Theoretical Win Wagering Strategy Question

[dvlander]

Suppose you could isolate three contenders in a large group of races. 80% of the time, one of your contenders wins. The field sizes for the races and the odds for the three contenders can be all over the place but as in any meaningful sample, the 3-5 contenders will win a lot more often than the 10-1 contenders. Other than that, I'm just trying to take advantage of the statistic that one of the three will win at an 80% clip without trying to further rank or separate the three contenders.

There are a lot of creative minds on this board so I'd like to solicit some win wagering strategies to best take advantage of this statistic as-is.

Thanks in advance for your input.

Dale

[cj]

Bet the longest price...or the second longest, but not the shortest.

[jcrabboy]

Hi dvlander:

With an 80% hit rate in your top three contenders I would bet them all if odds are 7/2 or better. Drop the low odds horse if under 7/2. If all your contenders are at low odds pass the race.

A lot of guys don't like betting more than one horse in a race. I think if you show a profit and it is compatible with your betting personality then it is a perfectly acceptable strategy. A big bonus would be a significant decrease in long runouts, which can be psychologically devastating.

Jimmie

[Dave Schwartz]

With an 80% hit rate in 3 horses you could simply bet all three. Period.

[Cratos]

Suppose you could isolate three contenders in a large group of races. 80% of the time, one of your contenders wins. The field sizes for the races and the odds for the three contenders can be all over the place but as in any meaningful sample, the 3-5 contenders will win a lot more often than the 10-1 contenders. Other than that, I'm just trying to take advantage of the statistic that one of the three will win at an 80% clip without trying to further rank or separate the three contenders.

There are a lot of creative minds on this board so I'd like to solicit some win wagering strategies to best take advantage of this statistic as-is.

Thanks in advance for your input.

Dale

I believe the way to view your problem would be to understand that your probability of randomly picking a contender from a large group will be 3/X. Therefore a group the size of the KY Derby field it would be 3/20 or about 6.60-1. However if the field shrinks to a small size of 5 your odds increase to a 3-5 chance of selecting your contenders or about 60%.

But, there is the second part which you say one of the 3 will win 80% of the time. Therefore you would have in a 20 horse field a 12% chance of winning, but the horse that you bet should be not less than 8.40-1 for a long run breakeven.

If you have a 5 horse field you will have a 48% chance of winning and your betting odds should be not less than 2-1 to breakeven.

In summary as your field size increases you need to increase your odds proportional to the breakeven point and conversely as the size becomes smaller you can accept lower odds and still breakeven

[dvlander]

Dave, are you saying that assuming the 80% hit rate is valid, betting all three horses each time is superior long-term to any odds-dependent exclusion?

Are you suggesting flat-betting the three regardless of odds?

Thanks,

Dale

[dvlander]

But, there is the second part which you say one of the 3 will win 80% of the time. Therefore you would have in a 20 horse field a 12% chance of winning, but the horse that you bet should be not less than 8.40-1 for a long run breakeven.

If you have a 5 horse field you will have a 48% chance of winning and your betting odds should be not less than 2-1 to breakeven.

In summary as your field size increases you need to increase your odds proportional to the breakeven point and conversely as the size becomes smaller you can accept lower odds and still breakeven

Cratos, your stats are logical. In retrospect, I would guess that 5-horse fields would offer little value. However, I am quite sure that an 80% overall hit rate would be a bit higher for 5-horse fields and quite a bit lower for 12+ fields. I posted my question today before breaking the data down by field size although I know the average field size in the entire sample was 8.1.

I appreciate your input.

Dale

[rrbauer]

With an 80% hit rate in 3 horses you could simply bet all three. Period.

Agree with Dave subject to the following:

You need composite-odds of 4/5 to break-even (at an 80% hit rate). So anytime the composite odds of the three horses is greater than 4/5 you have a 3-horse dutch play.

[podonne]

Agree with Dave subject to the following:

You need composite-odds of 4/5 to break-even (at an 80% hit rate). So anytime the composite odds of the three horses is greater than 4/5 you have a 3-horse dutch play.

Sorry, just wanted to understand this. By composite you mean additive, like if you added the pools of the three contenders together and treated them as one imaginary super-horse, that super-horse should have odds higher than 4/5?

