have no understanding of horse racing or the business of horse racing. I don't think the crowd understands a horse can go from being a 15 length loser to a 5 length winner (in many cases), in a follow up race at a different condition.

Not only does the crowd have no beans to add to beans because it doesn't understand a bet represents a business; But when it sees something unusual (shippers, layoffs, changes in distances) the crowds going to ignore it
completely, as if it's a negative. The crowd gets it wrong.

Betting on a horse is not anything close to putting your money into a self correcting, forward looking instrument of deposit. It's closer to buying tulips in a panic bubble, from a blind man-- a market which never will reach equilibrium. It crashes. And you hope you've gotten out alive.

[thaskalos]

Let me correct the abopve statement. TrifectaMike is not suggesting that a group of speed are the sole factor utilized to estimate the winning chances of the corresponding horses. TrifectaMike is using speed ratings to determine a probability of winning based on a single variable, speed rating gaps as an example. That is not to say that speed ratings gaps are the only factors considered. Actually the truth is far from using only one variable.

Mike (Dr Beav)
I was not picking on you Mike...nor did I indicate that the speed figures would be the "sole factor" you would utilize in the creation of your winning probabilities.

I was just saying that speed figures say very little without the proper interpretation...and that, IMO...this interpretetion requires a certain level of skill and experience, along with whatever mathematical acumen the horseplayer might possess.

Again...I am not implying that you yourself don't possess this skill and experience (I know that you have been playing this game for 45 years)...just making a general comment.

[Nitro]

I will never, ever, understand the mindset of someone like Nitro, who insists on poo-pooing the thoughts of those who have embraced a mathematical solution to this game. And usually, the only defense you have is "they're animals...flesh and blood...you can't possibly incorporate all the intangibles into your equations."

News flash: HE DOESN'T HAVE TO IN ORDER TO PROFIT HANDSOMELYI can't comment on the so-called "Profit Handsomely" supposition, but I’m at a loss as to why you can’t accept the realities of the game. If those outside the game decide that using Math is their answer to finding the value that they need to succeed, then I say let them keep searching. When they discover that their quests are in vain, they’ll either become frustrated and in all likelihood look for another approach or quit the game entirely.

Friend - The Truth requires no Defense. As far as I'm concerned the burden of proof is on those who lay claim to using Math in order to succeed, and not on those using a common sense approach.


Then how are we suppose to handicap the races, with ESP or astrology? Right now my ESP is flat and my horoscope is bad. It sounds like you have a very limited background in horseracing. Instead of following the “numbers” crowd, why not take the time to learn a bit more about the horses themselves and those who control every aspect of their racing careers? You just might come away a different perspective on the entire game!


After reading your entire post it was the last paragraph that got to me and in particular when you say: “If you believe that the numbers, statistical models and probability theories will someday overcome or better yet explain nature, then I've got some land on the Moon you might be interested in.” That statement clearly tells me that you have a lack of understanding of mathematical or statistical modeling.

Modeling is about the ASSUMPTIONS you make and the mathematical or statistical formulas you use. Horseracing is not much different from any other dynamic endeavor when it comes to a predictive model. Physics is the “mother” science in these endeavors and math and statistics are the tools which are used to tighten the screws, but if your assumptions are incorrect you will strip the threads to use a mechanical analogy.

In horseracing we have the variables: velocity (it is speed when not directional), acceleration, distance, and load (weight). The trick is what assumptions you make about those variables with respect to environmental conditions, breeding, track surface, and track geometry.

Yes, there are the non-parametric variables like class and style that can cause the model to go awry. Also there is the variability caused by human intervention (e.g. trainer, owner, and jockey), but people like Mike with his knowledge of Bayesian statistics would be very helpful in weighing those variables.

In summary a carefully designed model with good ASSUMPTIONS is the best predictor of future events no matter if it is horseracing, the weather, the stock market, etc.
You are correct. It’s not that I’m not aware of the techniques that you’ve described. It’s just that I’ve left all of my knowledge of statistics and modeling back in my days in college. They’ve never been of much use to me.

But, I have a real problem with your suggestion that we base our play simply on “Assumptions” derived from Statistical Models. In fact TrifectaMike seems to imply something just the opposite from what your saying, when he mentioned that, “Bayesian probability models are based on evidence...not the past.” As I responded: I have real problem with this rationale! Are you trying to tell us that “evidence” by its very definition is not based on something factual that already exists? I think common sense says otherwise!

Also by definition there seems to be a conflict between what you’re asking us to believe about how successful these methods might be and their actual basis for implementation.

