Small Money Wagering Plan

[RaceBookJoe]

Depending on pool size of the track, i use a different wagering plan. With smaller tracks, here is one i pulled out of the closet that i used a long long time ago especially when i played the dogs. I call it a 3 Column Round Robin, its not an original plan and maybe some of you have tried it?

You set up 3 seperate columns and a starting bet for each column. After a loss in any column you either add 1 unit, or you can up the bet by 20%..i add 1 unit for easy math. Fun little plan that allows you to cash a lot of tickets as long as your horse hits the board. You can start this with $2, but my phone account only allows $5 minimum bets so thats what i use for now. Any winning column goes back to the base bet. You set up the column like this

Bet Column A Column B Column C
1 $5 W $5 P $5 S if you lose this bet, then
2 $10 P $10 S $10 W your horse places
3 $5 S $5 W $15 P etc etc

columns keep rotating W P S W P S bets change depending on which column hits. I either circle or highlight winning bets so i adjust the bet correctly. Wont make you rich, but a fun little plan. Its just one of a few different plans i use depending on track size/bankroll size i start with. With decent selections and a few decent prices you should avg 20-30% roi..just a guess but thats what a few recent test runs have shown.

[davew]

progression systems will always lose eventually if you have a negative ROI

[traynor]

progression systems will always lose eventually if you have a negative ROI

That is "kind of" true, but not absolute. The people declaring such usually do so in relation to casino games, most notably the Labouchere (cancellation) method of wagering at roulette. A lot of people have made a lot of money with such systems before they run into the "choppy" sequences that cause a long series of escalated wagers similar to the "due column" system used by a number of very knowledgeable (and very profitable) wagering syndicates.

While "true" (with caveats) that progression systems "invariably lose" in casinos, it is much less so when wagering on pari-mutuel races. In casinos, the normal payoff is the amount of the wager. The odds available in pari-mutuel wagering create a leverage unavailable in casinos. The only seriously deficient situation is when the wagers get so big that the odds are diminished below what they "should" be. Other than that, "computer simulations" that declare progression betting as invariably losing are interesting, but there are exceptions.

In the real world, I think it is probably more difficult to make a profit on a small positive ROI than it is to make a profit using a progression system on an even or slightly negative (from past races) ROI. In general, the winning wagers are always slightly bigger than the losing wagers.

[Some_One]

It's true when when the long run is an infinite number of events in the series. Of course in the short run, deviations can occur, and I think that's what those people who tout their alternate money management systems rely on.

[Robert Fischer]

At least this is aptly named. ;)

[RaceBookJoe]

At least this is aptly named. ;)

Yeah it wont make a fortune, but it isnt designed to. I also should have said that you need to have a good in-the-money hit rate, but it allowed me to play every race with not much drawdown at all. I was a kid when i was doing this, just recently started thinking about it again and was wondering if anyone had tried it. Cant remember the last test i did where it went even as high as the 5th level. Again, its not my main wagering plan..just one that i found to be fun and can give a lot of action for those who like that. Off the top of my head, the last 10 race test/card i did was total bets of $250 returned $342 with no big prices coming in. Fun for me to play dogs,harness this way since i cant handicap them :)

[davew]

what do you do for low payers?

say your horse paid $5.4 $3.2 $2.1
and your bet was

S$15 W$15 P$15 after 2 losses (W$5P$5S$5, P$10S$10W$10)

do all columns clear?

[RaceBookJoe]

what do you do for low payers?

say your horse paid $5.4 $3.2 $2.1
and your bet was

S$15 W$15 P$15 after 2 losses (W$5P$5S$5, P$10S$10W$10)

do all columns clear?

All columns clear back to base bet in that case because of cashing in all 3 slots. I have never tried going down 1 level. In your example your next bets would be $5W $5P $5S because the columns keeping rotating to the next type bet wpswps etc. Now if that horse in your example had placed...next bets would be $5W $20P $5S because you had a winner in column A and C. Just something fun for me when i am trying new selection processes or "system"..i like to use real money even if its small amounts. Originally i started with $2 base bets back in the 70's when i learned it, then adjusted it starting with $5 base bet with a 20% increase ( 5,8,11 etc ), then for ease of math just add 1 unit with a return to base on a winning column. My main play had always been a receding due column with a takeout which worked great ( used a variation for craps/baccarat also), but now i mainly use a 1unit/3unit win/place bet and a few exotic plays.

[JackS]

A simple method using the same premisis except the progression of bet amounts is to bet less as you lose and a large bet following a win. i.e.- A Place bet method which better than average handicappers should be able to hit 50% of the time. Sugessted rules- Toss all horses with a morning line above 10-1 and assure yourself before hand that the place pool guarantees a price of $4.00 or better. Example- Race 1, $20 to Place. If lost, $15 to Place in Race 2. If either of these races are hit, the bet doubles and regardless of the outcome, the sequence starts again. The difference between this and the Martingale is the reduction on the second bet and returning to the basic sequence. In effect you are playing on an educated assumption (your records) that you can pick a winner 50% of the time..
Note- I do not play this method myself and would urge anyone interested to scan thier own past performance or begin to develop a large enough database before proceeding.

