assigning chances to odds

[cordep17]

I've read that, if a horse is at its equillibrium price (not an underlay or overlay), it has a specific chance of winning a race.
2-1...33%
4-1...20%
9-1...10%
Generally, I just kind of go with the feel of it. If they seem like favorable odds, I take them.
What I want to know is whether there is an educated way of assiging these percentages, or is it just a gut feeling?

[Robert Fischer]

I've read that, if a horse is at its equillibrium price (not an underlay or overlay), it has a specific chance of winning a race.
2-1...33%
4-1...20%
9-1...10%
Generally, I just kind of go with the feel of it. If they seem like favorable odds, I take them.
What I want to know is whether there is an educated way of assigning these percentages, or is it just a gut feeling?

You have to develop a feel.

It depends how much effort you want to put into it.

There are also programs that attempt to automatically assign odds.

You can also do things like rank the horses in order, and then compare your rankings to the odds.

[Robert Fischer]

i think when I first started to mess around with thinking in terms of percentages, I had to learn to make a "conservative estimate".

At first, a lot of 2-1 shots looked like they had a 40-50% chance to win. I realized 50% was not conservative enough for my estimate. In time I got a better feel.

You don't have to assign odds to every horse, in every race, if that doesn't fit your style of play.
At this point, I am doing a lot of specific plays and specific bet-sizes(% of bankroll). These plays(and specifically with those ranges of bet-size) work best in certain odds ranges, and I have a good feel for those certain odds ranges, and whether or not a given horse is a match.

[cordep17]

You can also do things like rank the horses in order, and then compare your rankings to the odds.

This assigning percentages really sort of caught me off guard. I pretty much do that already, but lately I've been looking a lot more for underlays, so hopefully they will pop out at me more.

I appreciate the response, thanks. I'll be sure to keep em' conservative.

[davew]

underlays are bets that are overbet compared to their probabilities -


some people like these because they feel people know something they do not and it has smart action

other people look at these and say, if this horse is overbet, are there other opportunities in the race for me?...

[HUSKER55]

OVERLAYS and underlays are relative figures because I don't believe there is a fixed value. In other words if the public says 3:1 and you think the horse is 5:1 then it is over bet. But if I think the horse should be 2:1 then the horse is an underlay.

it boils down to you and your line.

[Overlay]

I would say that, although each individual race is a unique, never-to-be-repeated set of circumstances, it is still possible to develop objective criteria that can determine whether a horse is overlaid or underlaid with sufficient reliability to make profitable long-term play possible. But, to me, it's not possible for one person to say that a specific horse in a specific race is overlaid, and for another to say that the same horse is underlaid, and for both of them to be right, even though that's what their different methodologies may be telling them. That's not to say that they can't both be successful in the long run, but (in my opinion), if they're discussing the same horse, one of them has to be wrong (at least as far as that race is concerned).

[Capper Al]

You have to develop a feel.

It depends how much effort you want to put into it.

There are also programs that attempt to automatically assign odds.

You can also do things like rank the horses in order, and then compare your rankings to the odds.

A feel is the best one can do for odds. The premise of an oddsline, this type of horse wins one out four times, is false. We never really know the likliness of the situation. I have made models that parallel the trackman's, but it didn't work in the long run. The best that I heard came from Dave's advertisement about one of his new books. If a horse is going to be under even money he should have the best speed figure, but a long shot going off at 10/1 or better need only to have the third best speed figure.

[jerry-g]

I use a spread sheet program to figure my own odds line. I just love it.
No two people will come up with the same odds unless they too are
using your handicapping factors in the same way. I use about 5 factors
with one optional to assign each horse a rank in numbers for his past
performance with that factor. Then my program sums up that column
for all the horses being considered and then divides each horses sum
of the factors by the sum of all the horses to get a probability figure.
Then V-LOOKUP goes to my odds table and finds the corresponding
odds for that figure and displays it for that horse. My top horse may
not be shown as low odds as I get 3-1 to 5-1 sometimes. The beauty
of it all is that I can judge how to bet by engaging only for the times
the horse goes off at my odds or higher. Thus it keeps me out of races
where I would have to bet lots of money to make a dime. The guys
that made the program say you will loose 4 out of 5 races using it,
but in the long run that one winner will pay off all your loses and make
you a profit.
I also have a program that keeps tabs of how your odds line is doing
so you can analyze how it is performing.
It all boils down to how good you are at picking the handicapping factors
to use for your odds line.

[burnsy]

You have to develop a feel.



You can also do things like rank the horses in order, and then compare your rankings to the odds.

This what i do...i don't use programs. When your rankings are out of whack with the odds of the favorite and a horse you consider a serious contender. Those are the best to jump on....it requires analyses and developing good intuition. Sometimes forgetting the numbers and statistics is a good thing, developing a horse player "street smarts" is huge in this. Being a "fan" of certain horses is fine, but if thats also your betting or handicapping method you are lost already. Reading posts where horses are bigger than life or "can't lose" just builds the serious handicappers confidence....shit happens, its still a sport and like any sport, you rarely, rarely win all the time. Even the Mighty Casey struck out.......in other words be objective and avoid the band wagon at all costs....most people that gamble, lose all the time, theres your bandwagon, a pleasant ride. If you can't critically think for yourself, you should be pushing a button on a slot machine. I've never got rich doing this, but i've never borrowed a dime at the track either. One of the reasons i think i've survived is because i can and will think outside the box. Its something thats so lacking in this society that its pathetic.....this is not just horse racing but politics, social policy and EVERYTHING. Be a people not a sheeple, take out the emotion, the bigger than life hype of this world and think for yourself. I'm not the smartest man in the world but Christ when i see what goes on and the crap people BELIEVE IN.....its so easy to be arrogant and a jack ass which i know i am. Most people feed off what they hear and see, look (research) for yourself, think for yourself. If you have been betting horses for over 5 years and you take "face value" what other "experts" say...you are not learning to bet for yourself. Numbers are great but you have to "feel" how sports and competition works, the levels , the conditions, the surfaces and the distances........FOR YOURSELF, BY YOURSELF AT SOME POINT.

[bob60566]

Here is intresting Angle in the concept with Ml odds and percentages for smaller tracks.
Your contenders for each race are all horses 4/1 and including 8/1. your contenders are now in the 20% to 11% range, Very simple and effective.

[traynor]

A feel is the best one can do for odds. The premise of an oddsline, this type of horse wins one out four times, is false. We never really know the likliness of the situation. I have made models that parallel the trackman's, but it didn't work in the long run. The best that I heard came from Dave's advertisement about one of his new books. If a horse is going to be under even money he should have the best speed figure, but a long shot going off at 10/1 or better need only to have the third best speed figure.

I think it may be conceptually impoverished to extrapolate one's own experience into a generic that applies to all. You might not know that Horse A has a probability of winning of X, and that Horse B has a probability of winning of Y, but I think anyone who uses data seriously has that information--and uses it.

[thaskalos]

A feel is the best one can do for odds. The premise of an oddsline, this type of horse wins one out four times, is false. We never really know the likliness of the situation. I have made models that parallel the trackman's, but it didn't work in the long run. The best that I heard came from Dave's advertisement about one of his new books. If a horse is going to be under even money he should have the best speed figure, but a long shot going off at 10/1 or better need only to have the third best speed figure.

From what I've read...this is how the whales operate.

Their computer systems come up with odds lines and exotics probable payoffs...which they compare to the actual odds and probable payoffs in the race.

Just because we can't do it doesn't meant that no one can.

[pondman]

i think when I first started to mess around with thinking in terms of percentages, I had to learn to make a "conservative estimate".

At first, a lot of 2-1 shots looked like they had a 40-50% chance to win. I realized 50% was not conservative enough for my estimate. In time I got a better feel.