Assumption:
1) Something taken for granted; ie: Something that is believed to be true without proof
2) Inclination to high expectations, ie: The tendency to expect too much.
3) Unproved starting point, ie: In Logic - something taken as a starting point of a logical proof rather then given as a premise.

Synonyms: Belief – Guess - Best Guess – Notion – Idea –Theory – Possibility – Hypothesis –Deduction - Presumption – Inference – Speculation - Conjecture

Antonym: Fact

[Nitro]

What I am saying is that based on odds data, analysis of the data shows that horse players (obviously not all) are still betting on horses that have maximum probability of winning instead of betting on horses whose expected rewards are maximum (winning probability multiplied by odds). If players bet in a manner to maximize their rewards, it would have to be equivalent for every horse in every race that has been decided.
Mike (Dr Beav) I completely agree that the successful players do not gauge their betting activities by looking at the “maximum probability of winning”. Those who have a relatively good track record for making selections simply look at the odds value on a play by play basis to determine if the risk warrants a play. At least that’s what I do.


In my opinion, the really good players do not give away winners because of price. The best players pick enough big price horses, so that they don't have to throw out their short price horses because of lack of value. If you are a player whose best picks are only going to be in the even money, two to 1 range, you have no shot. If the only time you can get value is by going with your 4th or 5th choice you are in trouble. I tend to agree with your idea about only making value plays, but I would only make them when they’re the 1st and 2nd choice. That’s assuming that you might have a method that will point them out on a consistent basis.

Aside from my enjoyment in playing exotics, I also like making routine long-shot spot plays. Very often these returns fuel my bankroll to play exotics properly. As an example, I’d like to show you my records for play during the month of June at Churchill Downs. I’ve included only the “meat” portion of my table. My systematic selection process is always value driven, so I never really worry about that aspect of the game. (See attachment)

I won’t go into what dictates my selection process (for obvious reasons), but you’ll notice that sometimes I have to play 2 entries in a single race. Since I’m only playing long-shots it really makes no difference. My main reason for posting this is to demonstrate that not only does it provide a decent return, it tells me something even more important: When to pull the trigger! I don’t have to play every race and that’s important to improve the returns even more! The chart is laid out in a simple sequential order by dates and races. Based on the frequency of the hits, I’m able to determine when my next play should be in order to maximize the profit margin. (Even you Math guys should have field day with that !) I believe it’s important to not only focus on what to play, but when to play it. I also use a continuous month-by-month charts to determine how much to play based on the overall hit frequency. This whole charting scheme also is a great tool for managing your money.

BTW in the past, I’ve read here and there on the PA forums about some people having extended losing streaks of 20 to 30 races! My suggestion would be to go back to the drawing board and try to not only improve the selection process.

For those of you still scratching your heads wondering about using some sort of mathematical approach, I’m happy to say that I got away from that many years ago for many of the reasons previously mentioned. However there is one area where its still very useful, and that’s having the ability to count your money!

[PaceAdvantage]

I can't comment on the so-called "Profit Handsomely" supposition, but I’m at a loss as to why you can’t accept the realities of the game. If those outside the game decide that using Math is their answer to finding the value that they need to succeed, then I say let them keep searching. When they discover that their quests are in vain, they’ll either become frustrated and in all likelihood look for another approach or quit the game entirely.So what you are saying is that it is impossible to profit long term by using Math.

If that's not what you're saying, then you are a poor communicator, because that is what is coming across.

Why exactly can't you profit long term by using math? And how did you arrive at this conclusion exactly?

[Native Texan III]

Allow me to present this argument in a slightly different way...

Some of us "intuitive" handicappers do not see this game as a mathematical puzzle...because we think that the INTERPRETATION of the figures is more important than the figures themselves.

In another thread, TrifectaMike suggested that a group of speed figures can be used to estimate the winning chances of the corresponding horses.

To some of us...speed ratings are only the byproduct of the DYNAMICS which govern the race. And these dynamics cannot be properly understood unless the handicapper has the knowledge and the experience to properly analyze the CIRCUMSTANCES under which these figures were earned.

Some folks too often see things only in black or white terms. I am not saying use models as the one and only tool. I use statistics to reveal the actual hard truth of the past. Models to see what is the range of things that might happen in the dynamics of today's race, and computers to give an unbiased arithmetically accurate values of prices that are proven long term to be value. I use intuition/ experience, whatever you call it, to alert me to something that is not quite right.

If you have not got the solid background data analysis informing you in the first place, then intuition has no solid basis and folks can react like headless chickens to things which turn out to be irrelevant (Wall St most every minute of every day). You need a solid base of facts against which to judge rationally the possible range effect of circumstances. Note the words range and possible effects.

What meaning has the interpretation of figures other than maths /statistics?

