Steve Fierro a quote

[karlskorner]

"The majority of handicappers (notice I didn't say investors) spend far too much time on unimportant items they have been led to believe are important"

I read it, reread it and concluded he's right. How much information has been gathered and used because the handicapper "has been led to believe that it is important".

www.netcapper.com/TrackTracts.htm

Steve Fierro's article "Financial goals for a Horse Investing business". My goals are daily, his is for a longer period of time.

[blind squirrel]

GREAT BOOK...finally motivated me to keep "precise'
records.

[hurrikane]

Steve is a great guy.

He plays the horses for a living. He doesn't write books for a living.

It pays to listen to what he says

[Dave Schwartz]

"Steve is a great guy.

He plays the horses for a living. He doesn't write books for a living."

Steve IS a great guy and a very sharp handicapper. And his book started me on a path of self-discovery that has led to profitibility. But, by his own admission, he does not make a "living" at the races.


Dave Schwartz

[JimG]

If I only listened and learned from people who made their living at the races over my lifetime, I would not have done any listening and learning over the last 30 years.

Jim

[Dave Schwartz]

JimG,

Agreed.

I hope it did not sound like I was discrediting him in any way.


Dave

[karlskorner]

This is the 2nd time in the past 2 months that you have mentioned your handicapping turned around. You stated "self-discovery has led to profitibilty". that sounds mysterious. Was it the book, a chapter in the book or just his general theories ?

[cj]

Karl,

Odd you would consistently quote Fierro. He is definitely a "value" guy. You have repeatedly said you don't believe in that. What gives?

[karlskorner]

I don't own the book, never read it. What I posted came from Gordon Pines Netcapper.com. I have posted several things that Netcapper.com has posted by Fiorrie. I like what he says. To the best of my memory nothing he wrote or was quoted mentioned "value". I did mention that I would browse through the book at Barnes & Nobles and you were kind enough to mention that B&N doesn't sell it. That kills that. If "value" is the basis of his book, the I repeat his quote "handicappers spend too much time on unimportant items they have been led to believe are important".

[Dave Schwartz]

Karl,

First, it is good to hear from you. Not enough posts lately, old friend.

Fierro... I liked his book. Got two things from it to add to my arsenal: His odds templates and what he calls "filtering."

The odds templates caused me to look at the odds line(s) generated by HSH. First thing I did was try to shoehorn HSH's probabilities into his templates. Realized very quickly that our odds lines was better than the templates (as they should be if you think about it).

But more importantly it became clear that "value" in the sanse of overlays was the only way to play. (I realize that many people discovered this long ago... it is the correct application that has eluded most people.)

So, I set about trying to buiild a new science... the science of value.

The work has gone very well but is not quite ready for prime time, although several of our HSH users have reported astonishing results just from changing their orientation from selection to value.

By the way, the reason most odds line/probability based systems don't work is because they do not modify the probabilities WITH the tote board.


The second thing about Steve's book: Filtering. In a nutshell, he said that his experience was that when HE had a horse at 3/2 or less the race was not profitable (whether he played the low horse or the "overlay(s)").

This screamed out to me that if we look at a subset of our bets there are certainly some races/plays that we could EASILY deem "pass races." Subtracting those races from the sample makes the rest of the sample playable.

Okay, so to summarize what I REALLY took from it:


1. After we have done the handicapping, we then must find a way to properly "exploit" the race. Obviously some races are less "exploitable" than others and some not at all. These become the pass races.

2. We must search our betting records to locate areas where our race assessments are simply wrong. Amazingly, you cannot do this just from looking at who we wager on. It starts with the success of our contenders.

So, these two points... and I went down a path of self-improvement. The performance got better and better... on paper. Time for some real betting and, for the first time EVER (for me, anyway) the real bets went as good (actually better) than the tests!

When I went to Saratoga it was an opportunity to do it publicly. I cut my bet size back to around $30 per race to reduce the pressure and make it fun instead of just work. To say it went well would be a gross understatement. (I am sure that some of the 'Toga guys can verify this.)

BTW, anyone that wants to read the story of my personal Saratoga adventure may do so at this link:

http://www.horsestreet.com/ubb/Forum4/HTML/000868.html



So, what are we doing to share this? First, I am collaborating with Red Knave (new poster on this board) to develop the ultimate bet analyzing software. (I don't want to say much but I will say that there will not be any manual entry of anything for starters.)

Second, I am working on a great new way to deliver this "science of exploitation" to the handicapping public. (Think cyber-cast.)



Regards,
Dave Schwartz

[BillW]

Originally posted by Dave Schwartz

By the way, the reason most odds line/probability based systems don't work is because they do not modify the probabilities WITH the tote board.

Regards,
Dave Schwartz

Dave,

Are you talking the sum of the probabilities (i.e. 80% line vs. a 70% line) or the individual probabilities independent of the total?

Bill

[karlskorner]

Thank you for the reply. I printed it out, took it to the track and over a bowl of Chowder and a Boston Creme, a computer person friend explained it to me.

On your site about the Toga trip you stated that your profit was $1022.00 with $700.00 invested. Originally you stated you intended to cut back to $30.00 a play. Apparently Fiorrie has caused you to become an exotics player since you won 11-28 exotic plays, plus occasional win/place wagers, where I always understood you played to win only.

Assuming you had approximately 35 plays, including your occasional win/place wagers. since your big win play payed $106.00, what would your profit have been if you had played $30.00 to win per race over the approximste 35 plays ? The $106.00 horse would have paid $1590.00 with a $30.00 punch and leave you 34 more plays which you probably would have won at least a 25%. thereof.

[Dave Schwartz]

Bill,

The way I see it, the sum of the probabilities must equal 100% (including non-contenders).

Of course, "line" versus "probabilites" are really interchangeable, but probabilitiies are easier to adjust.

The point is that two horses lined at 2/1, one goes off at 30/1 and the other goes off at 6/5, their lines need to change. The 6/5 prbably needs to come down and the 30/1 needs to go up (from the original 2/1, of course).


Dave

[Dave Schwartz]

Karl,

I did not actually calculate the individual bets... I was having too good a time. I just know that I started with a $50 voucher and never looked back. <G>

My standard bet was:

1. Box all my contenders for $2 (4 horses in a 8+ field, 3 horses in 7 or less)
2. Play any horse that went off at 9/1 or higher to w/p for $2.

Thus, in an 8-horse field I'd bet $24 on exactas and (potentially) up to 4 (in reality about 1.5) horses w/p. In a 7-or-less field I'd bet only $12 on exactas.

In addition, I desgnated a handful of races "crush races" where I not only had huge overlay exactas (without ever looking at the payoffs, I might add) and a high hit rate. In these I bet $3 units instead of $2.

The first exacta I hit was a "crush" race, as was the last one. In both of those I got the "short end," with low-priced horses ruining my hit. That last exacta was a heart breaker; closest photo I have ever lost. Still went $160 but the other way it would have been $310 (plus the $60 winner instead of "just" the place).

I think I actually wagered more like $800, by the way.


Dave

[karlskorner]

Dave;

I realize you went there for the enjoyment, but my point is you "were" a win only player, you changed horses mid-stream (pun) and became an exotic player. Your decision to cut back to $30.00 a play would have been my decision also at a new track, by your 3rd play (if I read your post right) you had a possible $1590.00 win payoff with a $30.00 punch, you would have been ahead over $1500.00 At that point I personally would have folded the tent and bought everybody a beer and go sit in the sun and watch the ladies, but that is my own method of play. Somewhere in the next few days you had to have a couple if not more "win" plays. To quote you elsewhere "handicapping is easy, it's the winning thats hard". But, hey, you had a good time, made some nice money and proved your program works. Like I use to say when in business "You gotta look good to be good" Looking good Dave, looking good.

Karl

[JustMissed]

What the heck is this $30 win bet business about? I thought guys at your level bet $100 to $300 minimum for just the win.