[BlueShoe]

My main play had always been a receding due column with a takeout which worked great ( used a variation for craps/baccarat also)
Sounds like that old very good plan from LGH? Speaking of Holloway, regarding wagering plans for negative expectation casino games, he insisted that actual real life patterns and table results behaved differently than random computer generated simulations, thus making such games beatable with skillfully used methods.

[RaceBookJoe]

Sounds like that old very good plan from LGH? Speaking of Holloway, regarding wagering plans for negative expectation casino games, he insisted that actual real life patterns and table results behaved differently than random computer generated simulations, thus making such games beatable with skillfully used methods.

Bingo..its the old LGH plan, and also correct on his patterns which he had 5. I would use a different wagering plan for horses than i would when i played craps/baccarat. Both worked well because the "takeout" grinded out a profit either on each winner or after a certain amount of hands/rolls/spins etc depending on which plan you used. I still use both but also incorporate a few other things.

[Handiman]

There's software I wrote called Bet Magic. I made it available in a thread back some time ago. Don't remember where but PA might know where it is at if you are interested. It essentially does what you are talking about. It's good for a string of about 25 bets I believe.

Handi:)

[PaceAdvantage]

There's software I wrote called Bet Magic. I made it available in a thread back some time ago. Don't remember where but PA might know where it is at if you are interested. It essentially does what you are talking about. It's good for a string of about 25 bets I believe.

Handi:)Here you go:

http://www.paceadvantage.com/forum/showthread.php?postid=1168951#post1168951

Downloaded 587 times...not bad... :ThmbUp:

[DigitalDownsJoe]

I wonder how the system would do for greyhounds? I have a really high ITM strike rate with them..Wish I could pick the horses half as good. The problem with horses I think is I dont lay off races like I should and I hate betting favorites. Im always looking for the bud longshot! :D I never tracked my ROI though. Dogs, I generally make money when I go to bet, but few big days. Horses i lose prob 5 out of 6 times i go to play horses(i never mix) but i have some really good days were i make good profit(when its not chalk city)

[DigitalDownsJoe]

Do you have rules as far as the range of odds you will bet/not bet to be profitable?

[traynor]

I wonder how the system would do for greyhounds? I have a really high ITM strike rate with them..Wish I could pick the horses half as good. The problem with horses I think is I dont lay off races like I should and I hate betting favorites. Im always looking for the bud longshot! :D I never tracked my ROI though. Dogs, I generally make money when I go to bet, but few big days. Horses i lose prob 5 out of 6 times i go to play horses(i never mix) but i have some really good days were i make good profit(when its not chalk city)

Probably a lot better than it does on thoroughbreds. Especially if you stick to A and AA. Because of the high win percentage and consistent performance, we have used a similar wagering plan for a couple of years with good results. The only problem is that your win bets tend to tweak the win pools if you bet too much. Great for steady, modest wagers to generate a solid bread-and-butter profit. Almost like free money. Especially on Florida tracks.

[myhorse1]

from traynor #3.
i'm sorry i entered this in the wrong place.


[In the real world, I think it is probably more difficult to make a profit on a small positive ROI than it is to make a profit using a progression system on an even or slightly negative (from past races) ROI. In general, the winning wagers are always slightly bigger than the losing wagers.[/QUOTE

i am not quite sure how using a progression system makes things better.

what would qualify as slightly negative --would 5 to 7% qualify as slightly negative.

even though the winning wagers are larger if you are making these wagers with an overall negative roi it seems to me that the whole set of wagers has a negative roi and that by using a progression system you would have the same
negative roi with a lot more pain at the end.

thank you for your many thoughtful and enlightening posts.

[traynor]

from traynor #3.
i'm sorry i entered this in the wrong place.


[In the real world, I think it is probably more difficult to make a profit on a small positive ROI than it is to make a profit using a progression system on an even or slightly negative (from past races) ROI. In general, the winning wagers are always slightly bigger than the losing wagers.[/QUOTE

i am not quite sure how using a progression system makes things better.

what would qualify as slightly negative --would 5 to 7% qualify as slightly negative.

even though the winning wagers are larger if you are making these wagers with an overall negative roi it seems to me that the whole set of wagers has a negative roi and that by using a progression system you would have the same
negative roi with a lot more pain at the end.

thank you for your many thoughtful and enlightening posts.

Your assumptions are correct in regard to "even money" wagering in casinos. If you have a negative expectation at roulette, craps, or whatever, you will invariably lose over time. In pari-mutuel racing, the return is flexible--you can get more than even money in return for your wagers.