You don't have to assign odds to every horse, in every race, if that doesn't fit your style of play.
At this point, I am doing a lot of specific plays and specific bet-sizes(% of bankroll). These plays(and specifically with those ranges of bet-size) work best in certain odds ranges, and I have a good feel for those certain odds ranges, and whether or not a given horse is a match.

This is the best response I've ever read at PA. After years of analysis, an individual should be able to estimate at what percentile a horse belongs. I personally have a 10%, 20%, 30%, and 40% percentile. I don't believe a horse will ever get to the 50% percentile. All of my bets go on horses at the +40% percentile. I avoid any horse with less. So any within that range needs to be at least 9-5. But I want an extra 3 points. So I sit out until +5-1. This is long term approach. I could play much more than I do, but I want to swing for the fences on a 5-1, rather than stress on a 2-1, which over time will make 10 cents. I'll hang my money out there on a 20-1 shot, but when I do I'm going to win +40%. The other 60% of the time you've got me.

[pondman]

Their computer systems come up with odds lines and exotics probable payoffs...which they compare to the actual odds and probable payoffs in the race.

.

They do it for ever horse in the race. If the horse is 80-1, and their personal line is 20-1, they'll have a proportional wage. This however creates an inefficiency in itself, because sometimes others are in agreement with you and there's nothing you can do. Your line is 8-1. The horse is at 7-1. What do they do? And then the horse wins. That's the problem with high end precision scalping methods. Often having whales in the game helps the small player.

[MightBeSosa]

I'll hang my money out there on a 20-1 shot, but when I do I'm going to win +40%. The other 60% of the time you've got me.

Are you saying you have a subset of plays with a sample size >10 that wins at a 40% rate at 20-1+?

[HUSKER55]

that is a nice edge and keeps a player in the black

[pondman]

Are you saying you have a subset of plays with a sample size >10 that wins at a 40% rate at 20-1+?

No. I'm saying I only play a subset as a group if it hits at +40% rate. I have more than one type of group that hits high enough. My play is different at every track-- my play is entirely different at Hollywood than at Keenland. Because I dropped performance rating from my vocabulary, and don't ever look at how the horse performs, the crowd rarely is with me. If they are it's usually because of a super jock or super trainer, at which point I pass. On one of these groups it's nothing for me to walk away with more than 10k on a $300 bet. The rate is independent of the odds, which will go to 20-1 enough times in a year. My minimum is +5-1.That's standard for me.

[traynor]

No. I'm saying I only play a subset as a group if it hits at +40% rate. I have more than one type of group that hits high enough. My play is different at every track-- my play is entirely different at Hollywood than at Keenland. Because I dropped performance rating from my vocabulary, and don't ever look at how the horse performs, the crowd rarely is with me. If they are it's usually because of a super jock or super trainer, at which point I pass. On one of these groups it's nothing for me to walk away with more than 10k on a $300 bet. The rate is independent of the odds, which will go to 20-1 enough times in a year. My minimum is +5-1.That's standard for me.

I humbly suggest that anyone serious about making money wagering on horse races pay close attention to what this gentleman is saying. And less attention to agonizing over trivial differences between what Horse A did three weeks ago at one track compared to what Horse B did two weeks ago at a different track.

I don't advocate completely ignoring performance, because it can be indicative of potential. I think performance can be evaluated, quantified, and used--but not in the ways it is typically used in "handicapping."

As I mentioned elsewhere, when one can hold the only winning tickets in the exacta pool of a small to medium size track with wagers of $100 or less (commonly referred to as "a pool cleaner"), one begins to understand that one is NOT betting on horses. One is betting against the other bettors.

With all due respect to semi-mythical tales of whales and their huge wagers (implying size of wager equates to testosterone level, as if wagering were a surrogate for more interesting activities), the people who seem to be taking the most money out of the pools are doing so with leverage--not large wagers.

[PaceAdvantage]

With all due respect to semi-mythical tales of whales and their huge wagers (implying size of wager equates to testosterone level, as if wagering were a surrogate for more interesting activities), the people who seem to be taking the most money out of the pools are doing so with leverage--not large wagers.You're the only one who thinks or talks about testosterone levels when referencing whales...

And they're not mythical.

[pondman]

the people who seem to be taking the most money out of the pools are doing so with leverage--not large wagers.

I'm not sure what this means-- leverage. Any way you look at it, you've got to cash for more than you played. You've got to be on the right side of the gamble more often than the wrong side. Just as there are huge owners with piles of money, there are huge whales. Fortunately the rules of gambling apply to them as well. They've been know to lose as times. But they've got the money to go after the carry overs.

[traynor]

You're the only one who thinks or talks about testosterone levels when referencing whales...

And they're not mythical.

I did not say that whales were mythical--only that some of the tales about whales are semi-mythical.

[traynor]

I'm not sure what this means-- leverage. Any way you look at it, you've got to cash for more than you played. You've got to be on the right side of the gamble more often than the wrong side. Just as there are huge owners with piles of money, there are huge whales. Fortunately the rules of gambling apply to them as well. They've been know to lose as times. But they've got the money to go after the carry overs.

Small wagers that equal large returns are leverage. Large wagers that return much less (in percentages relative to the size of the wager) provide less leverage. With a small margin of profit, there is greater risk. And less leverage.

[thaskalos]

I humbly suggest that anyone serious about making money wagering on horse races pay close attention to what this gentleman is saying. And less attention to agonizing over trivial differences between what Horse A did three weeks ago at one track compared to what Horse B did two weeks ago at a different track.

I don't advocate completely ignoring performance, because it can be indicative of potential. I think performance can be evaluated, quantified, and used--but not in the ways it is typically used in "handicapping."



The gentleman is saying that he doesn't bother at all with any aspect of a horse's past performances when contemplating his wagers.

Could you please tell us what it is about his comment that you agree with?

[traynor]

The gentleman is saying that he doesn't bother at all with any aspect of a horse's past performances when contemplating his wagers.

Could you please tell us what it is about his comment that you agree with?

Factors other than the past performances of the horses may be relevant to the outcome of a race. As I said, I don't think PPs should be ignored completely. However, other factors may be more significant than the PPs in determining the outcome of a given race.

[thaskalos]

Factors other than the past performances of the horses may be relevant to the outcome of a race. As I said, I don't think PPs should be ignored completely. However, other factors may be more significant than the PPs in determining the outcome of a given race.

Yes...but Pondman's comment -- which you emphasized -- dealt with ignoring PPs completely. And you advised us to pay more attention to what he had to say...and less attention to the business of "agonizing over trivial differences" between horses.

I've read what Pondman wrote very carefully...and I have a hard time believing that he cashes 40% of his bets at odds ranging between 5-1 and 20-1.

Do you think that such results are achievable?

Cashing 40% of your bets while refusing to wager on anything at less than 5-1 odds?

[PaceAdvantage]

Yes...but Pondman's comment -- which you emphasized -- dealt with ignoring PPs completely. And you advised us to pay more attention to what he had to say...and less attention to the business of "agonizing over trivial differences" between horses.

I've read what Pondman wrote very carefully...and I have a hard time believing that he cashes 40% of his bets at odds ranging between 5-1 and 20-1.

Do you think that such results are achievable?

Cashing 40% of your bets while refusing to wager on anything at less than 5-1 odds?And remember, Traynor is not a man who easily accepts semi-mythical tales, of which this one just might qualify...

[cj]

I've read what Pondman wrote very carefully...and I have a hard time believing that he cashes 40% of his bets at odds ranging between 5-1 and 20-1.

Do you think that such results are achievable?

Cashing 40% of your bets while refusing to wager on anything at less than 5-1 odds?

Of course not, it is ridiculous over any decent sized sample. I defy anyone to do it at 5-1 alone, let alone including the higher odds.

[classhandicapper]

Of course not, it is ridiculous over any decent sized sample. I defy anyone to do it at 5-1 alone, let alone including the higher odds.