Why do you think / presume that the dynamics of races are ignored in any aspect of computer rating, let alone speed figure estimation?

Why do you presume that people who use more sophisticated analysis tools have no understanding nor experience of racing?

Intuition only folks will have to provide some concrete examples of why they think ignoring all the rest of the tools in the armoury actually works.

[Tom]

There are a lot of ways to make money in this game, no one has a monopoly.
Just because we do not understand all of them doesn't make them any less profitable for those who do.

Best to limit our statements to what we understand and not make ourselves look foolish to those doing what we say can't be done.

[thaskalos]

Some folks too often see things only in black or white terms. I am not saying use models as the one and only tool. I use statistics to reveal the actual hard truth of the past. Models to see what is the range of things that might happen in the dynamics of today's race, and computers to give an unbiased arithmetically accurate values of prices that are proven long term to be value. I use intuition/ experience, whatever you call it, to alert me to something that is not quite right.

If you have not got the solid background data analysis informing you in the first place, then intuition has no solid basis and folks can react like headless chickens to things which turn out to be irrelevant (Wall St most every minute of every day). You need a solid base of facts against which to judge rationally the possible range effect of circumstances. Note the words range and possible effects.

What meaning has the interpretation of figures other than maths /statistics? When I say "interpretation of figures", I mean knowing when they are more or less impressive than they appear to be...due to the circumstances under which they were earned. Factors like TRIPS and TRACK BIASES often render these speed figures meaningless...and these factors have little to do with math/statistics, IMO.

I have been "interpreting figures" for years now...eventhough I have long forgotten any math/statistics I learnt in college.

Why do you think / presume that the dynamics of races are ignored in any aspect of computer rating, let alone speed figure estimation? IMO, these "dynamics" are very SUBJECTIVE in nature...and not easily measured by machanical means. But I could be wrong...

Why do you presume that people who use more sophisticated analysis tools have no understanding nor experience of racing? I meant to imply no such thing...as I already stated to TrifectaMike in my most recent post to him. And - as I posted in another post (#119) - I think that those players with BOTH, vast mathematical knowledge AND experience, are probably the game's biggest winners.

Intuition only folks will have to provide some concrete examples of why they think ignoring all the rest of the tools in the armoury actually works. I think it goes back to the old argument of whether handicapping is an ART or a SCIENCE. I happen to think it's an ART.

Of course, my own mathematical deficiency might have something to do with this opinion. :)

Texan, I have tried (above) to answer your questions as best I could...and would like to add that I meant no disrespect to those handicappers with a mathematical preference.

I have the utmost respect for ALL winning players - regardless of methodology.

[thaskalos]

There are a lot of ways to make money in this game, no one has a monopoly.
Just because we do not understand all of them doesn't make them any less profitable for those who do.

Best to limit our statements to what we understand and not make ourselves look foolish to those doing what we say can't be done.
Great advice Tom!

I think I will take it to heart from now on...:)

[TimesTheyRAChangin]

BTW in the past, I’ve read here and there on the PA forums about some people having extended losing streaks of 20 to 30 races! My suggestion would be to go back to the drawing board and try to not only improve the selection process.

Do you know that your very own example,that you posted here,is on a streak of 30 races w/o picking a winner?

TTRAC

[thaskalos]

Do you know that your very own example,that you posted here,is on a streak of 30 races w/o picking a winner?

TTRAC:lol:

I counted 30 concecutive losing bets, spread-out over 27 races.

But, having read Nitro's disparaging remark about the long losing streaks...I thought I made a mistake.

[TimesTheyRAChangin]

Do you know that your very own example,that you posted here,is on a streak of 30 wagers w/o picking a winner?

TTRAC

Thanks T!
Fixed it.

TTRAC

[Capper Al]

All this heavy talk about math and modeling. Yet the simple example that shows how a standard deviation of 1 length between only two horses might cause the slower horse to win about 1/3 of the time wasn't talked about. If a deviation of 1 length could possibly have such a dramatic effect between 2 horses, how can we figure value in today's race with 8 horses? At least, the modelers admit they can't on today's race but will get value in the long run. There isn't any value to be figured in today's race. The value is in the series of plays if it's to be had.

[proximity]

Now here's the important part. The quantifying of value that we did wasn't done through feel or intuition. Over the entire course of the PAIHL contest I never once looked at past performances and hope my teammates never once looked at them either. (I'm not kidding about that last part.)

We were able to accomplish what we did through number crunching... using math and algorithms exclusively.


if you're good enough to win like that (and i dont doubt that you are) then i'm sure your judgement is good enough that any application of it would not hurt your long term results.