The definition of of "negative expectation" is historical--it doesn't become negative until it happens. It could just as easily be positive. Lots of people lose lots of money believing they have a "statistical edge" betting horses. Other people win lots of money by understanding that horse races are unique events that can be leveraged profitably.

That is not a flippant comment. A negative expectation (in pari-mutuel horse races) does not exist until it happens--a sequence of bets that return less than they cost. Unfortunately for all the hardcore bean counters of the world, making money on the (assumed) positive expectation based on past race results is every bit as elusive as the negative expectation.

[myhorse1]

thank you for your thoughtful reply.

i'll have to think about how or if i can apply this.

[porkchop]

The single most important part of any progressive wagering plan is to know the probabilities of your personal handicapping methods . If you are a player who prefers higher payouts the percentge of winning bets is lower than someone who backs favorites. Without keeping records of all wagers chances are you will fall into a streak of unexpected runouts that will more than likely destroy your bankroll. Knowing your own capabilities will help keep any due column method in line with your goal of showing a profit.

[TrifectaMike]

The definition of of "negative expectation" is historical--it doesn't become negative until it happens. It could just as easily be positive. Lots of people lose lots of money believing they have a "statistical edge" betting horses. Other people win lots of money by understanding that horse races are unique events that can be leveraged profitably.

Agreed a horse race is a singular, non-reproducibe unique event. However, in order to determine expectation (negative or positive) for a race, you are required to determine the probability of each horse's probability of winning for that race.

How do you propose one determines these probabilities?

Mike (Dr Beav)

[traynor]

The single most important part of any progressive wagering plan is to know the probabilities of your personal handicapping methods . If you are a player who prefers higher payouts the percentge of winning bets is lower than someone who backs favorites. Without keeping records of all wagers chances are you will fall into a streak of unexpected runouts that will more than likely destroy your bankroll. Knowing your own capabilities will help keep any due column method in line with your goal of showing a profit.

That is where the rubber meets the road. "Knowing the probabilities of your personal handicapping methods" implies that some static value exists that can be so labeled, and the future depended on to replicate the past events from which it is derived. Nice and neat in theory. Not so nice and kind of sloppy in the real world.

[traynor]

Agreed a horse race is a singular, non-reproducibe unique event. However, in order to determine expectation (negative or positive) for a race, you are required to determine the probability of each horse's probability of winning for that race.

How do you propose one determines these probabilities?

Mike (Dr Beav)

To quote Howard Sartin, "You do the best you can with what you have to work with."

After reading a number of your very illuminating posts, my current field of study is implementation of Bayesian inference (in my own handicapping). It is a fascinating field, and one I hope to effectively utilize in the near future. At this point, it is still limited to study.

[TrifectaMike]

To quote Howard Sartin, "You do the best you can with what you have to work with."

After reading a number of your very illuminating posts, my current field of study is implementation of Bayesian inference (in my own handicapping). It is a fascinating field, and one I hope to effectively utilize in the near future. At this point, it is still limited to study.
I assume you were able to filter the noise generated in those threads to get at the heart of the matter.

Mike (Dr Beav)

[traynor]

I assume you were able to filter the noise generated in those threads to get at the heart of the matter.

Mike (Dr Beav)

Yes. Textual deconstruction was one of my most interesting activities in graduate school. It is quite useful.

[myhorse1]

That is not a flippant comment. A negative expectation (in pari-mutuel horse races) does not exist until it happens--a sequence of bets that return less than they cost. Unfortunately for all the hardcore bean counters of the world, making money on the (assumed) positive expectation based on past race results is every bit as elusive as the negative expectation.

1. am i understanding you to say that the sequence of bets could take place over more than one race--whereas the first race might negative combining with race 2 or 3 could be positive.

2. do you normally bet more than one horse in race?

3. are your bets odds related e.g. at 4/5 no bet but 8/5 you have a bet(or the reverse at 8/5 no bet and 4/5 you bet (depending on the wisdom of the bettors).

4.do you set an odds line that the bet has to be above?

5.do you have a minimum odds for a bet(e.g.2/1)?

any answers would be appreciated.

[traynor]

1. am i understanding you to say that the sequence of bets could take place over more than one race--whereas the first race might negative combining with race 2 or 3 could be positive.

2. do you normally bet more than one horse in race?

3. are your bets odds related e.g. at 4/5 no bet but 8/5 you have a bet(or the reverse at 8/5 no bet and 4/5 you bet (depending on the wisdom of the bettors).

4.do you set an odds line that the bet has to be above?

5.do you have a minimum odds for a bet(e.g.2/1)?

any answers would be appreciated.

I do this for a living, not as a hobby. That provides me with a very pragmatic view of theories, ideas, and concepts. Basically, if they don't work, I don't use them. Unfortunately, that includes a great deal of what many handicappers believe is Truth With A Big T. I care not one bit that someone may have spent 10 or 20 years chasing rainbows, programming nonsense into his or her computer, or possibly even picked a few winners--if I can't use it to make money, it is (to me) worthless.