You might be able to do it if you made 3-5 plays a year, but we'd have to wait a 100 years to have a large enough sample to prove it. ;)

[traynor]

Yes...but Pondman's comment -- which you emphasized -- dealt with ignoring PPs completely. And you advised us to pay more attention to what he had to say...and less attention to the business of "agonizing over trivial differences" between horses.

I've read what Pondman wrote very carefully...and I have a hard time believing that he cashes 40% of his bets at odds ranging between 5-1 and 20-1.

Do you think that such results are achievable?

Cashing 40% of your bets while refusing to wager on anything at less than 5-1 odds?

You are asking me to validate pondman's claims, which is unrelated to what I said.

I humbly suggest that anyone serious about making money wagering on horse races pay close attention to what this gentleman is saying. And less attention to agonizing over trivial differences between what Horse A did three weeks ago at one track compared to what Horse B did two weeks ago at a different track.

Specifically, it might be useful to consider that there are factors influencing horse races beyond what is seen in the past performance lines. Some of those factors may take precedent over everything else when considering wagers.

I don't think that could be considered an endorsement, or that whatever numbers are tossed out should be accepted uncritically as fact.

[jerry-g]

Here are some things that I think you should check out after
you're finished with the PP's.

Last minute change to a high % Joc.
Surface change before race.
How does the horse look....today.
Crowd elects a heavy favorite.
Possibility of rain.
A horse becomes factitious and may interfere with others.

There's probably more concerns but just a few I can think of.

[traynor]

Here are some things that I think you should check out after
you're finished with the PP's.

Last minute change to a high % Joc.
Surface change before race.
How does the horse look....today.
Crowd elects a heavy favorite.
Possibility of rain.
A horse becomes factitious and may interfere with others.

There's probably more concerns but just a few I can think of.

To the above list might be added trainers at a given track that do unusually well with specific categories of horses or races. Anyone who played the Laurel/Pimlico circuit years ago was well aware of the name Bonds in relation to two-year old "baby races."

Several trainers on the southwest circuit excel at "ready to run" entries that look like complete losers in the past performances. No chicanery is suggested--only that there is more to horse racing than the numbers one sees on a past performance line.

[HUSKER55]

traynor, if you don't use pp's because of trivial numbers then can you give us a decent example of what you do use.

thanks :)

[thaskalos]

Here are some things that I think you should check out after
you're finished with the PP's.

Last minute change to a high % Joc.
Surface change before race.
How does the horse look....today.
Crowd elects a heavy favorite.
Possibility of rain.
A horse becomes factitious and may interfere with others.

There's probably more concerns but just a few I can think of.

Yes, but as you've already said...these are things to consider "AFTER you are finished with the PPs".

These secondary factors cannot replace the primary factors, which deal with the past performances of these horses.

[Robert Fischer]

you gotta use something
and you have to be good with it

it doesn't matter if it's PPs, physicality, connections, odds movement, etc.
It just has to work.

[thaskalos]

you gotta use something
and you have to be good with it

it doesn't matter if it's PPs, physicality, connections, odds movement, etc.
It just has to work.

IMO...you need to be able to handicap a field of horses, first and foremost.

Physicality, connections and odds movements can only enhance the handicapping process...they cannot replace it.

IMO...you cannot beat this game with "formula betting"; the game is much too complex for that.

For example...I find many trainer angles and jockey/trainer combinations which have shown great profits over 100-200 races. Are these the answer to my gambling dreams and aspirations?

Hardly.

100-200 more races...and the results of these angles are vastly different. The crowd somehow seems to catch up to these short-term profitable trends.

IMO...the horseplayer should strive for game mastery, even if full mastery of this game is sure to elude him.

You have to be a well-rounded player in order to survive in this game. You cannot rely on a few simplistic angles and hope to win...any more than you could have expected to beat Robert Fischer in a game of chess -- by learning a few fancy opening moves.

[Robert Fischer]

IMO...you need to be able to handicap a field of horses, first and foremost.

Physicality, connections and odds movements can only enhance the handicapping process...they cannot replace it.

IMO...you cannot beat this game with "formula betting"; the game is much too complex for that.

For example...I find many trainer angles and jockey/trainer combinations which have shown great profits over 100-200 races. Are these the answer to my gambling dreams and aspirations?

Hardly.

100-200 more races...and the results of these angles are vastly different. The crowd somehow seems to catch up to these short-term profitable trends.

IMO...the horseplayer should strive for game mastery, even if full mastery of this game is sure to elude him.

You have to be a well-rounded player in order to survive in this game. You cannot rely on a few simplistic angles and hope to win...any more than you could have expected to beat Robert Fischer in a game of chess -- by learning a few fancy opening moves.


I 100% agree that you have to be well rounded with a strong knowledge of the fundamentals.

Even if you end up specializing in a contrarian approach, your approach is going to have to be created with, and shaped by that strong knowledge of the fundamentals. It's almost like a yin-yang concept in that regard.

[traynor]

Yes, but as you've already said...these are things to consider "AFTER you are finished with the PPs".

These secondary factors cannot replace the primary factors, which deal with the past performances of these horses.

One might argue (as Bob Purdy has in the past with great vigor) that the most basic question in race analysis is not the past performances of the entries, because horses do not enter themselves in races. People do. So the most basic question is, "Why was this horse entered in this race?" Once that question is answered, the past performance issues tend to be more clarified.

An example might be, "As a prep and tightener for the upcoming XYZ Stakes at Santa Anita." That puts a slightly different spin on the numbers involved.

[traynor]

traynor, if you don't use pp's because of trivial numbers then can you give us a decent example of what you do use.

thanks :)

When did I say I don't use PPs? There is a big difference between "not using" and "not obsessing over" PPs. My error--I thought I made that clear.

[thaskalos]

One might argue (as Bob Purdy has in the past with great vigor) that the most basic question in race analysis is not the past performances of the entries, because horses do not enter themselves in races. People do. So the most basic question is, "Why was this horse entered in this race?" Once that question is answered, the past performance issues tend to be more clarified.

An example might be, "As a prep and tightener for the upcoming XYZ Stakes at Santa Anita." That puts a slightly different spin on the numbers involved.
I would be willing to compare handicapping skills with any "race conditions" handicapper in the country.

[Saratoga_Mike]

When did I say I don't use PPs? There is a big difference between "not using" and "not obsessing over" PPs. My error--I thought I made that clear.

This is a very interesting discussion. I like to learn about new approaches. Would you be willing to explain what your process is before making a bet? I'm not asking for painstaking detail, just an overview.

[traynor]

I would be willing to compare handicapping skills with any "race conditions" handicapper in the country.

I don't know anyone who handicaps based on race conditions exclusively. I was suggesting that it might be more important to start race analysis from the understanding that one is studying a group of animals entered in a specific race by people--hopefully for some rational reason.

My mention of Bob Purdy was because of a very specific technique he recommended for race analysis: thinking of the current race and every preceding race from the trainer's standpoint, beginning with the oldest race shown for that horse, and working upwards to the latest race, in an attempt to understand why that horse was entered--not just in today's race--but in every race in its past performance lines. A tedious process initially, but guaranteed to expose details that would not otherwise be as clear.

[JackS]

The least you can do is return all horses to "actuall odds". Track take is about 20% thus a 2-1 favorite real odds would be closer to 5/2. This 5/2 would only be fair odds and if you trust the morning liner, in the long run you will break even.
You could TRY an adjustment above 20% in hopes that this adjustment might put you in the black. If this did work for your track, it would no doubt take patiences and would first have to be proved by some rather extensive record keeping covering many races.

[traynor]

This is a very interesting discussion. I like to learn about new approaches. Would you be willing to explain what your process is before making a bet? I'm not asking for painstaking detail, just an overview.