[Nitro]

So what you are saying is that it is impossible to profit long term by using Math.

If that's not what you're saying, then you are a poor communicator, because that is what is coming across.

Why exactly can't you profit long term by using math? And how did you arrive at this conclusion exactly?From what I’ve gathered on this thread the question raised was about whether or not some Mathematical formula, process, system, model, etc. could be used for finding valued plays. My personal opinion is that in the long run this approach will fail, and as far as I’m concerned the burden of proof is not those who have similar beliefs.

I’m certainly not a Ronald Regan when it comes to articulating an argument, but if that’s what you’re getting out of my comments, then maybe what I’ve written so far is hitting the mark. (As painful as it might be for some)

As an avid Speed man from the 70’s, I truly believed (as many apparently do today) that the only way to make sense all of the various horseracing variables was to apply properly weighted numerical values. This way we would have the capability of somehow measuring ability and utilizing this information in combination with the racing conditions at hand. The intention of course was to come up with a way of making reasonably good decisions after the appropriate comparisons were made.

I recognized that after a lot of trial and effort, that no matter how well I scrutinized my calculations that there always seemed to be something missing from the equation. I later realized that very often the intangibles in this game far out weigh anything that could be accomplished by using the “numbers” alone. Because these intangible factors are unquantifiable the entire mathematical scenario in many instances becomes invalid. I can quantify this statement by saying that there are certainly instances when dealing with certain Classes in racing that the “numbers” might hold up. However, since we’re talking about value here, I would also have to say that in these instances I was not the only one who foresaw the same outcome. The results of course are low valued plays.

No matter what any Mathematician, Statistician, or Probability Analyst says, they’re ALL using historical data from the past. They’re thinking (and apparently trying to convince everyone else) that because it happened that way, it must be factual evidence that can be reused to determine the outcome of a future event. Their bottom-line is an attempt to manipulate this data in order to pigeon hole each animal with specific characteristics based on their previous racing accomplishments. The belief is that once this is done, it’s just a matter of assessing how the potential of each horse fits today’s competition. Unfortunately to arrive at some conclusion there’s a lot of subjective speculation involved.

Those that recognize these shortcomings will (sometimes) compensate by attempting to compare their assessments to the odds board in order to find a valued play. The thinking is that if their contenders are underrated according to the odds value, that they’ve found an overlay. It never occurs to some that there might be a legitimate objective reason why the betting action doesn’t correspond to the results of their mathematical appraisals. Even if this is not always the case, the underlying problem here is achieving some degree of consistency. Without it, planning a reliable betting scheme that produces is very difficult.

And talking about betting schemes, I see that we have a couple of people who may not only have a problem with Math, but obviously with Reading too.

Do you know that your very own example, that you posted here, is on a streak of 30 races w/o picking a winner?
TTRAC
I counted 30 consecutive losing bets, spread-out over 27 races.

But, having read Nitro's disparaging remark about the long losing streaks...I thought I made a mistake. So let’s review what I previously posted to see if their comments are valid:
My main reason for posting this (chart) is to demonstrate that not only does it provide a decent return, it tells me something even more important: When to pull the trigger! I don’t have to play every race and that’s important to improve the returns even more! The chart is laid out in a simple sequential order by dates and races. Based on the frequency of the hits, I’m able to determine when my next play should be in order to maximize the profit margin. (Even you Math guys should have field day with that !) I believe it’s important to not only focus on what to play, but when to play it.

I had purposely left out my betting plan so that people might appreciate the raw value available over the past month. Instead I get a couple of so-called players that seem to think that to only way to skin a cat is by picking winners. Tell that to any player that understands and appreciates the concept of “hedging” when real value is at stake and you might have a hint as to what their reaction might be.

I also believe that the betting side of the game is more important then the selection process. It’s only when a player recognizes the limits of whatever selection method they’re using can they develop a successful betting strategy. For that reason, I’ve once again attached my previous chart in order to offer a simple demonstration of the monetary results when a predetermined betting scheme is included. So that there’s no confusion, I’ve reduced each playable race to a $20 maximum bet. This total is arrived at by breaking it down into a typical 1W x 3P x 6S unit type bet (of course others are also viable). Those races showing “$0.00” indicate a “No Play” situation for reasons that are best understood by those who acknowledge the importance of trying to reach a desired goal for the day.

I’m about through with this thread (at probably the relief of some), but I just wanted to add that I’m the last person to discourage anyone from attempting something that they believe is worthwhile. However, I’m also realistic. There’s comes a point whenever the results of an endeavor appear discouraging or overwhelming that you have to re-think your approach in order to survive in this game.