Maybe you are just trying to hideout from "Big Brother"!!!!!!

Karl-Glad to see you are back posting-I have missed your input.

JustMissed
;)

[hurrikane]

That is interesting Dave. I met Steve a couple of times and was certain that he did make his living handicappint.

I know he has other things he does but I was under the impression that the majority of his income came from horses..or at least did.

Just goes to show you how it is easy to misunderstand what people.

One thing is certain. He is a great guy.

[VetScratch]

I think there are several value-play systems that have shown a profit year after year.

The only way I can understand why this phenomenon exists has to do with the psychology of most horseplayers.

Picking a winner feeds our egos whereas letting other players dictate our bets is not as exhilarating as watching our picks confirm our handicapping prowess.

All successful value-play systems entail a rather rigid methodology for passing/screening races, qualifying overlay contenders, and responding to what other players do with a prescribed wagering strategy.

Following such a system becomes such a tedious and mechanical exercise that few players are willing to persist and strictly comply with the methodology... in other words, while everyone gives lip service to a profit motive, the majority of players are not willing to grind out a profit with a tedious system that lacks the passion, thrills, and ego-involvement associated with traditional handicapping. Value-play is a passive exercise that takes the assertive intellectual "fun" out of winning.

This is why existing value-play systems have not been diluted by publication to the extent we might expect.

However, within the next year, I think we will see commercial value-play systems integrated with on-line toteboards and wagering services such that users will be relieved of all the work that makes such systems tedious. When this happens, widespread availability should kill the goose that laid the golden egg. It will be interesting to see what happens when the first system that allows numerous users to go golfing while their PC plays the races comes into the marketplace.

[karlskorner]

Another post I am going to print out and explained to me over chowder and a Boston Creme

[Pace Cap'n]

From an article posted not long ago about online tip sheets:

The service – which employs Bruno De Julio, Steve Fierro and six other veteran horse evaluators – is one of the most successful online race-tip guides, with about 8,000 clients paying $12 or more a day for cards from Del Mar, California Fairs, Arlington and Saratoga, as well as coverage of venues in Australia and Europe.

That's right at $10,000 per day. Not too shabby.

Full story:



www.signonsandiego.com/news/business/20030802-9999_1b2tipsheet.html

[VetScratch]

Pace Cap'n,
The service – which employs Bruno De Julio, Steve Fierro and six other veteran horse evaluators...I see no harm in mentioning that PA member Lyn Konkoly is one of the featured handicappers on Bernsen's web site. She is a trainer's daughter, a friend of Purple Power, and seems to be doing quite well:

Lyn Konkoly
Lyn's Saturday Ohio 7&7 card keeps on going!!! 9/20= 10 winners on the card (71%), 8 winners on top (57%); 3 exactas; 4 trifectas; 1 superfecta; 1 daily double; 3 pick 3s; and the pick 4!! Best plays= Race 5 TRIFECTA paid a whopping $3,020.60! She also had the $22.00 winner on the card in Race 6 (her 2nd choice), and she gave the pick 4 with ALL top picks! I say, "You go girl!"

BTW, the pricing is per card per day with some special packages. Single cards seem to range between $6 and $12 per track, which is in the same general range as BRIS and other tout sheet providers.

[Valuist]

Vet Scratch-

You defined "value capping" as more of a passive activity. I disagree; as you can see from my moniker, I'm a big believer in value capping. In true value handicapping, one makes up their own odds then looks for discrepancies. One must be anticipative; I like to check out the public selectors to get a feel for how the public will bet a race. Much of my betting is done blind; I often will look for what I feel are the best plays, bet them, and get out without seeing the odds or the race. Let's face it; when we do cash a ticket, we have to get our money's worth; cashing underlays does us no good in the long run, which is the only run that matters.

[alysheba88]

There is nothing tedious about a value line and furthermore there it is not something that can be replicated by a computer. It is completely subjective. When I make an odds line it is my personal line/opinion. Not something a computer spits out.

It is a great feeling when you have a horse like Wonder Again at 5-2 on your line and see the horse go off at 6-1.

To me blindly betting computer or data base picks is tedious. The human element is what makes this game the best. Matching wits against each other. Creating a value line is part of that process.

[VetScratch]

Valuist,
Agreed, value-play is not purely passive for the few who start from scratch each day and formulate their own odds-lines. However, many private and most commericial value-play systems rely on software to automatically calculate odds-lines.

In any case, after the odds-lines are calculated, the process does tend to be passive because bets are dictated by others, whether those others are public oddsmakers (as you indicated) or the betting public (by monitoring real-time toteboard odds). All value-play decisions boil down to avoiding underlays, finding overlays, and judging whether to pass or play each race.

Fierro's 3/2 angle illustrates the merit of monitoring real-time odds to screen out races that are not worth playing when results from many such races are studied. Not all value players may draw the line at 3/2 because a different value (e.g., 2/1) may be optimal at a specific track. However, I am not a spokesperson for Fierro, so I recommend that doubters buy his book and decide for themselves whether they agree or not.

[Lefty]

Alysheba, there can be a lot of subjectivety in a value line but don't you think it's possible for a computer to come up with an objective line that works? That's my only hope as my intuition just plain stinks.

[lynette]

Vet-

Thanks for the plug!

Lyn

[VetScratch]

Originally posted by lynette
Vet-

Thanks for the plug!

Lyn You are welcome. I imagine Fierro and the rest of the boys feel flattered to be featured with you. :)

[lynette]

Vet-
You are very kind :) I believe the feeling probably goes both ways between all of us on that site.
Lyn

btw, this is a direct quote from David Bernsen, with his permission: "That article (was) way misquoted. (There are) 8,000 subscribers ranging from free (the vast vast majority) to the ability to purchase from $5 to $20 or more. "

[dbernsen]

The author for the San Diego Union misquoted me (considerably).

While TrackExperts.com is quite popular and we have over 8,000 subscribers, that entails our Free Daily Special available at http://trackexperts.com

However, 8,000 people are not paying $12 a day. We have a variety of services and some subscribe on monthly, quarter or annual basis, and others maybe purchase a couple of times a year.

By the way, this is a nice forum.

David Bernsen

Originally posted by Pace Cap'n
From an article posted not long ago about online tip sheets:



That's right at $10,000 per day. Not too shabby.

Full story:



www.signonsandiego.com/news/business/20030802-9999_1b2tipsheet.html

[Pace Cap'n]

My apologies for passing along mis-information.

I should have known, by virtue of the medium, to allow for the hyperbole factor.

Money aside, with 8,000 subscribers you must be doing something right!

[GR1@HTR]

Question re odds line adding up to 100.

Run this drill. I've done it a few times.

1) Get a Results Chart
2) Calculate the prob. of the horses to win based on off odds.
3) Add the total of the prob. of all the horses in the race.

You should find that the total probability of all the horses to win a race does not add up to 100. It is about 118 to 125.

Hence, shouldn't we base our odds line/probabilties based on a total of 118 to 125 vs 100?


BTW, I also have met the all mightly fierro. If I am correct he earns his imcome from horse racing. Some of that might consist of betting, consulting for racebooks and providing public selections.

[VetScratch]

Originally posted by GR1@HTR
Question re odds line adding up to 100.

Run this drill. I've done it a few times.

1) Get a Results Chart
2) Calculate the prob. of the horses to win based on off odds.
3) Add the total of the prob. of all the horses in the race.

You should find that the total probability of all the horses to win a race does not add up to 100. It is about 118 to 125.

Hence, shouldn't we base our odds line/probabilties based on a total of 118 to 125 vs 100?


BTW, I also have met the all mightly fierro. If I am correct he earns his imcome from horse racing. Some of that might consist of betting, consulting for racebooks and providing public selections. Are you considering takeout and breakage when you calculate probabilities from final odds?

A workable formula for dime breakage is: (1 - takeout)/(odds +1.05).