I've been looking for an easy way to intrinsicly tell wherther a situation is "dutchable" (without monte-carloing the hell out of the entire thing)

[rrbauer]

Sorry, just wanted to understand this. By composite you mean additive, like if you added the pools of the three contenders together and treated them as one imaginary super-horse, that super-horse should have odds higher than 4/5?

I've been looking for an easy way to intrinsicly tell wherther a situation is "dutchable" (without monte-carloing the hell out of the entire thing)

You can do it that way, or you can normalize the odds to 1, and see if the three horses you like have less than 80% of the total. For example, let's say there's one with .10 of the total and another with .20 and another with .30. That's .60 of the total which is less than .80, so (roughly) you would bet 1/6 of your bet on the .10 horse; 2/6 on the .20 horse and 3/6 on the .30 horse to complete the dutch.





(As an aside, monte carlo has nothing to do with it)

[Dave Schwartz]

Dave, are you saying that assuming the 80% hit rate is valid, betting all three horses each time is superior long-term to any odds-dependent exclusion?

Are you suggesting flat-betting the three regardless of odds?

I am.

Dick's suggestion about dutching is a good one as well, but ... the nature of dutching is that you will now be wagering more money on the lower-priced horses and less on the higher ones.

Logically, the 1st and 2nd choices will almost always be part of your contenders. What will likely make your system profitable are the slightly higher-priced horses that make it into your top 3.

As an example, suppose we considered only horses that went off at 8/1. One would assume that they would, as a group, lose around 20%. But you would have two classes of 8/1 horses: those that are conetnders and those that aren't. One would (logically) assume that the contender group would significantly outperform the non-contenders.



All this being said, I will offer you another alternative: Monty Hall.
http://math.ucsd.edu/~crypto/Monty/monty.html

I have a methodology that will concentrate two-thirds of the winners into one-third of the horses. In other words, it is a "final decision" approach that works. If you really are capable of getting 80% winners in three contenders, then a realistic expectation when Monty is applied is about 50% winners with one horse.


On the 23rd of August we will hold a pay-for workshop on this precise topic.
http://www.horsestreet.com/ubb/Forum32/HTML/000053.html

While this seminar is aimed at HSH users, it is not mandatory that you have HSH to apply the principles. Read what one user had to say after attending this workshop back in March:

http://www.horsestreet.com/ubb/Forum4/HTML/001822.html

Here is the thread of the original in-person workshop we did in January. If you scroll towards the bottom you will see the results achieved by the actual attendees. Note that at that workshop I played live - every race - 30 races, 1 horse per race, and cashed 17 tickets. The other players had similar results.

Read about it here:
http://www.horsestreet.com/ubb/Forum4/HTML/001722.html



Regards,
Dave Schwartz

[Cratos]

With an 80% hit rate in 3 horses you could simply bet all three. Period.


As I understand the problem, it is not a 80% hit rate of a given race, it is a 80% hit rate of a subset of a given race and the probability of that subset having a contender is 3/X where “X” is the size of the field. This problem becomes a conditional probability or Bayesian analysis.

[Dave Schwartz]

What I heard was that his contenders win 80% of their races. Is that incorrect?

[Cratos]

What I heard was that his contenders win 80% of their races. Is that incorrect?

That is correct, but he is selecting contenders at 3/X which says the larger the field, the harder it is to select contenders which is typically true in horseracing. Also a small field is virtually a shoo in with a contender.

[juanepstein]

i play three horses in one race alot. but i look at horses physically and make evaluations. if you can cap horses physically and evaluate them in the paddock or post parade then your gonna do much better in your handicapping.

#1- i look at the favorite and see if he is false and i look at his appearance and see if looks out of shape and beatable.

#2- if the first favorite looks bad and is beatable i go to the 2nd favorite. if its the same with the 2nd favorite then i go to the 3rd favorite.

#3- once ive found which favorite looks top notch. i look at his odds pay and divide it by 2. that will tell me how many horse i can bet with that favorite.

so if the 2nd favorite looks impressive on the track and he is say 3/1- 2 goes in to 8 four times. so i can bet the 2nd favorite with 3 other horses to beat the 1st favorite. if the 2nd favorite comes in ive pushed if any of my other horses come in ive made a profit. usually in american racing i go three horses unless its a big field and/or i see potential bombers then i go for it.

this is the way i love to play australia with their 16 to 20 horse fields.