Whatever works best at the moment. And that is not intended as a facetious response. I am not in love with any particular method, or any particular approach. The process of wagering on races is about the competition between people (the other bettors) rather than solely the competition between horses.

Things that worked really well last month, or last year, tend to be non-profit this month/year. It is a lot like warfare. Tactics and strategies that worked well in the last war can be disastrous when the opposition adapts tactics and strategies to negate the advantage.

[thaskalos]

Whatever works best at the moment. And that is not intended as a facetious response. I am not in love with any particular method, or any particular approach. The process of wagering on races is about the competition between people (the other bettors) rather than solely the competition between horses.

Things that worked really well last month, or last year, tend to be non-profit this month/year. It is a lot like warfare. Tactics and strategies that worked well in the last war can be disastrous when the opposition adapts tactics and strategies to negate the advantage.
I use a somewhat more simplified approach. I analyze the race thoroughly...and try to find the best horses in the race...with regard to the circumstances that they are being asked to face.

I then structure my wagers...while balancing my handicapping opinion with the payoffs offered to me on the board.

As long as my handicapping methods go beyond the obvious...I figure I'm okay.

[classhandicapper]

I identify the contenders and try to get a feel for the ranking of the horses and the gaps between them. Then I look at the odds to see if anything screams "bet" to me.

The other thing I do is try to focus my attention on horses or conditions where I believe I might have an insight that most handicapper won't. That's where I am most likely to find the kind of value that screams to me. If almost everyone is looking at the same information I am, it's unlikely I can outhandicap them by enough to have a significant edge.

[pondman]

An example might be, "As a prep and tightener for the upcoming XYZ Stakes at Santa Anita." That puts a slightly different spin on the numbers involved.

A better example would be:

The horse wins a MSW70 at Santa Anita. Nothing is impressive as far as performance-- it's mediocre. What's the trainer going to do next? Is this a high dollar horse? Forget about breaking it's fractions into pace! Forget about the speed rating! Forget about performance. You either know what to look for or you don't. The overwhelming majority of horse players don't have any data to support a horse under these conditions. They don't even know what to look for outside a beyer's rating. They are going to see lower ratings, not bet the horse, and then get smoked by the horse at 15-1.

[Overlay]

They are going to see lower ratings, not bet the horse, and then get smoked by the horse at 15-1.
That's because they will be looking only to isolate the one most likely winner while eliminating every other horse from consideration, instead of evaluating each horse's chances in light of its odds.

[pondman]

These secondary factors cannot replace the primary factors, which deal with the past performances of these horses.

There's a where and a when to the equation. And a condition. Those are important. But for most of my singles, I could take a black marker and draw through the performance. The how isn't important. In fact, reviewing the performance itself would be detrimental to the play.

There are only a few races, generally the high end stakes (Triple Crown, Breeder Cup), that I do consider the performance. But that equals to less than 5% of my play.

[pondman]

That's because they will be looking only to isolate the one most likely winner while eliminating every other horse from consideration, instead of evaluating each horse's chances in light of its odds.

Actually a repeat after a maiden victory can be an angle by itself at some tracks. It can be a continual single-- year round-- and you don't have to do anything else. There isn't any reason to consider any other horse. But that's part of a handicapper's skill, knowing stuff, rather than looking at a performance. You as the player need to know stuff like this, at the track you are playing, or you'll lose your wallet.

[Overlay]

That's because they will be looking only to isolate the one most likely winner while eliminating every other horse from consideration, instead of evaluating each horse's chances in light of its odds.
Let me amend that to, "That's because they will be looking only to isolate the one most likely winner while eliminating every other horse from consideration, instead of evaluating each horse's actual/true chance of winning in comparison to its odds."

[pondman]

Let me amend that to, "That's because they will be looking only to isolate the one most likely winner while eliminating every other horse from consideration, instead of evaluating each horse's actual/true chance of winning in comparison to its odds."

Aren't these true odds subject to each and every player's own opinion?

I'm not sophisticated enough to know if a horse should be valued at 7-1, and play it if it's at 8-1. I don't know anyone who is playing under those conditions, playing multiple overlays. That's a tricky proposition, for the simple reason, you are continually examining the crowd's behavior. But I do know when a horse has a 40% chance of winning, can ignore everything else, and take a big shot if it's at 15-1. As long as I'm getting +9-5, I would be happy at the end of the season, but I'm not going to touch it until 5-1. I'm passing up a few dimes, but have much less stress.

[Overlay]

Aren't these true odds subject to each and every player's own opinion?
My goal is to eliminate subjectivity from my assessment to the greatest degree possible through the use of statistics, which still allows the possibility of personally tailoring a wagering plan (as you suggest) through waiting for a pre-determined edge before wagering and/or sizing wagers based on the disparity between true odds and actual odds, while doing so with greater confidence than if relying on opinion or qualitative analysis.

[traynor]

My goal is to eliminate subjectivity from my assessment to the greatest degree possible through the use of statistics, which still allows the possibility of personally tailoring a wagering plan (as you suggest) through waiting for a pre-determined edge before wagering and/or sizing wagers based on the disparity between true odds and actual odds, while doing so with greater confidence than if relying on opinion or qualitative analysis.

Yeah, what he said.

Every entry can be considered a composite of attributes with varying degrees of intensity present. The composite of those attributes can be quantified for each entry, and expressed as a percentage probability of that entry winning, relative to the other entries in that specific race. That composite value can access a decision matrix of projected ROI for that composite based on empirical values (not pulled out of one's hat or elsewhere). The result (if coded properly, and based on basic statistical processes and procedures) can be expressed as a simple Go/No Go of mutuel prices.

Just as Thaskalos has pointed out elsewhere, it comes down to "I gotta get 3/1 or better to make this bet worth making." The only difference is that it can be done with a computer much, much faster. Perhaps even more accurately, because the values used in the decision matrices are known values, not estimates.

If you are curious what such a matrix looks like, it is a set of toggles:
MAY022013-YR-R08-Pace-PP1|1|4.20|F|F|F|F|F|T|T|F|F|F|T|T|T|F|F|F|F|F|F|F |F|F|F|F|T|F|T|F|T|F|F|F|F|T|F|T|T|T|T|F|F|T|F|F|F |T|F|F|T|F|F|T|F|F|F|T|F|F|T|F|F|T|T|F|F|F|T|F|F|T |F|F|T|T|T|F|F|T|F|F|F|F|F|F|F|F|F|F|F|F|T|T|F|F|F |F|F|F|F|F|F|F|F|F|F|T|T|T|F|F|F|F|F|F|F|F|F|F|F|F |F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|T|F|F|F|F|F |F|F|T|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F |F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F |F|F|F|F|F|F|F|F|F|F|F

In the race example, T is the boolean value True (indicating that particular entry had the attribute associated with that element), F is the boolean value False (indicating that particular entry did not have the attribute associated with that element). By keeping separate records of "all entries" and "winners," layered by track, class, distance, and whatever else, comparison values can be generated for each element. Specifically, the ratio of each element present in winners as opposed to that element being present in all entries.

That information can in turn be used to create a decision matrix of the probability of each entry winning, based on its attributes ("handicapping" or "race analysis"), combined with a projected ROI value to generate what amounts to an odds line. It takes a bit of work, but the effort is well rewarded.

[pondman]

My goal is to eliminate subjectivity from my assessment to the greatest degree possible through the use of statistics, .

Statistics? Are you randomly choosing a sample?

Or are you observing a variable and then observing the value of that variable over time under different conditions. For example if a horse runs a race at FG (or perhaps just trains) , and then ships into an Allowance race at Mountaineer, it's going to have about a 28% of winning.

Because many of the threads I've read at PaceAdvantage.com have major errors when I read the word statistics. They generally try to apply a non-random event to a population and they fall flat. Because these precise non-random observations, such as picking from a group of favorites, have much more unknown in them, than valuing a complex multiple variable problem, such as the horse running at mountaineer. Those types of gimmicky statements over time hold truer to their value.