So, if takeout is 17% and final odds are 4.80 then: (1 - .17)/(4.80 +1.05) = .14188 = 14.2% = probability.

Now you should get 100% when you sum all probabilities.

[Bob Harris]

<BTW, I also have met the all mightly fierro. If I am correct he earns his imcome from horse racing. Some of that might consist of betting, consulting for racebooks and providing public selections.>


You are correct. Steve resigned from the 9-5 world a few years back and started playing full-time using the same templates which are outlined in his book. As his success became known, other horse racing related opportunities came his way...ESPN contacted him to do a weekly radio show, Today's Racing Digest contacted him to publish his betting line, the Reno Hilton asked him to hold seminars for their guests, etc.

With each opportunity, he always asked himself the same thing: "Will this opportunity still allow me to bet daily or will I just end up as another talking head who rarely bets?"

I know of several lucrative things which he's turned down simply because it would keep him away from the windows. The other stuff is nice but his first love will always be handicapping a race and backing up his decision with cold, hard cash.

[Dave Schwartz]

GR1,

You are making this more difficult than it needs to be.

A horse at 2/1 IS a 33% horse.

A horse at 9/1 IS a 10% horse.

A field must have 100%.

The goal is not to predict the odds. The goal is to determine the horses' chances. (At least that is my goal.)


Dave

PS: Your assessment of Steve Fierro is right on. And, again, I want to make it clear that I have no intention of disparaging him or his work in any way. He is a competant and profitable handicapper and (I have heard) a very good teacher.

[VetScratch]

Originally posted by Dave Schwartz
GR1,

You are making this more difficult than it needs to be.

A horse at 2/1 IS a 33% horse.

A horse at 9/1 IS a 10% horse.

A field must have 100%.

The goal is not to predict the odds. The goal is to determine the horses' chances. (At least that is my goal.)


Dave

PS: Your assessment of Steve Fierro is right on. And, again, I want to make it clear that I have no intention of disparaging him or his work in any way. He is a competant and profitable handicapper and (I have heard) a very good teacher.
Dave,
As you might guess, I disagree with you because the purpose of formulating a value odds-line is to compare these odds to track odds. Thus, since it is so simple, you might as well conform to the parimutuel system when you compute your odds line. Otherwise, you end up comparing raisins to grapes.

If you give a horse a 10% probability of winning, and the betting public also gives the horse a 10% probability, your odds "should" be the same as the parimutuel odds.

For 10% probability with 17% takeout and dime breakage, the parimutuel odds will be 7.3/1 while your method gives odds of 9/1. On the toteboard, 7.3/1 is 7/1 and not 9/1.

If accurate pace and speed figures are desirable, why slack off when calculating value odds-lines. Are you suggesting 6th-grade math is above our heads?

IMHO, a value odds-line that doesn't add up to 100% probability just as the parimutuel odds add up to 100% probability, is not worth bothering with, especially when it's so easy to do this.

[GameTheory]

I thought the purpose was to compare your probabilities (adding up to 100%) with the *prices* given by the toteboard (adds up to more than 100%).

If I make my line so it adds up to 118% or so, then my 10% horses will be equal with the toteboard's 10% horses, and it would then appear I have a fair (break-even) bet, when of course that is not true. The public has to underestimate the horse's chances merely to create a break-even bet.

Actually, I think we all agree, but VS misunderstood Dave in her zeal to disagree with him...

[Dave Schwartz]

GT,

Nice call.


Sincerely,
Dave

<G>

[VetScratch]

Originally posted by GameTheory
I thought the purpose was to compare your probabilities (adding up to 100%) with the *prices* given by the toteboard (adds up to more than 100%).

If I make my line so it adds up to 118% or so, then my 10% horses will be equal with the toteboard's 10% horses, and it would then appear I have a fair (break-even) bet, when of course that is not true. The public has to underestimate the horse's chances merely to create a break-even bet.

Actually, I think we all agree, but VS misunderstood Dave in her zeal to disagree with him... Hi Game,

<g> Actually, you've aroused my zeal. <g>

At the core of value handicapping is the assertion that the handicapper can outperform the betting public with respect to formulating accurate probabilities for a field of starters.

I have no doubt that Dave (HSH) can accomplish this and has been able to validate his superior degree of accuracy against a database that reflects actual results and the probabilities produced by the betting public.

On the surface, it would seem that all you need to do is arm users with natural odds (10% = 9/1) because these odds represent the breakeven point. How simple! If your 10% horse is going off at 18/1, your proven history of accuracy should alert any user to an overlay opportunity. Of course, there are other considerations, like dutching or picking one play when a race offers more than one overlay opportunity. Here again, by going back to database analysis, you can evaluate these situations and formulate recommended strategies to cope with these circumstances.

From what Dave posted about working on his new "science of value," I assume that he is on the verge of "prime time" readiness with a value system backed by such thorough research that users should soon be banking profits and doing back-over flips.

What can possibly go wrong? How much simpler does it have to get? Dave formulates highly accurate probabilities, and it is patently obvious that 9/1 is the breakeven fulcrum for a horse with a 10% probability of winning. Again, what can go wrong?

What can go wrong is what has bedeviled others who have been there and done that. The problem with value systems is that you can lead a horse to water but you can't make him drink. The least experienced horseplayers will be happiest, but more experienced horseplayers will often falter because old habits are hard to break. You simply can't overestimate how hard it is for players who have been picking winners for twenty years to convert themselves into value players.

The result can be very frustrating. Some users (mostly novices) are delighted, while others (with the most experience) will challenge the credibility of the system. For reasons most vendors are familiar with, it is not a simple task to get to the root of some problems via customer service emails (or even phone calls).

On the one hand, you have a system that seems simple to implement and is verifiably profitable. On the other hand, you have a body of users where the experienced majority often cannot reproduce the results enjoyed by novices (i.e., the same results that are still profitable when you track them via your database).

Because I had done such work the past, the publisher of a value system asked me to review and possibly revise the documentation in an effort to make it easier to use. However, what already existed was so straightforward that I was baffled as to how anyone could misunderstand the documentation. The end result was that I was not very helpful.

About a year later, a minor enhancement was added to the system. Much to everyone's surprise, user satisfaction instantly improved for no quantifiable reason. The nature of the minor enhancement was to publish a true parimutuel odds line. This minor added feature made a big difference.

It turned out that users fared much better when they were given parimutuel odds to compare with toteboard odds. If anyone is interested, I will gladly post examples that illustrate why users were having problems and why parimutuel odds made a big difference. :)

IMHO, Dave's users would also benefit if he started them out with a parimutuel odds line because I suggest he will want to add this feature sooner or later. It really helps traditional players convert successfully to value players.

[VetScratch]

Heck, here are simplfied examples of what I meant in my previous post.

Definitions:
Track Takeout = 17%
Natural Odds = breakeven odds for win probabilities
Toteboard = live toteboard odds
Parimutuel = parimutuel odds for win probabilities

TRADITIONAL Value-Play Projections:
Horse Win Natural Tote-
Prob Odds board
Amble 33.3% 2/1 8/5
Biker 20.0% 4/1 3/0
Caper 14.3% 6/1 9/2
... ... ... ...
Deano 10.0% 9/1 6/1
Early 10.0% 9/1 14/1
... ... ... ...
No matter how clearly you explain your system purpose:
Some rushed players will falsely interpret Amble as an underlay.
Some will also overestimate the degree to which Biker, Caper, and Deano may be (or actually are) underlays.
All players will see you gave Deano and Early the same probabilities and natural odds.
Reverting to old habits, some players will bypass Early and bet Deano, thinking "smart money" bet Deano down.
Others will bet Early as an overlay but waste money by also betting Deano and/or Caper as an insurance hedge.