[thaskalos]

And talking about betting schemes, I see that we have a couple of people who may not only have a problem with Math, but obviously with Reading too.
So let’s review what I previously posted to see if their comments are valid:
My main reason for posting this (chart) is to demonstrate that not only does it provide a decent return, it tells me something even more important: When to pull the trigger! I don’t have to play every race and that’s important to improve the returns even more! The chart is laid out in a simple sequential order by dates and races. Based on the frequency of the hits, I’m able to determine when my next play should be in order to maximize the profit margin. (Even you Math guys should have field day with that !) I believe it’s important to not only focus on what to play, but when to play it.

I had purposely left out my betting plan so that people might appreciate the raw value available over the past month. Instead I get a couple of so-called players that seem to think that to only way to skin a cat is by picking winners. Tell that to any player that understands and appreciates the concept of “hedging” when real value is at stake and you might have a hint as to what their reaction might be.

I also believe that the betting side of the game is more important then the selection process. It’s only when a player recognizes the limits of whatever selection method they’re using can they develop a successful betting strategy. For that reason, I’ve once again attached my previous chart in order to offer a simple demonstration of the monetary results when a predetermined betting scheme is included. So that there’s no confusion, I’ve reduced each playable race to a $20 maximum bet. This total is arrived at by breaking it down into a typical 1W x 3P x 6S unit type bet (of course others are also viable). Those races showing “$0.00” indicate a “No Play” situation for reasons that are best understood by those who acknowledge the importance of trying to reach a desired goal for the day.

I’m about through with this thread (at probably the relief of some), but I just wanted to add that I’m the last person to discourage anyone from attempting something that they believe is worthwhile. However, I’m also realistic. There’s comes a point whenever the results of an endeavor appear discouraging or overwhelming that you have to re-think your approach in order to survive in this game.
Having tangled with you in the past, I should know better than to take your posts seriously...

First you make an off-the-cuff remark advising those bettors who experience 20-30 race losing streaks to "go back to the drawing board"...

And then you provide us with a record of your OWN play -- listing only the "MEAT" portion -- where it clearly shows what appears to be a 30 bet losing streak, over that last 27 races of your sample. The plays are marked as WIN, PLACE, SHOW and OUT...some races include multiple wagers...and some races are skipped altogether -- just like a normal log book would look like.

But now you offer us a REVISED version, where you claim that MORE of these races were "passed", and in which you reveal yourself to be essentially a "SHOW" bettor -- in this sample at least...

I say "essentially a SHOW bettor"...because the majority of your wager is bet to SHOW.

And I ask:

When you suggested that those players experiencing extended losing streaks should "go back to the drawing board"...did you think that THOSE bettors were SHOW bettors too?

Does it take a mathematical genius to figure out that a bettor who wagers ONE unit to win, THREE units to place, and SIX units to show...will avoid the long runouts which plague the more "conventional" players?

Don't blame US for misunderstanding you, Nitro.

IMO... you should worry LESS about impressing us with your mastery of the English language...and MORE about getting your point across.

[Jeff P]

The title of this thread is "Can one really figure value?"

For those of you who may not know, the late Alan Woods was a successful computer player who emphasized value in his models and occasionally posted here at Paceadvantage using the handle name Entropy.

I Googled "Alan Woods" and came up with the following article:
http://www.tonywilson.com.au/writing/alanwoods.php

It's September 2005 and I'm sitting in a beanbag on the 32nd floor of a luxury Manila apartment block, structuring a sentence that I suspect is going to end with 'if that isn't a rude question'. I know the man opposite me is worth at least US$150 million, I know the vast majority of his fortune has been won on the Hong Kong races, and I know he dislikes being asked how much he is worth. Nevertheless I have to go for it - because the how-Alan-Woods-became-a-successful-gambler story doesn't pack the same punch without the how much.

I ask the question and we bathe in its stench for a few seconds. Alan runs a hand through his white game-show host hair.

'Hundreds of millions Australian,' he says. 'Much less than a billion.'



-jp

.

[Cratos]

The title of this thread is "Can one really figure value?"

For those of you who may not know, the late Alan Woods was a successful computer player who emphasized value in his models and occasionally posted here at Paceadvantage using the handle name Entropy.

I Googled "Alan Woods" and came up with the following article:
http://www.tonywilson.com.au/writing/alanwoods.php





-jp

.
Avery good read; thanks for posting it.

[dnlgfnk]

I've read everything I could about the Woods/Benter adventures previously.
In rereading however, the funniest line was,

"I ask Alan whether he ever has any moral pangs about what he does", after the author went to great pains to document Mr. Woods' lack thereof.

May the souls of the departed rest in peace.