[pondman]

Every entry can be considered a composite of attributes with varying degrees of intensity present.

Why worry about every entry? There are combinations of variable that produce greater than 40% wins over time. Why not wait for those horses? On the other hand there are variables (I believe) which signal instant death. Why not focus on either the real positives, or the real negatives in the case of a 3/5 shot showing sign of an instant death variable, or less than 10% chance over time. Why not focus just on the races that have either of these, or in some cases have both, and then you walk out with a bag full of cash. Why spend your day trudging through a 5 contender race, trying to hedge and cover any loss?

[Overlay]

Statistics? Are you randomly choosing a sample?

Or are you observing a variable and then observing the value of that variable over time under different conditions. For example if a horse runs a race at FG (or perhaps just trains) , and then ships into an Allowance race at Mountaineer, it's going to have about a 28% of winning.

Because many of the threads I've read at PaceAdvantage.com have major errors when I read the word statistics. They generally try to apply a non-random event to a population and they fall flat. Because these precise non-random observations, such as picking from a group of favorites, have much more unknown in them, than valuing a complex multiple variable problem, such as the horse running at mountaineer. Those types of gimmicky statements over time hold truer to their value.
I'm applying the type of methodology presented by William Quirin in Winning at the Races, where he discussed both the identification of specific variables as independent, based on the degree to which their association with positive or negative race results exceeded the normal/expected rate of occurrence; and the use of multiple regression to blend variables together into a properly weighted composite formula that could be employed to predict both the probable order of finish in a race, and also the fair odds of each horse entered.

[thaskalos]

I'm applying the type of methodology presented by William Quirin in Winning at the Races, where he discussed both the identification of specific variables as independent, based on the degree to which their association with positive or negative race results exceeded the normal/expected rate of occurrence; and the use of multiple regression to blend variables together into a properly weighted composite formula that could be employed to predict both the probable order of finish in a race, and also the fair odds of each horse entered.

Are you convinced that Quirin's methodology, as portrayed in Winning at the Races, has withstood the test of time?

[Overlay]

Are you convinced that Quirin's methodology, as portrayed in Winning at the Races, has withstood the test of time?
Some of Quirin's specific statistical findings may no longer be precisely accurate due to changes in the game over the past 30+ years, but I don't think that his discussions of the general mathematical bases for the determination of statistical independence and the application of multiple regression would be subject to such obsolescence.

[thaskalos]

Some of Quirin's specific statistical findings may no longer be precisely accurate due to changes in the game over the past 30+ years, but I don't think that his discussions of the general mathematical bases for the determination of statistical independence and the application of multiple regression would be subject to such obsolescence.

I have in on good authority that Quirin's statistical conclusions were later repudiated by the very person who hired him to conduct the research in the first place.

[traynor]

Why worry about every entry? There are combinations of variable that produce greater than 40% wins over time. Why not wait for those horses? On the other hand there are variables (I believe) which signal instant death. Why not focus on either the real positives, or the real negatives in the case of a 3/5 shot showing sign of an instant death variable, or less than 10% chance over time. Why not focus just on the races that have either of these, or in some cases have both, and then you walk out with a bag full of cash. Why spend your day trudging through a 5 contender race, trying to hedge and cover any loss?

It is not worry. All real world events in horse racing are composites that include confounding variables that are largely ignored by the average bettor. As well as being largely ignored by (or unknown to) the majority of serious bettors. It is the influence of those confounding variables that most research and studies fail to consider.

In the 40% scenarios, most would fail to grasp the difference between records in which specific, knowable, but generally considered confounding, variables do and do not exist. It is often the absence or presence of those confounding variables that determines whether a particular race is in the 40% group or the 60% group. Failure to understand that can be costly.

[Overlay]

I have in on good authority that Quirin's statistical conclusions were later repudiated by the very person who hired him to conduct the research in the first place.
By that, I assume that you're referring to the accuracy of the impact values and percentages presented as findings, rather than the validity of the mathematical formulae applied to the original raw data.

[thaskalos]

By that, I assume that you're referring to the accuracy of the impact values and percentages presented as findings, rather than the validity of the mathematical formulae applied to the original raw data.

Didn't the impact values play a key role in the application of the mathematical formulae?

[cordep17]

Yeah, what he said.

Every entry can be considered a composite of attributes with varying degrees of intensity present. The composite of those attributes can be quantified for each entry, and expressed as a percentage probability of that entry winning, relative to the other entries in that specific race. That composite value can access a decision matrix of projected ROI for that composite based on empirical values (not pulled out of one's hat or elsewhere). The result (if coded properly, and based on basic statistical processes and procedures) can be expressed as a simple Go/No Go of mutuel prices.

Just as Thaskalos has pointed out elsewhere, it comes down to "I gotta get 3/1 or better to make this bet worth making." The only difference is that it can be done with a computer much, much faster. Perhaps even more accurately, because the values used in the decision matrices are known values, not estimates.

If you are curious what such a matrix looks like, it is a set of toggles:
MAY022013-YR-R08-Pace-PP1|1|4.20|F|F|F|F|F|T|T|F|F|F|T|T|T|F|F|F|F|F|F|F |F|F|F|F|T|F|T|F|T|F|F|F|F|T|F|T|T|T|T|F|F|T|F|F|F |T|F|F|T|F|F|T|F|F|F|T|F|F|T|F|F|T|T|F|F|F|T|F|F|T |F|F|T|T|T|F|F|T|F|F|F|F|F|F|F|F|F|F|F|F|T|T|F|F|F |F|F|F|F|F|F|F|F|F|F|T|T|T|F|F|F|F|F|F|F|F|F|F|F|F |F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|T|F|F|F|F|F |F|F|T|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F |F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F|F |F|F|F|F|F|F|F|F|F|F|F

In the race example, T is the boolean value True (indicating that particular entry had the attribute associated with that element), F is the boolean value False (indicating that particular entry did not have the attribute associated with that element). By keeping separate records of "all entries" and "winners," layered by track, class, distance, and whatever else, comparison values can be generated for each element. Specifically, the ratio of each element present in winners as opposed to that element being present in all entries.

That information can in turn be used to create a decision matrix of the probability of each entry winning, based on its attributes ("handicapping" or "race analysis"), combined with a projected ROI value to generate what amounts to an odds line. It takes a bit of work, but the effort is well rewarded.

So, if you dont use a computer, you can still do this correct? Maintain the PPs for winners of each kind of race at the track and use them to guesstimate the odds of a certain horse vs. them?

[Magister Ludi]

I've read that, if a horse is at its equillibrium price (not an underlay or overlay), it has a specific chance of winning a race.
2-1...33%
4-1...20%
9-1...10%
Generally, I just kind of go with the feel of it. If they seem like favorable odds, I take them.
What I want to know is whether there is an educated way of assiging these percentages, or is it just a gut feeling?

I use a mathematical model which we’ll call the Fair Market Model (FMM) to measure the public information associated with a horse racing event. The difference between the FMM and the actual tote odds we’ll call delta(FMM). Delta(FMM) measures the information that has been incorporated into the tote odds which is not reflected in the FMM.

I have several models that measure the information which is peculiar to a particular horse racing event and not already incorporated in the FMM. We’ll call these models Racing Event Private Information Models (REPRIM). The difference between delta(FMM) and REPRIM we’ll call delta(info). Delta(info) is simply the information reflected in the tote odds which is not accounted for in the FMM or in the REPRIM. Delta(info) and tote modelling are used to evaluate underlays and overlays.