ENHANCED Value Play Projections:
Horse Win Pari- Natural Tote-
Prob mutuel Odds board
Amble 33.3% 1.60 2/1 8/5
Biker 20.0% 3.30 4/1 3/0
Caper 14.3% 5.00 6/1 9/2
... ... ... ... ...
Deano 10.0% 7.30 9/1 6/1
Early 10.0% 7.30 9/1 14/1
... ... ... ... ...
By adding parimutuel odds, it becomes much easier to see that:
Amble is not an underlay.
Biker, Caper, and Deano are either not underlays or not significant underlays.
Given live toteboard odds of 6/1, Deano's true odds may in fact be very close to 7.30/1 (e.g., 6.90/1).
Early is an obvious single play as an overlay opportunity.

(In both examples, other horses not necessary to illustrate anything are not shown.)

[alysheba88]

If anyone has read Fierro's book and used his templates you will find that his contender odds line adds up to 79-81%. Which provides a nice cushion.

[VetScratch]

My examples: I knew something looked amiss, now I see what.
I mistakenly used 13% takeout for Amble, Biker, and Caper. Their 17% takeout parimutuel odds should have been 1.40, 3.10, and 4.80. For Deano and Early I did use 17% takeout. I guess such things shouldn't be attempted on a calculator... at least by me.:)

Horse Win Pari- Natural Tote-
Prob mutuel Odds board
Amble 33.3% 1.40 2/1 8/5
Biker 20.0% 3.10 4/1 3/0
Caper 14.3% 4.80 6/1 9/2
... ... ... ... ...
Deano 10.0% 7.30 9/1 6/1
Early 10.0% 7.30 9/1 14/1
... ... ... ... ...

[JustMissed]

Best I can tell Vetscratch is correct in her math and I also believe correct in her conclusion that the odds you(the player) calculate should be calculated in the same manner as the tote odds in order to avoid comparing apples and oranges.

Sometimes you guys just crack me up; you worry yourself silly about the accuracy of pars, beaten lengths, distance of run ups, pace figs, speed figs, etc. and then when it gets to the most important decision of the day, whether or not to make that bet, you purport to make that decision based on an inaccurate odds line.

Just my take on the subject,

JustMissed

[hurrikane]

Write or wrong I believe you are trying to kill flys with a cannon here.

It is very seldom that the tote does not change 17-18% after the gate opens. All the precision in the world is not going to get you any more value.

[JustMissed]

Originally posted by hurrikane@HTR
Write or wrong I believe you are trying to kill flys with a cannon here.

It is very seldom that the tote does not change 17-18% after the gate opens. All the precision in the world is not going to get you any more value.

I totally agree with you. My point was that if you are going to promote or tout a system, at least promote a system that is calculated correctly.

I used a "value system" sometime ago and gave it up when I found I could do just as well by labeling my contenders in projected winning order, A B C D. If I felt strongly that A could win, I bet, if not, I passed the race or looked for an exotic play.

You have no idea how much money I have lost because I bet the B horse because it had more "value" than the A horse who would finish lengths ahead.

Not saying that value is not valid but I won't be convinced till someone can show me their betting records to prove it works.

JustMissed

[GameTheory]

I'm trying to figure out why Amble is "not an underlay" when anything less than 2/1 is an underlay.

I frankly see no purpose for calculating the winning probabilities for each horse and not have them add up to 100%. Having "pari-mutuel" odds doesn't seem to be any help whatsoever -- what extra information am I supposed to be getting? (It seems like you really want to be comparing pool percentages to probabilities instead of working in odds, which is more confusing. I see percentages as probabilities and odds as prices.)

I didn't realize we were talking about what the best thing to do for a "value system seller" was rather than what the best thing to do for a value player making their own line is -- someone making their own line obviously can tell what is an overlay & what is an underlay based on that line. I don't need to adjust the line merely to play a psychological trick on myself so I will be less confused -- I know what I'm doing.

The other important point Dave made is that you can't just set your line and leave it be, unless you build-in very large margins of error (ala Fierro, apparently) -- you must adjust it somewhat along with the actual toteboard odds as they flucuate.

So there are two things I would use the toteboard for:

-- Firstly, determine the % of the pool bet on each horse, and use that to adjust my odds line accordingly (according to some function, which could be a simple linear weighting, or the percentages could be used as a variable in a more complex non-linear adjustment with a neural network or something [that's what I do personally]). If this whole system is not automated/computer-assisted, you'll probably be skipping this step, in which case you might want to have your line add up to only 80% or so as suggested, and you also must be on the lookout for "false overlays" (a 15/1 on-the-toteboard horse that you have at 4/1 is probably a bad bet, for example).

-- My line thus adjusted (which is just a way of making the line more accurate and less biased), I look at the actual toteboard odds (or decimal representations of same for greater precision) to see what each horse is paying (or to guess what it will be paying, actually).

Then, bet the overlays over a certain threshold of margin.

[VetScratch]

Originally posted by hurrikane@HTR
Write or wrong I believe you are trying to kill flys with a cannon here.

It is very seldom that the tote does not change 17-18% after the gate opens. All the precision in the world is not going to get you any more value. I tend to think you need the cannon for the hard part, which is handicapping to produce accurate probabilities. After that, only a few trivial lines of code are needed to translate probabilities into both natural and parimutuel odds. :)

[VetScratch]

Game,
I'm trying to figure out why Amble is "not an underlay" when anything less than 2/1 is an underlay.You can certainly look at it that way. I use the following simple definitions:
Underlay - overbet by the public.
Overlay - underbet by the public.

If 33% is the probability yardstick, odds equivalent to 33% of the win pool indicate whether the public has overbet or underbet what you expect. The parimutuel odds for 33% are 1.40/1.

I have always looked at it that way because it's easier to see where the parimutuel toteboard is going versus your parimutuel odds. When it comes time to bet or pass, you still have your breakeven points as a reference.

Without belaboring the point, I believe an attractive overlay has more credibility if it is mirrored by several small underlays as opposed to one or two big underlays (barn buzz, juice rumors, plunging owners, etc.). To me, this seems easier to grasp if you always compare toteboard odds against parimutuel odds until it becomes time to make your bet or pass decision.

[VetScratch]

Game,
Originally posted by GameTheory
I frankly see no purpose for calculating the winning probabilities for each horse and not have them add up to 100%. Having "pari-mutuel" odds doesn't seem to be any help whatsoever -- what extra information am I supposed to be getting? (It seems like you really want to be comparing pool percentages to probabilities instead of working in odds, which is more confusing. I see percentages as probabilities and odds as prices.)The probabilities should always add up to 100% or something is wrong. Why would anyone do otherwise?

For any set of probabilities adding up to 100%, you simply need to calculate two odds representations for each probability.

Natural Odds = (1/Probability)-1
Parimutuel Odds = ((1-Takeout)/Probability)-1.05

If a probability established by the betting public is 33%, then that horse comprises 33% of the win pool. Unfortunately, the parimutuel odds will not be 2/1.

Isn't this fun! :)

[alysheba88]

Anyone who focuses on "picking winners" without respect to value cannot win in the long term. Just don't see how its possible (unless playing exotics and get a huge hit).

There seems to be a misconception that value means "longshot"

I'll use my Wonder Again example again. I had the horse 5-2 on my line. So at 6-1 it was clearly a great value play. I bet, she won. Now if the horse was 7-5 I would not have bet her. She wasn't worth 7-5 and taking shorter odds than your estimation of her chances of winning is a sure road to ruin. If the horse had won at 7-5 I would have lived with it.

Not sure why this is so complicated for some. I mean you may like the Falcons this week plus 6 points. Would you make the same bet if the Falcons were only getting 1? What about if they were favored by 4? Would you still bet the same as you did when you were getting 6? Or how about a MLB game? You like a team +135 on the money line. Would you make the same bet if someone told you could only have -160 instead? Of course not, because betting a true underdog at favorite odds will deplete your bankroll very fast

[VetScratch]

Game,

The other important point Dave made is that you can't just set your line and leave it be, unless you build-in very large margins of error (ala Fierro, apparently) -- you must adjust it somewhat along with the actual toteboard odds as they flucuate.