I do what I do because it works for me. It may not work for anyone else, because my motivation is strictly profit. That is, I have no interest whatsoever in the sport, grandeur, excitement, or whatever else motivates people (other than profit) to bet on horse races. That makes my view of racing quite unlike most other people's view of racing.

Without going into a long, boring discussion of statistics vs reality, the short form is that what many consider a "positive expectation" or "negative expectation" is derived from a small clump of races. That small clump of races is uncritically assumed to be a perfect replica of all other races. Whether the small clump of races is 10, 100, or several thousand is irrelevant (with all due respect to those who believe they can "prove mathematically" that their sample is a perfect representation of future events). The key point is that the distribution of events in the sample only suggests direction--it does not guaranty that future events will be distributed precisely as they are in the sample.

To answer your questions more specifically, I ignore the morning line, and--with the exception of relatively large wagers at one or two tracks--I pay no attention to post time odds. Specifically, most of the wagers I make are based on my own estimation of a horse's probability of winning, and the "available odds" are based on the average mutuel that my records indicate is "normal" for that specific configuration of factors, in that specific type of race, at that specific track, currently.

Sounds silly, huh? Crazy, even? How can anyone possibly hope to make a profit betting on horse races if he or she ignores the odds that the horse will pay if it wins? Flip that coin. How can anyone possibly believe that the "law of averages" or "statistical edges" apply to the results of the specific race she or he is betting on today? For the average bettor--including many fairly serious professionals--the results of the whole gaggle of bets they make in a month, or a year, are "statistically insignificant."

The idea that "if you only bet when you have a positive expectation, profit is almost guaranteed" seems to make sense. Especially if a bettor believes that a "positive expectation" magically controls which horse will win the next race. It doesn't matter if the horse you bet pays less than the average mutuel you think it should pay in that specific race, as long as you don't have to give back the extra money won when it pays more.

I don't especially care about the results of one specific race, or even of 10 races. I don't agonize over trivia. I am in for the long haul, and that means I have to pay attention to making money, not the results or available odds on any one or two or ten races.

[myhorse1]

Specifically, most of the wagers I make are based on my own estimation of a horse's probability of winning, and the "available odds" are based on the average mutuel that my records indicate is "normal" for that specific configuration of factors, in that specific type of race, at that specific track, currently.
.
does that mean if your estimate of the horse's probability of winning is 30% and your records indicate that the current configuration of factors would be even money (not the tote board) you would pass the race -or am i misunderstanding.

[traynor]

does that mean if your estimate of the horse's probability of winning is 30% and your records indicate that the current configuration of factors would be even money (not the tote board) you would pass the race -or am i misunderstanding.

That is correct--I make wagers according to my figures, not the tote board. If my records indicate a 33% win probability for a particular configuration, the average mutuel (corrected for outliers and anomalies) for that configuration has to be appreciably better than 2-1 to be worth betting.

The upside is that I get to bypass the whole bleak scenario of a dip just before the off that "prevents me from betting because of unacceptably low odds" that is reversed by someone dumping a bundle on another horse at the last second and causing what had been "unacceptably low odds" to jump into the profitable range--when it is too late to bet. The lack of knowledge of what the final odds will be (in any individual race) tends to make obsessing over those odds an exercise in futility. ("Fudging" is not a strategy.)

A point that many fail to grasp is that they are attempting to apply the results of a large sample to a few individual events. They might do far better if they focused more on long-term results rather than on individual races and the ups and downs of the tote board--the very things the "positive expectation" of their models is intended to smooth out.

[myhorse1]

traynor

if not to presumptuous would you say that the average odds on the horses you bet are greater or less than 3 or 4 to one.

[Capper Al]

I play shots with few exceptions.

[traynor]

traynor

if not to presumptuous would you say that the average odds on the horses you bet are greater or less than 3 or 4 to one.

That depends on the win%. The application I use for modeling "corrects for outliers" in mutuel prices, to avoid the high hopes and expectations created by a few large mutuels mixed in with a lot of chalk.

Briefly, the app makes a sorted list of all the mutuel prices for each pattern, tosses the highest quarter and lowest quarter, averages the mutuels in the middle two quarters, then sets a value 1.5 times that average as high end. The mutuel prices are adjusted, and any mutuel price higher than that is entered (for modeling purposes) as 1.5 times "the mean of the interquartile range" of values. That corrected figure is the "average mutuel."

Multiplying the win percentage by the average mutuel/2 indicates the potential return. Example: 0.33 (7.50/2) = 1.2375. That is the expected ROI for betting that pattern.

In the real world, I ignore any ROI less than 1.10 at any except a handful of major tracks. The reason is simple. In most cases, I did not bet on the races used for modeling. Profit margins of less than 10% tend to evaporate when bets are made in the real world, rather than on paper or in regression studies. Lots of people lose lots of money chasing trivial "positive expectations" that are based on races they didn't bet on.