[asH]

wait for a solid single then play a multiple pk 3 for 3X1X3 = $9, ...

if you are certain of your percentages 3X3X3 = $27 at a 51% hit rate, you cant lose...over time

to beter increase your odds:
you need to know which kind of races produce your higher win percentages and center your pk (X) around that kind/type of race ie MSW, allowance NW1, 6F, etc...

[rrbauer]

As I understand the problem, it is not a 80% hit rate of a given race, it is a 80% hit rate of a subset of a given race and the probability of that subset having a contender is 3/X where “X” is the size of the field. This problem becomes a conditional probability or Bayesian analysis.

If he was selecting the horses at random, the probability would be as you state, 3/X. However, he is not selecting them at random. The probability of his selection method producing a winner in various size fields has not been established. He simply states that field sizes and winners' odds varies and that low-priced horses win more frequently than high-priced horses.

[cutchemist42]

In one of James Quinns books, he actuqlly says shooting for an 80% hit rate on your top 2/3/4 contenders is something to strive for. I remember the book saying the idea came from elsewhere.

[davew]

With an 80% hit rate in 3 horses you could simply bet all three. Period.

I would flat bet, but only when the lowest odds of the 3 was 2/1

I am not sure how many races this will leave you to bet ... but gets discouraging paying $6 and returning $3.80 (times your bet level).

[BCOURTNEY]

Calculate your historical edge for every odds range for whatever process you used to produce the contenders. Bet all the horses you have an edge on as a function of your current bankroll. e.g. if you have a 3% edge, place a 3% bet of bankroll on the horse. In any given race you might have no horses to play, might have 3 or might have 6. A good selection process will produce a variable amount of edge wagers in any given race. From a 30,000 foot view a 80% hit rate in your top 3 is great, but how many times a month, week, day or hour does that come up?

[formula_2002]

Suppose you could isolate three contenders in a large group of races. 80% of the time, one of your contenders wins. The field sizes for the races and the odds for the three contenders can be all over the place but as in any meaningful sample, the 3-5 contenders will win a lot more often than the 10-1 contenders. Other than that, I'm just trying to take advantage of the statistic that one of the three will win at an 80% clip without trying to further rank or separate the three contenders.

There are a lot of creative minds on this board so I'd like to solicit some win wagering strategies to best take advantage of this statistic as-is.

Thanks in advance for your input.

Dale

for most of us, winning 80 percent of time will still coat us the price of the track take out. :pout:

[therover]

I would go back over the course of a month or two and look back at the winners of those races. Determine if your handicapping method is picking winners more in route races vs sprints, dirt vs turf, etc.

I would then look at how to possibly isolate one of the 3. Are there any angles you are using to handicap or strictly via a pace/speed/race set up ? I am an angle guy. I may not have a high win percentage but the average price of my winners makes up for it. There are times when my contender looks brutal but wins and pays $40-$50.

Look for positive jockey switches. Recent work outs. Trainer changes, distance changes. Anything to get an edge.

Lastly....check out the prices on those winners. Are they in a specific range. Are you hitting a lot of low odds favorites or are your winners paying $8-$12.

Betting horses in a specific PTO range is always tough. Me....I can't bet on 2-1 or less. I pass the race most times. Same on the other end...30-1 is the top end. I look for the sweet spot between 4-1 through 12-1. I have found that most of my winners that I pick using my specific angle pay win at 5-1 through 8-1. They are my prime bets. I will double up on my bet if the horses odds are in that range. Also, the method I use works about 5 times better on route races so I pretty much stick to them now.

3 contenders with an 80% win rate is great. I would try to do some additional analysis to see if there is additional data or a methodology to isolate one of them. I know that's tough work but it may be worth it. You may end up with a sweet spot play method.

[tshinder]

All this assumes that you know what the closing odds are - which is almost impossible these days.

[mistergee]

with the stats given of 3 bets per race and 80% winners wouldnt anything over an average payoff of $7.50 be profit?