So there are two things I would use the toteboard for:

-- Firstly, determine the % of the pool bet on each horse, and use that to adjust my odds line accordingly (according to some function, which could be a simple linear weighting, or the percentages could be used as a variable in a more complex non-linear adjustment with a neural network or something [that's what I do personally]). If this whole system is not automated/computer-assisted, you'll probably be skipping this step, in which case you might want to have your line add up to only 80% or so as suggested, and you also must be on the lookout for "false overlays" (a 15/1 on-the-toteboard horse that you have at 4/1 is probably a bad bet, for example).

-- My line thus adjusted (which is just a way of making the line more accurate and less biased), I look at the actual toteboard odds (or decimal representations of same for greater precision) to see what each horse is paying (or to guess what it will be paying, actually).
Why would you routinely change "your" odds-line? The premise is that you have handicapped the race and formulated accurate probabilities. Since your odds are based on these probabilities, only program changes would precipitate the need to re-handicap the race.

Of course, public odds will fluctuate, but as post-time approaches the objective is to find value by comparing your odds to public odds. Furthermore, public odds may fluctate even after you must bet or pass the race. For this reason, successful value players look for an acceptable premium above the breakeven point by setting parameters based on experience. If co-mingling becomes a recurring problem, they raise their premium margin or seek greener pastures. Some tracks are less susceptible to big past-post drops than others (at least so far).

I almost didn't address this issue, probably because I found it just as difficult following your reasoning as you did following mine. What I mean is, if you have confidence in the accuracy of your handicapped probabilities, why change them in response to public opinion. Conversely, if you have reason to lack confidence, isn't every side of this debate moot? I remain a little bit puzzled for this reason.

[alysheba88]

Originally posted by VetScratch
Game,


Of course, public odds will fluctuate, but as post-time approaches the objective is to find value by comparing your odds to public odds. Furthermore, public odds may fluctate even after you must bet or pass the race. For this reason, successful value players look for an acceptable premium above the breakeven point by setting parameters based on experience. If co-mingling becomes a recurring problem, they raise their premium margin or seek greener pastures. Some tracks are less susceptible to big past-post drops than others (at least so far).



GREAT GREAT POST

[VetScratch]

alysheba88 (MW),
There seems to be a misconception that value means "longshot"Every successful value player I have ever encountered agrees. Most successful plays are like Wonder Again.

My examples may have been misleading (14/1 on Early) because I used previously mentioned probabilities for the sake of continuity within this thread (i.e., 33% & 10%).

[alysheba88]

Well first of all I would not categorize a 6-1 as a longshot. To me double digits and up is a longshot but thats quibbling and obviously a matter of interpretation.

More to the point, if I have a horse at 2-1 and he is 3-1 in actuality is that a longshot? He is 50% overlayed and a great value bet.

Many of the value bets I make are on horses 7-2,4-1, 5-1, odds like that. Far from classic longshots. Usually the result of the favorite being overbet.

Fierro has some great stuff on how you choose between multiple horses that are overlayed.

I think why some scorn value betting is the think it means playing 20-1 and up horses. Or being doomed to infrequent hit rates in search of double digit longshots. Not the case at all.

[Dave Schwartz]

VS,

You said: "Why would you routinely change "your" odds-line? The premise is that you have handicapped the race and formulated accurate probabilities. Since your odds are based on these probabilities, only program changes would precipitate the need to re-handicap the race."


This is why odds line approaches don't work.

To my knowledge there is only one other person that has said anything about changing his probailities/odds line as the tote changes: Bill Benter.

Personally, I like my code better than his. <G>


Regards,
Dave Schwartz

[alysheba88]

Personally I hope most players stick to picking winners regardless of odds. Is to my bankroll's benefit.

[VetScratch]

Originally posted by Dave Schwartz
VS,

You said: "Why would you routinely change "your" odds-line? The premise is that you have handicapped the race and formulated accurate probabilities. Since your odds are based on these probabilities, only program changes would precipitate the need to re-handicap the race."


This is why odds line approaches don't work.

To my knowledge there is only one other person that has said anything about changing his probailities/odds line as the tote changes: Bill Benter.

Personally, I like my code better than his. <G>


Regards,
Dave Schwartz "Your" didn't mean you (or anyone specific except whoever was calculating odds-lines). I was responding to a post by Game Theory, but I don't see any reason to change handicapped probabilities unless there are program changes (post-time scratches and jockey changes). Why does Benter advocate fiddling with handicapped probabilities/odds?

[Dave Schwartz]

Because if you have 2 horses that both have a 33% probability (or 2/1 line, if you prefer), and one goes off at 6/5 and the other goes off at 20/1, they do not win at the same rate.

Or, the simpler answer is, "because it works better."

[GameTheory]

Yeah, it works better. Much much much better, in fact. I would go so far as to say it is the big secret to getting an oddsline approach to work well.

I'll put in another way. When I make up probabilities for each horse, I have a number of variables that go into calculating that. Post time odds (as near as I can figure) is one of those variables, that's all.

Any time you make a line, you're going to bias in it. It can't account for everything -- it just can't. There is always extra info contained in the public odds that you won't have. So I incorporate that info and my line becomes bias-free. Once adjusted, I can consider my line "true" and I don't have to worry about any of the things you're talking about ("credible underlays", etc.). All info is contained in the line, and I just compare prices. An unadjusted line will often contain false overlays, an adjusted line will not.

-- If I bet on all the favorites, I will win about 33% of the time and lose around 17%.

-- If I bet on the top (highest probability) on my UNADJUSTED line (odds not considered), I will win about 26% of the time and lose around 10%.

-- After adjusting my line with the odds, if I bet on the top horse on my line, I will now win about 33% of the time and break even or make a slight profit.

That's what it does for me. Just betting blindly on the top horse every race I can now make break even or make slight profit with around the same winning percentage as favorites. Of course the real money comes from ferreting out the overlays.

[My handicapping + the public] compared to [the public alone] = big profit

[My handicapping alone] compared to [the public alone] = loss or little profit, unless I use all sorts of additional filters and qualifications to account for that extra information offered by the public line.


As Dave suggested, these ideas come from Bill Benter. He has made a gazillion dollars this way...

[Dave Schwartz]

Dang, have I let the cat out of the bag? I may have to slow down on putting proprietery stuff up here. <G>

(Dave actually did not say that the idea came FROM Benter. Only that Benter did something similar.)

Of course, the secret is HOW you adjust the line. We've been doing this for several years... well after Benter did something similar, but way before I ever saw Benter's code. I do not believe that Benter's approach is anywhere near as "smooth" as my approach.

Geez... I am starting to sound like the guy that says "I have all the answers but cannot tell you how it works because... well, I just can't."

Seriously, it is so amazingly simple, but I will wait for the book. <G>


Dave

[Tom]

Do you have a way to adjust the Beyers in your book? :rolleyes:

[GameTheory]

Originally posted by Dave Schwartz
(Dave actually did not say that the idea came FROM Benter. Only that Benter did something similar.)

Well, I got the idea from Benter. I believe my method is better than Benter's as well (better than the way he originally reported it). Of course, even Benter doesn't do it the way he used to anymore...

[VetScratch]

Dave,
Originally posted by Dave Schwartz
Because if you have 2 horses that both have a 33% probability (or 2/1 line, if you prefer), and one goes off at 6/5 and the other goes off at 20/1, they do not win at the same rate.

Or, the simpler answer is, "because it works better." I think I see the dilemma (and possibly why Game Theory has issues too). You are not really talking about about a value system at all. You are to trying to pick winners.

I say this because if you handicap a race and produce 33% as the win probability for two horses, you ARE asserting that both have an EQUAL chance of winning... pure and simple.