How is it possible to get average mutuels of $7.50 combined with a 33% strike rate? Look beyond the obvious, and avoid doing the same old-fashioned, worn out, over-the-hill "race analysis strategies" everyone else hammers to a pulp every racing day.

In short, study the experts, but think for yourself. It is your money you are betting.

[traynor]

traynor

if not to presumptuous would you say that the average odds on the horses you bet are greater or less than 3 or 4 to one.

In the example I used, the 33% strike rate would require an average mutuel higher than 2-1 ($6.00, which would be break even).

[myhorse1]

traynor

thank you for the detail in your replies-- you have given me a great deal to think about.

[traynor]

traynor

thank you for the detail in your replies-- you have given me a great deal to think about.

Thinking is good. If more bettors did it more often, they would lose less, win more, and life would be more difficult for people like me.

[myhorse1]

In short, study the experts, but think for yourself. It is your money you are betting.

which one or two experts would you recommend studying.

[thaskalos]

I do this for a living, not as a hobby. That provides me with a very pragmatic view of theories, ideas, and concepts. Basically, if they don't work, I don't use them. Unfortunately, that includes a great deal of what many handicappers believe is Truth With A Big T. I care not one bit that someone may have spent 10 or 20 years chasing rainbows, programming nonsense into his or her computer, or possibly even picked a few winners--if I can't use it to make money, it is (to me) worthless.

I do what I do because it works for me. It may not work for anyone else, because my motivation is strictly profit. That is, I have no interest whatsoever in the sport, grandeur, excitement, or whatever else motivates people (other than profit) to bet on horse races. That makes my view of racing quite unlike most other people's view of racing.

Without going into a long, boring discussion of statistics vs reality, the short form is that what many consider a "positive expectation" or "negative expectation" is derived from a small clump of races. That small clump of races is uncritically assumed to be a perfect replica of all other races. Whether the small clump of races is 10, 100, or several thousand is irrelevant (with all due respect to those who believe they can "prove mathematically" that their sample is a perfect representation of future events). The key point is that the distribution of events in the sample only suggests direction--it does not guaranty that future events will be distributed precisely as they are in the sample.

To answer your questions more specifically, I ignore the morning line, and--with the exception of relatively large wagers at one or two tracks--I pay no attention to post time odds. Specifically, most of the wagers I make are based on my own estimation of a horse's probability of winning, and the "available odds" are based on the average mutuel that my records indicate is "normal" for that specific configuration of factors, in that specific type of race, at that specific track, currently.

Sounds silly, huh? Crazy, even? How can anyone possibly hope to make a profit betting on horse races if he or she ignores the odds that the horse will pay if it wins? Flip that coin. How can anyone possibly believe that the "law of averages" or "statistical edges" apply to the results of the specific race she or he is betting on today? For the average bettor--including many fairly serious professionals--the results of the whole gaggle of bets they make in a month, or a year, are "statistically insignificant."

The idea that "if you only bet when you have a positive expectation, profit is almost guaranteed" seems to make sense. Especially if a bettor believes that a "positive expectation" magically controls which horse will win the next race. It doesn't matter if the horse you bet pays less than the average mutuel you think it should pay in that specific race, as long as you don't have to give back the extra money won when it pays more.

I don't especially care about the results of one specific race, or even of 10 races. I don't agonize over trivia. I am in for the long haul, and that means I have to pay attention to making money, not the results or available odds on any one or two or ten races.
I like what you say here, Traynor...and I would like to ask you a question or two, in an attempt to understand you better:

You say that you play for the long haul...not for the short term. What does "long haul" mean to you? As a professional bettor, would it alarm you to find yourself in the red after a period of, say...three months?

Also...do you find that your profits remain fairly consistent over time...or do they fluctuate wildly year to year?

Thanks in advance for your answers.

[traynor]

I like what you say here, Traynor...and I would like to ask you a question or two, in an attempt to understand you better:

You say that you play for the long haul...not for the short term. What does "long haul" mean to you? As a professional bettor, would it alarm you to find yourself in the red after a period of, say...three months?

Also...do you find that your profits remain fairly consistent over time...or do they fluctuate wildly year to year?

Thanks in advance for your answers.

If I were in the red for two consecutive weeks, I would stop betting (and have a number of times in the past) until I understood why, and corrected it.

Of all things, the truncating of windfall mutuels down to 1.5 times the mean of the interquartile range (described several times on various threads) has done more to stabilize return than any other single factor. The inclusion of anomalous outliers in a (relatively) small sample seriously distorts the potential return. By truncating outlier mutuels, I am often pleasantly surprised by returns greater than anticipated, and rarely disappointed by returns less than expected.

The last part is tough to answer, because "it depends." I bet based on models that are continually changed and upgraded, primarily according to the wagering patterns of the other bettors. A particular strategy (pattern) may generate a decent return for weeks, months, and (rarely) for years. Then enough of the other bettors either see the same thing or something similar, bet more on those particular pattern selections, and the profits drop to unacceptably low rates.