[Overlay]

Didn't the impact values play a key role in the application of the mathematical formulae?
For each given sample of races, the impact value for each characteristic examined was derived by dividing the percentage of winners of the sample races possessing the characteristic by the percentage of all horses in the sample races possessing the characteristic. By contrast, the formula for determining the expected winning percentage for horses possessing any given characteristic was PCT = .83 / (Odds-to-$1.00 + .05). (Quirin had already demonstrated the validity of the hypothesis that horses generally win at a rate corresponding to their toteboard odds, which reflect the effects of the take and breakage, according to the above formula.) Whether the actual percentage of winners differed from the expected percentage of inners to a statistically significant degree was then determined by the result of the equation

(NW - EW) divided by the square root of (EW x (1-EW/NH)).

Statistical independence of a given variable was determined by whether the above equation produced a value above +2.5 (for positive variables), or below -2.5 (for negative variables).

Whether Quirin accurately presented the data upon which he based his impact values, as well as the values themselves, I cannot say. But those impact values did not play any role in the determination of statistical independence (as shown by the fact that one could not predict, merely on the basis of an impact value appearing to be unusually high or low, whether it was associated with an independent variable or not). Nor do they affect the validity of the actual formula for calculating impact values (when correctly/honestly applied) as a means of measuring the correlation between a given variable and a horse's chance of winning.

[traynor]

So, if you dont use a computer, you can still do this correct? Maintain the PPs for winners of each kind of race at the track and use them to guesstimate the odds of a certain horse vs. them?

Yes, but it would be a huge amount of work for more than one or possibly two tracks. More than most would tackle without a computer. Way back, Paul Braseth sold something he called a Winner's Book (for the old Longacres track near Seattle). It was--in essence--the winning line of every horse, with the three or four or whatever preceding racelines leading up to that win.

The problem is that each horse is then only considered in isolation, and all the variables are considered only for that horse. It is the "confounding variables" present or absent in a given scenario that largely determine the outcome of races.

That is not to dissuade you from doing it by hand. Any efforts in that direction will only increase your skill in race analysis. That alone makes it worthwhile.

[thaskalos]

For each given sample of races, the impact value for each characteristic examined was derived by dividing the percentage of winners of the sample races possessing the characteristic by the percentage of all horses in the sample races possessing the characteristic. By contrast, the formula for determining the expected winning percentage for horses possessing any given characteristic was PCT = .83 / (Odds-to-$1.00 + .05). (Quirin had already demonstrated the validity of the hypothesis that horses generally win at a rate corresponding to their toteboard odds, which reflect the effects of the take and breakage, according to the above formula.) Whether the actual percentage of winners differed from the expected percentage of inners to a statistically significant degree was then determined by the result of the equation

(NW - EW) divided by the square root of (EW x (1-EW/NH)).

Statistical independence of a given variable was determined by whether the above equation produced a value above +2.5 (for positive variables), or below -2.5 (for negative variables).

Whether Quirin accurately presented the data upon which he based his impact values, as well as the values themselves, I cannot say. But those impact values did not play any role in the determination of statistical independence (as shown by the fact that one could not predict, merely on the basis of an impact value appearing to be unusually high or low, whether it was associated with an independent variable or not). Nor do they affect the validity of the actual formula for calculating impact values (when correctly/honestly applied) as a means of measuring the correlation between a given variable and a horse's chance of winning.

The way I remember it, impact values were created...and affixed to particular handicapping "angles".

These impact values and angles were then employed to create win probabilities for the horses in a race.

It would consequently seem to me that the accuracy of these win probabilities would hinge on the accuracy of the impact values.

I apologize if I am wrong; it's been some time since I read Quirin's book.

[Overlay]

The way I remember it, impact values were created...and affixed to particular handicapping "angles".

These impact values and angles were then employed to create win probabilities for the horses in a race.

It would consequently seem to me that the accuracy of these win probabilities would hinge on the accuracy of the impact values.

I apologize if I am wrong; it's been some time since I read Quirin's book.
For each attribute (factor or angle) or combination of attributes that Quirin examined, he presented the total number of horses possessing it/them in the sample of races that he examined; the number of winners of the sample races possessing it/them; the winning percentage of the horses possessing it/them; the percentage of all winners of the sample races represented by horses possessing it/them that won their races; the impact value for the attribute/combination of attributes, found by dividing the percentage of winners of the sample races with it/them by the percentage of all horses in the sample possessing it/them; and the net return (on a per-wager basis) as a result of placing a $2.00 win bet on each of the horses in the sample possessing the attribute or combination of attributes. He also tested each "number of winners" finding to determine if it qualified as an independent variable by differing to a statistically significant degree from the expected number of winners based on the odds of the horses in the sample races. Factors that did not qualify as independent variables by this criterion had less assurance of retaining their seemingly positive or negative character in future samples, since their performance was not attributable solely to the factor acting by itself, but was dependent upon the factor's interaction with other characteristics possessed by the horses in question, which might differ in future races.

Although Quirin included a winning percentage for the factors he presented (both individually and in combination), he did not base wagering recommendations upon those percentages, such as demanding a minimum odds level before betting. The only time that "fair odds" were discussed as a criterion for wagering decisions was in the section on "Creating an Odds Line" that was included in the discussion of his multiple regression formulas for sprints and routes, in which he presented a formula for converting a horse's multiple regression total into a winning probability, from which a fair-odds figure could be developed for use in making play/pass judgments.

[raybo]

Yes, but it would be a huge amount of work for more than one or possibly two tracks. More than most would tackle without a computer. Way back, Paul Braseth sold something he called a Winner's Book (for the old Longacres track near Seattle). It was--in essence--the winning line of every horse, with the three or four or whatever preceding racelines leading up to that win.

The problem is that each horse is then only considered in isolation, and all the variables are considered only for that horse. It is the "confounding variables" present or absent in a given scenario that largely determine the outcome of races.

That is not to dissuade you from doing it by hand. Any efforts in that direction will only increase your skill in race analysis. That alone makes it worthwhile.

I was going to say something similar. IMO, one must not only use factors of the winners, but also factors of the race he won in, the "confounding variable"? In other words, the factors of the horse, the winner, must be taken in context of the race it won in, the composite factors of all the horses in that field, as well as other "race" factors. In RS we use PPG as the "confounding variable", involving "ranges" of PPG values, for eliminations from win contention. In other words, we use the composite of the field's contribution to the race, not just individual horse attributes (that would get you nowhere close to being significant).

[pondman]

In other words, we use the composite of the field's contribution to the race, not just individual horse attributes (that would get you nowhere close to being significant).

What about when the connection knows the horse doesn't have what it takes to run at the A game level? But They'll continue to prep it, knowing they'll slip it to the B level in 10 weeks? So they run MSW twice at BEL and then send it to somewhere soft. I'd say the field of the race doesn't matter. I'd want it in against the best looking horses, as many false contenders as possible. I'd also say the conditions doesn't matter. And the distance can vary; They're going to enter it from what they've seen in the morning, not from a PP. They could probably win with any rider, just as long as they can hold on. This is a common scenario in racing. How are you going to come up with a composite? Strength of field? Class?

[Magister Ludi]

...In other words, we use the composite of the field's contribution to the race, not just individual horse attributes (that would get you nowhere close to being significant).

Exactly! Analyzing the total event is far more productive than analyzing the individual horses. Analyzing the bettors can be the most productive of all.

[raybo]

What about when the connection knows the horse doesn't have what it takes to run at the A game level? But They'll continue to prep it, knowing they'll slip it to the B level in 10 weeks? So they run MSW twice at BEL and then send it to somewhere soft. I'd say the field of the race doesn't matter. I'd want it in against the best looking horses, as many false contenders as possible. I'd also say the conditions doesn't matter. And the distance can vary; They're going to enter it from what they've seen in the morning, not from a PP. They could probably win with any rider, just as long as they can hold on. This is a common scenario in racing. How are you going to come up with a composite? Strength of field? Class?