The objective of handicapping is to determine the probability of winning for each contestant in an event. This has been true for thousands of years, and nothing magically changed at the start of this millenium.

As I said in an earlier post, "You simply can't overestimate how hard it is for players who have been picking winners for twenty years to convert themselves into value players."

I won't challenge your "What works better" assertion any more than I would challenge Rmania's assertions for sprints at SO-CAL. Everyone should use what works best for them.

It's interesting to hear your opinions, but I don't necessarily buy your conclusions.

In essence, both you and Game Theory have described approaches that handicap both the horses and the betting public as if they were entered in the same event. Value players look at the horseracing competion and the parimutuel competition as two separate events.

Call it what you will, but when you "merge the race with the totes," you are just another guy trying to chase the "smart money."

[Dave Schwartz]

VS,

You said: "Value players look at the horseracing competion and the parimutuel competition as two separate events."

Exactly! And that is why it doesn't work well.


You have chosen to not get it. That is fine.

I will be finished with trying to explain it to you. Just keep doing it your way.


Dave Schwartz

[GameTheory]

As usual, VS, I don't know what the heck you're talking about.

It is very simple. By considering the odds of the public, you can make more accurate probabilities than you can by not considering them. I am looking to extract handicapping information out of the toteboard that I have not accounted for. I am not making any assertions at all until I have the public odds. Those odds are just one of dozens of variables, but they happen to be a very key variable with a very special purpose. It "smooths" out the whole line, tugging this horse & that one closer to reality.

It is not about chasing the smart money because that implies that I would then wager on the "bet down" horses, when in fact you would never wager on such a horse unless you already had him at a high probability to begin with. The main benefit of this adjustment is actually to avoid making bets that might look like overlays but are actually not.

I don't understand your comment that it isn't really a value system at all. As usual, there is some overlap of the definition of the two terms ("picking winners" vs. "value system") since you can never profit without getting value somehow.

But "picking winners" implies that you find one or maybe two horses that you think are most likely to win, and you bet them. It also implies you don't neccessarily have a firm idea of just how often "horses like this one" will win (other than a vague, "I think it will win the most"). But even selection-oriented players often would require certain minimum odds on those selections.

A "value system" implies that you are open to betting on any of the horses at the right price (or possibly any of the horses with at least a certain minimum assigned probability, since the allowed margin of error becomes incredibly small at low probabilities), and very possibly several horses. For instance, it might not occur to a selection-oriented player that if you can eliminate a very strong favorite, you can profit simply by dutching the rest of the field.

I don't see how adjusting a line to make it more accurate is suddenly "picking winners" when you haven't even addressed who might be a good bet until after the point of adjustment. As I said, before the adjustment, you aren't asserting anything -- you essentially have no probabilities yet -- you only have potential probabilities or candidate probabilities.

There is no dilemma, there is nothing complicated, it is very simple. Considering the public odds will allow you to generate more accurate probabilities (more unbiased, really) than otherwise. That's it. The odds happen to contain relevant information that isn't available elsewhere. So I extract the information, and move on.

[VetScratch]

Dave & Game,

This isn't war or peace; just a difference of opinions. If I didn't respect you both, I wouldn't want to debate with you, or maybe I would ask Sengbush take my place. :)

Note that I don't claim that your approach "doesn't work." Only you can judge profitablility. Without any way of knowing, if I made that kind claim, it would be a sign of resignation (or desperation), as such claims are in most debates.

I do appreciate your explanation, Game, and I concur with some of it (but maybe not in a literal sense). Here is my take on toteboard action.

There is a portion of the handle that is bet on the basis of unpublished information (i.e., unpublished even in the largest downloads/databases supplemented by all ancillary references like Ragozin sheets or Takach newsletters).

In a typical race, more than one of these special wagering motivators will exert its influence on the parimutuels, often at cross purposes with respect to different horses.

A second dimension of special angles also comes into play when unpublished information is a negative factor. For example, some horses are dropped to lose them with little hope of winning, but that often discourages "insider" wagering on the race. In any case, money that is shunted away from a horse or from a race by unpublished information also influences the parimutuels.

At the same time, chunks of totally meaningless or deceptive "smart" money comes into the pools. Trainers and owners are often "natural enemies." Many trainers are extremely selective about sharing their true intentions. Owners get stiffed all the time (as well as all the friends that they tout). Many owners are "outsiders" in the eyes of the "smiling" backside community. This is a third dimension of special angles based on unpublished "misinformation."

I happen to believe that wagering based on misinformation is underestimated by handicappers. Greed is the chief motivator for bad buzz. However, if barns and jockeys are betting for reasons that are unseemly, security becomes a serious issue, and you can be sure that they will not be "broadcasting" anything but misinformation. Gill and Stronach may run their shows, but many owners are cordially tolerated but are actually treated like weekly renters at a Residence Inn.

You have apparently found a way to detect and interpret this mix of toteboard influences. I think that's great!

I know others who work hard at this, usually by trying to analyze toteboard activity as if they were monitoring the bid and asked quotation traffic that pours into NYSE or NASDAQ. In my opinion, this approach is losing value because of national simulcasting. Hubs and satellite wagering centers consolidate everything into dollar totals. I am told that this is not as useful as the bid/asked market wires, which do transmit discrete bid and asked quotations rather than summary information.

[GameTheory]

The only difference of opinion is that we're talking about opinions. I'm talking about hard facts.


If you make a line that does not consider the public odds, it can be improved upon by considering them. Period.

If you take all the lines you've produced for a bunch of races and look at all your 33% horses (or any specific percentage) and put them into groups according their final odds, you will find those horses that went off at higher odds won less often than those with lower odds, all in glorious disregard of your 33% rate they are all supposed to win at no matter what. I defy anyone to make a oddsline over a large set of races where this is not true if they did not consider the odds in the assignment of probabilities. Now those horses at higher odds may well outperform those odds, but each group will not be winning at a 33% rate (or any constant rate) I can assure you.

However, my 33% horses after this adjustment do win 33% of the time regardless of their final odds (more or less, nothing is perfect). Now at this point this is a slight problem with theory vs. practice because you obviously can't use the actual final odds if you want to assign your probability before the race starts, so you just use your best guess based on the current toteboard odds, with possibly an adjustment based on what you think the final odds will be (you use projected final odds). But my 33% statement holds true in theory if I do use final odds, and I just have to leave some margin for error in practice when betting actual races.

If you like, I can prove it to you. Send me a set of lines you made for any group of races, and I'll improve those lines by using the odds. It's childs play. (I would need a set of "past" races to use to determine what the proper formula should be for adjustment, and then a set of new races to use to show you the results of applying the formula. And of course the method used to derive your original lines must be consistent.)

[VetScratch]

Originally posted by GameTheory
The only difference of opinion is that we're talking about opinions. I'm talking about hard facts.


If you make a line that does not consider the public odds, it can be improved upon by considering them. Period.

If you take all the lines you've produced for a bunch of races and look at all your 33% horses (or any specific percentage) and put them into groups according their final odds, you will find those horses that went off at higher odds won less often than those with lower odds, all in glorious disregard of your 33% rate they are all supposed to win at no matter what. I defy anyone to make a oddsline over a large set of races where this is not true if they did not consider the odds in the assignment of probabilities. Now those horses at higher odds may well outperform those odds, but each group will not be winning at a 33% rate (or any constant rate) I can assure you.

However, my 33% horses after this adjustment do win 33% of the time regardless of their final odds (more or less, nothing is perfect). Now at this point this is a slight problem with theory vs. practice because you obviously can't use the actual final odds if you want to assign your probability before the race starts, so you just use your best guess based on the current toteboard odds, with possibly an adjustment based on what you think the final odds will be (you use projected final odds). But my 33% statement holds true in theory if I do use final odds, and I just have to leave some margin for error in practice when betting actual races.