The app I use "adapts to change." That means that if a particular pattern generates X ROI, the app monitors the return from that pattern in the results. If it detects an "unacceptable drop" (determined by lots of detailed range functions), it "sets a flag." It sounds slick, but it is fairly simple code. If the decline in ROI is detected, the app adds a warning--oncreen and in the printout--indicating such. The greater the drop, the sterner the warning. I used to do all that by hand, and I am much happier to have coded the functions that enable the app do it automatically.

"Consistent earnings"--for the reasons mentioned above, are not what some believe them to be. The returns are only consistent in that I only bet on productive patterns. Other than that, it varies, but no wild swings because I don't chase losing patterns in the hope that they will become productive again. I stop betting those patterns until they do.

I do not believe that any single strategy or single pattern is indefinitely profitable (other than in regression models that seem to indicate such). Bettors are a lot smarter than most give them credit for being. It doesn't take long for them to catch up, see the pattern, and bet it.

The simplest way to express it is that I bet a lot of races, but the reason I bet on each one of those races individually is continually changing.

[traynor]

which one or two experts would you recommend studying.

For starters:

http://www.nationalfiretraining.net/userfiles/GreatBasin/COMM/hammond_hidden_traps_in_decision_making.pdf

http://www.2shared.com/document/mNWOmA-I/Decision_Traps_-_The_Ten_Barri.html

http://psiexp.ss.uci.edu/research/teaching/Tversky_Kahneman_1974.pdf

For specifics about horse racing, this forum is probably your best resource. Ainslie, Beyer, Quirin, and Quinn all wrote interesting stuff, little of which is applicable to current racing circumstances, and should be regarded primarily as interesting background research rather than carved-in-stone Truth With a Big T. Brohamer's pace book is good, as well as the book on the same topic written by a group of Sartin's PIRCO members. Again, interesting, but outdated.

If you attend races personally, "inspection handicapping" can give you a lot of good information that others lack. The body language videos by Bonnie Ledbetter and Trillis Parker are both great. With all due respect to Joe Takach, I don't buy the idea that the quick glimpse of the horses provided by televised images is sufficient--watching the horses live is so much more useful that anything less seems foolish.

[TrifectaMike]

For starters:

http://psiexp.ss.uci.edu/research/teaching/Tversky_Kahneman_1974.pdf



Traynor, thanks for posting the links. The paper above is written very simply and has direct applications and implications in horse racing.

Everyone should read it.

Mike (Dr Beav)

[myhorse1]

For starters:

http://www.nationalfiretraining.net/userfiles/GreatBasin/COMM/hammond_hidden_traps_in_decision_making.pdf

http://www.2shared.com/document/mNWOmA-I/Decision_Traps_-_The_Ten_Barri.html

http://psiexp.ss.uci.edu/research/teaching/Tversky_Kahneman_1974.pdf

For specifics about horse racing, this forum is probably your best resource. Ainslie, Beyer, Quirin, and Quinn all wrote interesting stuff, little of which is applicable to current racing circumstances, and should be regarded primarily as interesting background research rather than carved-in-stone Truth With a Big T. Brohamer's pace book is good, as well as the book on the same topic written by a group of Sartin's PIRCO members. Again, interesting, but outdated.

If you attend races personally, "inspection handicapping" can give you a lot of good information that others lack. The body language videos by Bonnie Ledbetter and Trillis Parker are both great. With all due respect to Joe Takach, I don't buy the idea that the quick glimpse of the horses provided by televised images is sufficient--watching the horses live is so much more useful that anything less seems foolish.
traynor

thank you for the book suggestions and the links.
i have read the books you mentioned with the exception of the book by the PIRCO members.
i have downloaded the first and third links you provided and I am looking forward to see what they offer.
the middle link does not seem to be working-- when i downloaded it--i get 1978 pages but they are all blank-- maybe THAT'S the message or it may be because i'm using a mac.
anyway thanks again for your time.

[traynor]

traynor

thank you for the book suggestions and the links.
i have read the books you mentioned with the exception of the book by the PIRCO members.
i have downloaded the first and third links you provided and I am looking forward to see what they offer.
the middle link does not seem to be working-- when i downloaded it--i get 1978 pages but they are all blank-- maybe THAT'S the message or it may be because i'm using a mac.
anyway thanks again for your time.

Synopsis:
http://www.psclipper.com/DecisionTraps.asp

The book is available on Amazon for less than $2, and well worth reading. It has an especially interesting discussion of multiple factors as criteria for decisions about horse races, based on a study of public handicappers. Of note is the observation that the more information given to the handicappers, the more confidence they had in their predictions, but the less accurate the predictions became. Specifically, there was an inverse correlation between the "level of confidence" in the predictions, and the accuracy of those predictions in the real world.