You're talking about individual events, I'm not. Anyone who concentrates too much on individual events, instead of the long term, will probably be just another losing player in the end. Often I pass the individual events you're talking about, opting for races that favor my method, long term.

[raybo]

Exactly! Analyzing the total event is far more productive than analyzing the individual horses. Analyzing the bettors can be the most productive of all.

The bettors either give me what I want, or they don't, in which case I pass.

[pondman]

You're talking about individual events, I'm not. Anyone who concentrates too much on individual events, instead of the long term, will probably be just another losing player in the end. Often I pass the individual events you're talking about, opting for races that favor my method, long term.

Thank you for the answer!!!!!!!!

I've come here to simply throw out some ideas, and see what the crowd has to say.

I'm going to give people who read this something.

Be careful with your logic. I've read many egregious errors of logic. I'm certain you've just stepped into one. I've described what should be considered a yes or no. It's green light, red light. A player would either understand what I've just written or they don't. In other words it's highly playable, or deadly, with no middle ground. It's both long term and significant. This type of play fits well with this thread, and is easy to value.

[raybo]

Thank you for the answer!!!!!!!!

I've come here to simply throw out some ideas, and see what the crowd has to say.

I'm going to give people who read this something.

Be careful with your logic. I've read many egregious errors of logic. I'm certain you've just stepped into one. I've described what should be considered a yes or no. It's green light, red light. A player would either understand what I've just written or they don't. In other words it's highly playable, or deadly, with no middle ground. It's both long term and significant. This type of play fits well with this thread, and is easy to value.

You can talk about specific spot plays if you want, I'm talking about what most players do, playing as many races as they can, races that they believe they have an edge in, long term, not just looking for spot play or two, waiting many races between plays.

I'm sure one could produce a profit if they were disciplined enough to wait for many races in order to find a spot play, however, most players don't have the discipline to do that, nor the inclination to do so. Also, when a player can play many races, instead of only a couple, they are in a position to build bankroll much faster, with a smaller edge, than the spot player with few plays and a larger edge.

If you are going to play many races, you better be considering the composite factors of the field, not just individual horses, broken out by at least, track, distance (at least sprint or route), surface, time of year/time of meet, etc..

[pondman]

You can talk about specific spot plays

I'm talking about every horse that has left NY, S Cal, Gulfstream in the last 25 years... I can find a value of a horse as I've described, and can assign my own odds, and I don't limit it to just MSW. Can I do it and not consider the field at all? Yep. The field doesn't matter. In some cases the condition will make the horse a complete toss, so the conditions in some cases are important, but not all.

I'm just claiming to make bags of money doing the things you say that can't be done.

[thaskalos]

I'm talking about every horse that has left NY, S Cal, Gulfstream in the last 25 years... I can find a value of a horse as I've described, and can assign my own odds, and I don't limit it to just MSW. Can I do it and not consider the field at all? Yep. The field doesn't matter. In some cases the condition will make the horse a complete toss, so the conditions in some cases are important, but not all.

I'm just claiming to make bags of money doing the things you say that can't be done.

Is there some way we can persuade you to give us a small example of what you are talking about?

No "secrets", mind you. Just a few 5-1 or better horses, who win at a 40% clip regardless of the fields that they find themselves against.

[raybo]

I'm talking about every horse that has left NY, S Cal, Gulfstream in the last 25 years... I can find a value of a horse as I've described, and can assign my own odds, and I don't limit it to just MSW. Can I do it and not consider the field at all? Yep. The field doesn't matter. In some cases the condition will make the horse a complete toss, so the conditions in some cases are important, but not all.

I'm just claiming to make bags of money doing the things you say that can't be done.

Ok, congratulations!

[cordep17]

I'm talking about every horse that has left NY, S Cal, Gulfstream in the last 25 years... I can find a value of a horse as I've described, and can assign my own odds, and I don't limit it to just MSW. Can I do it and not consider the field at all? Yep. The field doesn't matter. In some cases the condition will make the horse a complete toss, so the conditions in some cases are important, but not all.

I'm just claiming to make bags of money doing the things you say that can't be done.


I may be wrong...
No, I can't be wrong
What you are saying cannot be the case.

[pondman]

Is there some way we can persuade you to give us a small example of what you are talking about?

No "secrets", mind you. Just a few 5-1 or better horses, who win at a 40% clip regardless of the fields that they find themselves against.

Doesn't matter which track you choose. N.Y shipping to Kentucky. I'll play 40-50 horses a year, at +5-1, and hit about 50%. I don't care what's in the race. The rest of my time is spent in well ventilated, and well lit environments. Just the way I've always played-- contrary to the number crunchers living in their mother's garages. I'm playing when you can play big. Avoid the game, when there isn't much. Not going to say I'm a great all around capper. But..I understand what it takes to move horses around on the highway. And that's at most levels and many tracks.

[Show Me the Wire]

A better example would be:

The horse wins a MSW70 at Santa Anita. Nothing is impressive as far as performance-- it's mediocre. What's the trainer going to do next? Is this a high dollar horse? Forget about breaking it's fractions into pace! Forget about the speed rating! Forget about performance. You either know what to look for or you don't. The overwhelming majority of horse players don't have any data to support a horse under these conditions. They don't even know what to look for outside a beyer's rating. They are going to see lower ratings, not bet the horse, and then get smoked by the horse at 15-1.

Assuming its BSF of 70, raw time 1:11 and change, and is not an expensive horse or a female with breeding potential, and not a Cal. bred run it in a 40k or less claimer, lose and enter into a $40k starter allowance.

[pondman]

Assuming its BSF of 70, raw time 1:11 and change, and is not an expensive horse or a female with breeding potential, and not a Cal. bred run it in a 40k or less claimer, lose and enter into a $40k starter allowance.

Actually the STA isn't a bad answer for repeats in a number of mcl instances. But I'm not interested in the claiming or losing part. If it doesn't win immediately off the maiden victory, I won't touch it. I don't look at the times or fractions...Times are going to get in the way.

[iceknight]

Actually the STA isn't a bad answer for repeats in a number of mcl instances. But I'm not interested in the claiming or losing part. If it doesn't win immediately off the maiden victory, I won't touch it. I don't look at the times or fractions...Times are going to get in the way. Great thread and solid discussions by you guys. Now, when you go figuring angles and then testing them.. you are basically trying to think like a trainer who gets wins (at a certain percentage level and at higher odds, right)? So, figuring out where to place an animal - based on the condition books at the track.

So, we are trying to develop a hypothetical composite trainer who has multiple live horses going off at certain odds or higher... but we are 40% certain that these horses have chances of winning?

And we wait for these chances..

I like this approach. This is what I have been training myself to do (ie not become weak and do random action bets) but wait for good value when your assessment is good based on more than the form.
:ThmbUp: :ThmbUp:

[pondman]

So, we are trying to develop a hypothetical composite trainer who has multiple live horses going off at certain odds or higher... but we are 40% certain that these horses have chances of winning?

:

As long as your reward is higher than the risk you'll make money. There are people who have precise methods that can split hairs. And so they can bet lower odds. They can bet 2-1 and 3-1. I play the other end. I prefer to look for the N. Cal horse repeating a maiden win at 25-1, ridden by someone unknown. I might get 6 shots each summer. And my confidence is lower-- perhaps at 28%. I might spend $1,500 each summer, but the reward is there. I've caught those types of horses at 60-1 with $300. I think this type of betting can be done at most tracks. You just need to stick with what works for you.

[JackS]

If your willing to play some horses with short odds assign your" can't be beat" favorite a mental odds line of 1to1. This becomes your assumption that this horse will win at a rate of 50%. If your can't be beat horse goes off at 6/5 or above, you have an overlay.
A simple little method that will require some patients.