If you like, I can prove it to you. Send me a set of lines you made for any group of races, and I'll improve those lines by using the odds. It's childs play. (I would need a set of "past" races to use to determine what the proper formula should be for adjustment, and then a set of new races to use to show you the results of applying the formula. And of course the method used to derive your original lines must be consistent.) Game,

Where did 33% become a magic number? That probability, as well as 10%, was introduced in this thread to illustrate calculations for natural and parimutuel odds. Odds don't get "made," they are calculated from probabilities.

Handicapping is NOT associated with any predetermined probabilities or odds-line patterns. In fact, the notion of shoe-horning a set of odds into a template or talk of "smoothing" odds-lines always makes me suspicious. If racetrack parimutuel systems operated that way, horseracing would soon perish.

We always seem to end up in the same loop. Handicapping is not about picking winners. The objective is to accurately determine probabilities. You and Dave seem to defy anyone to achieve absolute accuracy. That is like me defying you to absolutely pick all winners. In both cases, absolute perfection is like Plato's perfect triangle. We all know what Plato meant, but no one has every produced a perfect triangle. Perfection is approachable but not attainable.

I think value players have been maligned in the past because it is difficult to verify their results. They have no selections to post ahead of time, as is the case with players who attempt to pick winners. However, as I posted in the Handicapping Contest thread, I think this awkward circumstance is about to change.

I am confident that within the next year, one or more of the wagering services will introduce enhanced security accounts where the account holder will be able to let guests view the transaction history and financial summaries. Since guests are potential customers, the only constraining issues concern due diligence on privacy matters and the details of implementation.

This would enable everyone to sidestep all the tiresome arguments about what works best... let the results speak for themselves!

[alysheba88]

"A "value system" implies that you are open to betting on any of the horses at the right price"


That is exactly right Game Theory. Thats exactly what is done. Only thing I would add to that is the use of the word "any". I made lines on either three or four contenders (as Fiero suggests). So will be playing one of my contenders not any horse.

[GameTheory]

Originally posted by VetScratch
Where did 33% become a magic number? That probability, as well as 10%, was introduced in this thread to illustrate calculations for natural and parimutuel odds. Odds don't get "made," they are calculated from probabilities. It's not a magic number, just an example. My point is if you pick any set of horses from your odds line that all have the same approximate probability assigned to them (by you) and put them into groups based on their odds, the higher odds horses will win less often than the lower odds horses. If your probability is independent of the odds as you seem to think it could be, that shouldn't happen -- they should all win at more or less the same rate. But they won't.

I don't know what the most of the stuff you bring up has to do with the topic. I'm not sure you actually read the posts before you start writing your responses. Either that or I'm just being very unclear as you seem to misunderstand just about everything.


Handicapping is NOT associated with any predetermined probabilities or odds-line patterns. In fact, the notion of shoe-horning a set of odds into a template or talk of "smoothing" odds-lines always makes me suspicious. If racetrack parimutuel systems operated that way, horseracing would soon perish.I haven't the slightest clue what you're talking about here, especially the last sentence.


We always seem to end up in the same loop. Handicapping is not about picking winners. The objective is to accurately determine probabilities. You and Dave seem to defy anyone to achieve absolute accuracy.You're the one who keeps bringing up picking winners, and I don't know why. I don't deny anyone accuracy; I'm telling you how to achieve it. I am saying that any line that doesn't consider (in other words, ignores) the toteboard odds can be improved upon, yes. And I even offered to prove it empirically. Accuracy is the wrong word anyway -- it is bias I'm talking about. Accuracy is easy to achieve. Just divide 1.0 by the size of the field in every race and your lines will be 100% accurate. But they will be biased, just like your line, because horses that should be winning at the same rate won't be, and the discrepancies will mirror the odds.


In both cases, absolute perfection is like Plato's perfect triangle. We all know what Plato meant, but no one has every produced a perfect triangle. Perfection is approachable but not attainable.Did you notice how right there in my post I said, "Nothing is perfect"?


I think value players have been maligned in the past because it is difficult to verify their results. They have no selections to post ahead of time, as is the case with players who attempt to pick winners. However, as I posted in the Handicapping Contest thread, I think this awkward circumstance is about to change.

I am confident that within the next year, one or more of the wagering services will introduce enhanced security accounts where the account holder will be able to let guests view the transaction history and financial summaries. Since guests are potential customers, the only constraining issues concern due diligence on privacy matters and the details of implementation.

This would enable everyone to sidestep all the tiresome arguments about what works best... let the results speak for themselves!What *are* you talking about? Here, do this:

my_probabilities*.5 + public_probabilities*.5 = new hybrid probability

That is the beginning and end of my point^^^^

You take your probabilities (assigned by you based on your handicapping) and mix them with the public's *somehow*. The above is a very simple linear way to do it -- half yours/half the publics. I AM NOT SUGGESTING THIS IS A GOOD FORMULA. I am trying to illustrate what I am talking about -- you USE THE TOTEBOARD AS A HANDICAPPING FACTOR IN A PRECISE MATHEMATICAL WAY.

That's it. That's all. There is nothing else to it.

I'm not talking about picking winners vs. value. I'm not talking about emerging technologies from online wagering companies. I am not talking about smart money, dumb money, or misinformation. I am not talking about bid/ask market wires. I am not talking about Plato's theory of forms. And just to head you off at the pass, I will say in advance that neither I am talking about French food, nor the designated hitter rule, nor what happened to the missing 18 1/2 minutes of the Nixon tapes, nor super-string theory, nor whether the King's Gambit Declined Falkbeer Counter Gambit is viable for black!

I am simply saying you can improve your probabilities by using the public's to augment yours. I am saying that Bill Benter, who has probably made more money betting on horses than anyone in history, does the same thing. Maybe, just maybe, it is a good idea, despite the fact that it makes you "suspicious".

[GameTheory]

Originally posted by alysheba88 (MW)
"A "value system" implies that you are open to betting on any of the horses at the right price"

That is exactly right Game Theory. Thats exactly what is done. Only thing I would add to that is the use of the word "any". I made lines on either three or four contenders (as Fiero suggests). So will be playing one of my contenders not any horse.

Yes, that's what my comment about very small margins of error was about. Limit your betting possibilities to the contenders. Mark Cramer suggests considering any horse that is 6-1 or lower on your line (approx. 14% probability). At low levels of probability, a 2-3% error is huge -- the difference between a 15-1 horse and a 30-1 horse is only 3% probability.

[VetScratch]

You take your probabilities (assigned by you based on your handicapping) and mix them with the public's *somehow*. The above is a very simple linear way to do it -- half yours/half the publics. I AM NOT SUGGESTING THIS IS A GOOD FORMULA. I am trying to illustrate what I am talking about -- you USE THE TOTEBOARD AS A HANDICAPPING FACTOR IN A PRECISE MATHEMATICAL WAY.This is merging the race with the tote, which diminishes the variances between the two sets of probabilities. That would not be productive for me, is productive for you, and may or may not be useful to others.

When we go beyond that, we are dealing with opinions until there is an objective way for any stripe of horseplayer to demonstrate results. As I said, I believe this is forthcoming, and I presume you would also rather demonstrate than haggle.

[JustMissed]

Does it really matter what a person does to their personal odds line? After all, their odds line is developed by their personal handicapping methods and in fact doesn't mean squat to anyone else anyway.

After all, once the race is run-if you didn't have the winning horse at 1-99 your odds line was wrong to begin with.

JustMissed

[karlskorner]

I started this post with a quote from Steve Fierro

"The majority of handicappers spend far to much time on unimportant items THEY HAVE BEEN LED TO BELIEVE are important"

IMHO the 'word" VALUE falls into this category.

I am off to the track for a bowl of chowder. a Boston Creme doughnut and then the Paddock.

[alysheba88]

Well you must have missed the rest of Fierro's book if thats the only quote you remember
:)

[alysheba88]

Just missed, not at all.