The book is "simply written" but that should not dissuade readers. I first encountered it as a textbook in a graduate university class in executive decision-making processes. Like Musashi's Book of Five Rings, there is more there than there seems to be on the surface.

[classhandicapper]

The last part is tough to answer, because "it depends." I bet based on models that are continually changed and upgraded, primarily according to the wagering patterns of the other bettors. A particular strategy (pattern) may generate a decent return for weeks, months, and (rarely) for years. Then enough of the other bettors either see the same thing or something similar, bet more on those particular pattern selections, and the profits drop to unacceptably low rates.

The app I use "adapts to change."

This is the very reason some people try to make odds lines to begin with.

If you assume that the betting patterns of the players change, (and they do) then any specific angle that is profitable now may not be profitable a few weeks from now (or whenever).

However, if you "value" that angle correctly you don't care what the other players are doing. The odds dictate the decision. The way to overcome late odds changes and potential error is to insist on a large margin of safety. In other words, if you think horse "X is worth 2-1, don't bet him if he's 5-2. You insist on 7-2. Then if he gets bet down late or you aren't valuing him properly, you are still in safe territory.

I use the "this is nuts" method.

Without actually making a formal line, I have a mental feel for a range of odds that I would consider reasonable for a specific contender given what I know and what I am unsure of. Only when I look at a horse's odds and say "THIS IS NUTS" do I make a wager.

I also use a method similar to yours (less formal), but only for trainer patterns, because the betting trends change quickly as the sample size grows when it comes to trainers.

Some people are looking for large sample positive ROI trainer patterns. IMO, that's the past. Horses like that will typically get bet down to unprofitability as the sample size increases even though they were profitable in the past.

I am looking for a very small sample trainer stat that I have reason to believe is the beginning of a legitimate long term trainer pattern. Sometimes I misread things, but I am right often enough. Then as everyone else discovers it or finally believes it because the sample size is larger, I'm gone.

[thaskalos]

I am looking for a very small sample trainer stat that I have reason to believe is the beginning of a legitimate long term trainer pattern. Sometimes I misread things, but I am right often enough. Then as everyone else discovers it or finally believes it because the sample size is larger, I'm gone.

I admit that I have never felt comfortable with the idea of betting on a horse strictly because of the trainer's stats. Even when I encounter a new claim by a trainer who wins 40% of the time and shows a solid profit in these situations...I would rather pass the race than wager on this favorable stat. I know...money is money...but trainer handicapping does not sit well with my overall handicapping philosophy.

My question is...what do you do when your handicapping opinion of the ability of the horse flatly contradicts the trainer's winning pattern? What I mean is...let's say that you have reason to believe that the trainer is at the beginning of a legitimate long term training pattern with sprinters stretching out...but in this particular route, his horse does not look capable of pulling it off...for a variety of fundamental handicapping reasons.

Do you put the handicapping fundamentals aside, and wager on the horse with no reservations...or does your handicapping opinion of the horse overshadow the winning training pattern that you have uncovered?

I guess what I am asking is...I understand how handicappers who focus on trainer patterns approach this game...but how do more traditional handicappers -- who focus primarily on form, class, speed and pace -- handle these winning trainer patterns?

[classhandicapper]

I admit that I have never felt comfortable with the idea of betting on a horse strictly because of the trainer's stats. Even when I encounter a new claim by a trainer who wins 40% of the time and shows a solid profit in these situations...I would rather pass the race than wager on this favorable stat. I know...money is money...but trainer handicapping does not sit well with my overall handicapping philosophy.

My question is...what do you do when your handicapping opinion of the ability of the horse flatly contradicts the trainer's winning pattern? What I mean is...let's say that you have reason to believe that the trainer is at the beginning of a legitimate long term training pattern with sprinters stretching out...but in this particular route, his horse does not look capable of pulling it off...for a variety of fundamental handicapping reasons.

Do you put the handicapping fundamentals aside, and wager on the horse with no reservations...or does your handicapping opinion of the horse overshadow the winning training pattern that you have uncovered?

I guess what I am asking is...I understand how handicappers who focus on trainer patterns approach this game...but how do more traditional handicappers -- who focus primarily on form, class, speed and pace -- handle these winning trainer patterns?

I had the same problem for what seemed like an eternity. My ultimate solution to many of these problems won't be satisfactory to most players. I pass a lot of races.

I put a lot of value on "certainty". The more certain I am that I'm right about a race, the more willing I am to bet. Otherwise I pass the race.

I rarely key a trainer pattern blindly if the horse looks bad on paper. There are exceptions, but not many. I will "use" a horse like that if I'm playing a multi-race bet or hate the favorite so much I'm willing to bet more than one horse to win.

For me to "key" a trainer pattern horse I have to think:

1. The horse has reasonably recent back races that are good enough to win. In that case I am banking on the horse jumping back up to his "A" race.

2. The horse figures best already. In that case on am banking on the horse being way more likely to fire his "A" race than the typical horse.