[raybo]

If your willing to play some horses with short odds assign your" can't be beat" favorite a mental odds line of 1to1. This becomes your assumption that this horse will win at a rate of 50%. If your can't be beat horse goes off at 6/5 or above, you have an overlay.
A simple little method that will require some patients.

That's not bad, theoretically, but you're assuming the player assigning those 1/1 "mental odds" is an excellent handicapper, and that those 6/5 odds don't drop after you place your bet. Otherwise you better do a whole lot of fudging on your "mental odds". ;)

[traynor]

As long as your reward is higher than the risk you'll make money. There are people who have precise methods that can split hairs. And so they can bet lower odds. They can bet 2-1 and 3-1. I play the other end. I prefer to look for the N. Cal horse repeating a maiden win at 25-1, ridden by someone unknown. I might get 6 shots each summer. And my confidence is lower-- perhaps at 28%. I might spend $1,500 each summer, but the reward is there. I've caught those types of horses at 60-1 with $300. I think this type of betting can be done at most tracks. You just need to stick with what works for you.

I think a key point of the very solid information you are providing may be being overlooked. That is your emphasis on profit as the sole reason for wagering--not action, not love of the sport, not the challenge--the money. That mindest would automatically exclude many, many bettors from doing things that would dramatically improve their incomes.

[raybo]

I've read that, if a horse is at its equillibrium price (not an underlay or overlay), it has a specific chance of winning a race.
2-1...33%
4-1...20%
9-1...10%
Generally, I just kind of go with the feel of it. If they seem like favorable odds, I take them.
What I want to know is whether there is an educated way of assiging these percentages, or is it just a gut feeling?

Getting back to the OP's original premise and question, the premise is not correct, first of all. Unless the player is "perfect" there is no "equilibrium price". In the above "2/1...33%" example, the player is assuming that the horse has 1 chance of winning and only 2 chances of losing, but in real life he cannot know that to be absolutely true. So, one solution is to "fudge" those odds to account for the "imperfectness" of the player. Let's say that there is another horse in the race that also has a chance of winning the race, besides the, supposed, 3 that he has decided on. Maybe there are actually 4 real win contenders, instead of 3? Of course, there are some who think that they are good enough handicappers that they can actually split hairs and say a horse, among contenders has more of a chance of winning than the other contenders. But, is that actually something that the player is capable of doing? That's the question. In my mind, the answer is no. But, there are many out there who think they really can do this, consistently. Confidence? Good record keeping? Or, ego? Could be any of those. But, no matter what the reason, they are all "imperfect" estimates.

I think that anyone, who is really truthful with him/herself, will agree that they never know, 100%, which horse will win a particular race. Thorough record keeping can give one a very good idea of how often they are really "perfect" in assigning horses as win contenders. Then, that person can include that in the assigning of probabilities of any single horse, among those contenders, winning a specific race. And, if the player remains very consistent in his contender selection process, and things don't change in the interim (which is likely to happen anyway), the player has a starting point for a long term profitable method for finding "value" and making it work for him/her.

Bottom line is, it appears to me, that most people's (as in MOST), estimate of a horse's actual win probabilities are "off", many times extremely far off, and those probabilities should be adjusted to include more chance of losing each bet. Yes, you will have fewer plays, but if you can't handle that then you are just paying for entertainment anyway, whether that is the enjoyment of watching horses run, or alleviating the need to gamble, or enjoying the challenge, or whatever, you're just paying for the cost of pursuing a hobby.

What is true for you?

[thaskalos]

Getting back to the OP's original premise and question, the premise is not correct, first of all. Unless the player is "perfect" there is no "equilibrium price". In the above "2/1...33%" example, the player is assuming that the horse has 1 chance of winning and only 2 chances of losing, but in real life he cannot know that to be absolutely true. So, one solution is to "fudge" those odds to account for the "imperfectness" of the player. Let's say that there is another horse in the race that also has a chance of winning the race, besides the, supposed, 3 that he has decided on. Maybe there are actually 4 real win contenders, instead of 3? Of course, there are some who think that they are good enough handicappers that they can actually split hairs and say a horse, among contenders has more of a chance of winning than the other contenders. But, is that actually something that the player is capable of doing? That's the question. In my mind, the answer is no. But, there are many out there who think they really can do this, consistently. Confidence? Good record keeping? Or, ego? Could be any of those. But, no matter what the reason, they are all "imperfect" estimates.

I think that anyone, who is really truthful with him/herself, will agree that they never know, 100%, which horse will win a particular race. Thorough record keeping can give one a very good idea of how often they are really "perfect" in assigning horses as win contenders. Then, that person can include that in the assigning of probabilities of any single horse, among those contenders, winning a specific race. And, if the player remains very consistent in his contender selection process, and things don't change in the interim (which is likely to happen anyway), the player has a starting point for a long term profitable method for finding "value" and making it work for him/her.

Bottom line is, it appears to me, that most people's (as in MOST), estimate of a horse's actual win probabilities are "off", many times extremely far off, and those probabilities should be adjusted to include more chance of losing each bet. Yes, you will have fewer plays, but if you can't handle that then you are just paying for entertainment anyway, whether that is the enjoyment of watching horses run, or alleviating the need to gamble, or enjoying the challenge, or whatever, you're just paying for the cost of pursuing a hobby.

What is true for you?

Raybo...let me see if I am interpreting you correctly here.

Let's say that you are handicapping a 12-horse field, and you have been able to reduce this field down to 5 legitimate win contenders.

Do you automatically view all these contenders as equals...without concerning yourself with how these contenders rank among themselves, ability-wise?

[raybo]

Raybo...let me see if I am interpreting you correctly here.

Let's say that you are handicapping a 12-horse field, and you have been able to reduce this field down to 5 legitimate win contenders.

Do you automatically view all these contenders as equals...without concerning yourself with how these contenders rank among themselves, ability-wise?

Not necessarily, it depends on the handicapping skill level of the player. For the average player, I would say they have no idea which of those 5 contenders will win, or even if any of them will win. You, on the other hand, may be good enough to say one of those has a better chance than the others, but you still don't know if that horse will win that race, so you might, indeed, overestimate the horse's actual probability of winning.

My post was meant to point out that the original premise of the OP was not true, assuming that he actually determines win contenders, before assigning win probability to individual horses.

[JackS]

The public has a very good hit pct rate of 30-35% depending on track.
You might focus on the race type/class/distance of the other 66% that the public seems to have trouble with. For instance- The public may be hitting at a rate of close to 50% on certain types and close to 20% on the rest for an average of 30-35%. Bankroll manipulation and bet size should be based on generated facts with the thought that you will be able to change your original calculations as the season progresses and a need for change becomes obvious.

[raybo]

The public has a very good hit pct rate of 30-35% depending on track.
You might focus on the race type/class/distance of the other 66% that the public seems to have trouble with. For instance- The public may be hitting at a rate of close to 50% on certain types and close to 20% on the rest for an average of 30-35%. Bankroll manipulation and bet size should be based on generated facts with the thought that you will be able to change your original calculations as the season progresses and a need for change becomes obvious.

True, the public's accuracy in establishing horses' odds varies, due to a multitude of factors, some of which you mentioned, but there are of course, many others.

Long term, we know approximately the public's average hit rate, but we have no idea how accurate they are for individual races. This exacerbates an already difficult task, that being our estimate of "fair odds" or a horse's true probability of winning a race versus the price you will get (minus, or plus, any drop or rise in odds after you place your bet).

As I mentioned earlier, unless the player is a very good handicapper, the job of finding value, is at best, a crude guess. Even the very best handicappers are merely estimating a horse's win probability, and even if that estimate is perfectly correct, which is probably very rare, the horse can still lose the race. No horse is guaranteed to win, even a horse like Secretariat running in a $5k claimer, could lose, however remote that might be, it is still a possibility.

So, we're forced to work from long term averages, hoping to make a profit in that long run.