[RaceBookJoe]

Good question Thaskalos, i guess for me I would say that i use it either as a tie breaker (esp if odds are generous ), or if i have no "handicapping" factor to go on. Most likely makes it a non-playable single race ( i try to avoid out and out guessing ), but as part of a horizontal bet I might take a leap of faith. Sometimes for me its a gut feel, ie: narrow race down to 2 horses "close in ability" but one has weak connections, the other has trainer with high win% for this type of situation and/or the trainer/jockey combo shows positive roi. Kind of hard just to say, i do this here and that there.

[eurocapper]

From what I've seen the top trainers seem to generally have lots of works. Might be an overlooked angle, the amount of works instead of their speed. (Of course they are going to be overbet, but maybe some up and coming trainer who takes his horses seriously like they seem to do.) Also they seem to be careful about where it's going to run. It has seemed to me one can use such a horse in exotics even if it doesn't look like the best of the race.

[JackS]

I will play favorites but I always attempt to play something else first. Many times this is impossible as the favorite appears to be unbeatable. Most of these races will be obvious to the public and offer unplayable odds.
My stratagy in such races is to forego exactas and use this horse to begin a P3 and assuming that one or both winners of the next two races will not be the top favorite. This allows (most times) for a glorified DD so to speak.
I don't expect a huge payout but I do have hope for something very respectable using the "unbeatable" in the first leg.

[traynor]

This is the very reason some people try to make odds lines to begin with.

If you assume that the betting patterns of the players change, (and they do) then any specific angle that is profitable now may not be profitable a few weeks from now (or whenever).

However, if you "value" that angle correctly you don't care what the other players are doing. The odds dictate the decision. The way to overcome late odds changes and potential error is to insist on a large margin of safety. In other words, if you think horse "X is worth 2-1, don't bet him if he's 5-2. You insist on 7-2. Then if he gets bet down late or you aren't valuing him properly, you are still in safe territory.

I use the "this is nuts" method.

Without actually making a formal line, I have a mental feel for a range of odds that I would consider reasonable for a specific contender given what I know and what I am unsure of. Only when I look at a horse's odds and say "THIS IS NUTS" do I make a wager.

I also use a method similar to yours (less formal), but only for trainer patterns, because the betting trends change quickly as the sample size grows when it comes to trainers.

Some people are looking for large sample positive ROI trainer patterns. IMO, that's the past. Horses like that will typically get bet down to unprofitability as the sample size increases even though they were profitable in the past.

I am looking for a very small sample trainer stat that I have reason to believe is the beginning of a legitimate long term trainer pattern. Sometimes I misread things, but I am right often enough. Then as everyone else discovers it or finally believes it because the sample size is larger, I'm gone.

That makes a lot of very good sense. The one point I would differ on is the odds. That is not theoretical, but logistical--the way I bet would drive me crazy if I had to look at the post time odds of every entry before deciding whether or not to bet it, and I definitely do not trust a computer app to do my decision making in that regard.

In particular, leveraging a trend before everyone else sees that it is a trend can produce some very worthwhile results. Or dumping a trend while everyone else is betting it with both hands.

[myhorse1]

That makes a lot of very good sense. The one point I would differ on is the odds. That is not theoretical, but logistical--the way I bet would drive me crazy if I had to look at the post time odds of every entry before deciding whether or not to bet it, and I definitely do not trust a computer app to do my decision making in that regard.



do many of the horse you bet END UP being the morning line or post time favorites where it seems to me you would have the most concern about not getting your overlaid odds

[traynor]

do many of the horse you bet END UP being the morning line or post time favorites where it seems to me you would have the most concern about not getting your overlaid odds

Not at all. The "odds" that I use for betting are based on "expected average mutuel," which in turn is based on current betting trends at specific tracks. If a specific pattern of attributes has a profitable expected average mutuel, I bet it. Because the information I use is based on current races, if a trend changes, I know it fairly soon after the change. That is, I am not betting today based on "positive ROI" that is contingent on one or two anomalous mutuels from last month or three months back. It took quite awhile to get it right, but the information I use for betting is relatively current. If it falls off into the no-profit area, I don't bet it.

"Overlaid odds" are not as much a concern for me in any individual race, or set of races, as they would be for most bettors. Some pay less, some pay more, but the average usually exceeds the amount I calculate as the expected average mutuel because I truncate unusually high mutuels.

Again, it is the underlying concept of playing multiple races--rather than agonizing over tote board changes for a single race--that forms the basis for the algorithms I use in analyzing races. The idea of ignoring post time odds (for all practical purposes) might seem unusual, but it is an intrinsic part of how I wager.

I don't get anything from betting except profit. No thrill, no rush, no excitement, no grand feeling of having outslickered the rubes in the grandstand. It is a business, much like any other business. Profit is good. Loss is bad.

Specifically, the psychic reward that some derive from a momentary belief that they are "in control" because they picked a winner is worth much less to me than the money.