I bet Tampa Bay in last years Super Bowl with the points. Would I have made the same bet if they were giving 26 points even though it would have been a winner in the end? No, anyone who would cannot win money in the end.

As Cramer and others have said people will not change for the most part on this issue and will insist on picking winners (usually favorites with the highest numbers).

[JustMissed]

Originally posted by alysheba88 (MW)
Just missed, not at all.

I bet Tampa Bay in last years Super Bowl with the points. Would I have made the same bet if they were giving 26 points even though it would have been a winner in the end? No, anyone who would cannot win money in the end.

As Cramer and others have said people will not change for the most part on this issue and will insist on picking winners (usually favorites with the highest numbers).

Your assumption is incorrect. A sports point spread is not the same thing as pari-mutuel odds. A point spread can determine whether or not you win a sports bet. Pari-mutuel odds determine how much you win if you have made the correct choice.

Dale Baird ran Three Stikes in the 10th at the Mountain Tuesday night. A 30k claimer dropping to 4K(One of Dale's gimmick horses where he dares anyone to claim). Everyone and their brother knew this horse was going to win and the only decision was how much are you going to tip the teller when you cash your ticket.

Had I been making a line this horse would have been 1-99. When I hit the print ticket the tote was 1-1. The horse won going away and paid $3.

I suppose a value player would not take that bet had their line been 6-5 or higher but for the life of me I can't figure out why anyone would pass a "gimme race".

JustMissed

P.S. Three Stikes won by 5 1/2 lengths and was not claimed.

[alysheba88]

You won and thats great. But is that play a long-term winning one? That's what a value play asks himself/herself. Betting 1-2 shots who should be 6-5 or better is not a long-term winning play.

My sports analogy was this. I liked Tampa getting 4 points. Thought they had a very good chance of winning outright. But there is a point where I would not bet them. If Tampa had been favored by 6 I would not have bet, even though I thought they had a good chance of winning. Because the risk reward was not there.

Now if you focus only on the result, in this case giving 26 was a "good bet" since it "won".

Value is a universal concept, whether it be sports, horses, stocks houses.

Believe me I didn't always think this way, but am a firm believer in it now. Too much time is spent looking at the one time result, instead of the longer term tread. People make all kinds of strange decisions at the blackjack table. Just because it works out once doesn't mean its right.

Not to be arrogant but I don't mind missing out on 1-2 sure things. I do mind missing out on a 6-1 shot who should have been 2-1

[Red Knave]

This has been a fantastic thread!

[VetScratch]

Game,
I'm not talking about picking winners vs. value. I'm not talking about emerging technologies from online wagering companies. I am not talking about smart money, dumb money, or misinformation. I am not talking about bid/ask market wires. I am not talking about Plato's theory of forms. And just to head you off at the pass, I will say in advance that neither I am talking about French food, nor the designated hitter rule, nor what happened to the missing 18 1/2 minutes of the Nixon tapes, nor super-string theory, nor whether the King's Gambit Declined Falkbeer Counter Gambit is viable for black!Superb... Inspired... I loved it! :)

[JustMissed]

I would not disagree with one thing you said regarding your value play.

I think maybe we have moved over into the area of "style of play", and I would never question a person's style of play.

Perhaps I have been slow to identify value plays, as defined by Fiero, as a style of play, and if so I would apologize.

JustMissed

[GameTheory]

Originally posted by VetScratch
This is merging the race with the tote, which diminishes the variances between the two sets of probabilities. That would not be productive for me, is productive for you, and may or may not be useful to others.It would be productive for you (& everyone else), but you choose to dismiss it before even trying it just because it smells fishy to you. Let me ask you something -- have you ever found something to work well that you didn't expect to work well? Does everything you think will work well turn out to be great? Are you always right about everything without ever having to do tests? Have you ever discovered anything that you found counter-inituitive at first? Or is it always, "I've made my call on this, that's my final word, so don't bother me with evidence."?

We can cease these discussions now, but you can't make a definitive claim of the certainty of your position based on a "oh, that reduces the variances, so it's no good" when I'm willing to prove the utility to you using actual odds lines you've produced yourself (assuming you can produce such odds lines). Remember I said that the linear formula was just a example to get my point across -- in reality what happens with a proper blending formula is the false overlays disappear and the real overlays remain (maybe even becoming "bigger" overlays).

Anyway, suit yourself, I tried.

I'm starting to get that formula_2002 feeling again....

[GameTheory]

Originally posted by VetScratch
Game,
Superb... Inspired... I loved it! :)

Yes, I knew you'd enjoy that -- from one caustic wit to another...

[VetScratch]

Game,
Originally posted by GameTheory
It would be productive for you (& everyone else), but you choose to dismiss it before even trying it just because it smells fishy to you. Let me ask you something -- have you ever found something to work well that you didn't expect to work well? Does everything you think will work well turn out to be great? Are you always right about everything without ever having to do tests? Have you ever discovered anything that you found counter-inituitive at first? Or is it always, "I've made my call on this, that's my final word, so don't bother me with evidence."?

We can cease these discussions now, but you can't make a definitive claim of the certainty of your position based on a "oh, that reduces the variances, so it's no good" when I'm willing to prove the utility to you using actual odds lines you've produced yourself (assuming you can produce such odds lines). Remember I said that the linear formula was just a example to get my point across -- in reality what happens with a proper blending formula is the false overlays disappear and the real overlays remain (maybe even becoming "bigger" overlays).

Anyway, suit yourself, I tried.I actually made no definitive claim. What you quoted from me was the position I took from what has been said, which was:

"This is merging the race with the tote, which diminishes the variances between the two sets of probabilities. That would not be productive for me, is productive for you, and may or may not be useful to others."

The question at hand was what could be gained from a "proper blending formula" for a set of actual track odds and a set of handicapped value odds. If there was more to it than these data variables, I would have been intrigued, but neither did I say it does not benefit you or might not benefit others.

Did I misunderstand the proposition, or was it as I stated? All that is required are the two odds-lines? Example: Track Value
P# Odds Odds
1 4.3 2.7
11 105.5 31.2
4 4.2 4.6
2 4.0 4.8
9 43.3 15.7
10 103.8 99.0
12 21.5 24.0
7 99.5 82.3
6 9.0 9.2
8 67.0 31.1
5 109.5 13.9
3 1.6 3.9
Without a guiding theory or data not previously mentioned, how many folks would not be incredulous? I did not refute that you do benefit from blending two such odds lines, but it seems difficult to understand what "blending" by any formula can accomplish other than produce a new set of odds values.

Don't you acknowledge that it sounds a little like pulling a rabbit out of a hat?

Before the race was run, the credibility of both odds lines was in doubt, and the results were produced by the horses on the racetrack and not by either set of odds.

Unless "blending" incorporates other data, or is based on an assertion that was never mentioned, it sure seems as if blending two sets of odds can only produce a new odds line that must inherit the chief attribute of the originals, such that credibility is still in doubt before the race is run.

If there is more to it than "blending," then it has not been discussed. Otherwise, if there is no more to it than "blending," then I don't think my doubts are unreasonable.

[Dave Schwartz]

Since GT worked so hard at this, I will take one more small stab.

The goal in producing an odds line is to first reflect each animal's REAL chances in the race. THe second logical step (for most people) is to add a point at which they would bet.

What I believe GT is saying is that the more accurate the line is the better the results will be. (Yes, we realize that there may be no definitive way to test this; that would result in a different argument.)

So, what GT and I do, albeit differently, is to "smooth" the line towards reality.

You can present charts until the cows come home to explain why you think this approach doesn't/con't/won't work. The bottom line is that is does work FOR US. That is not open for debate. (And, no, I won't post 6 months of 4 tracks per day to prove it.)

VS, if you care to gain something from this thread, please do so. Otherwise, I will once again (but this time with more feeling) exit this thread.


Regards,
Dave Schwartz

[GameTheory]

Yes, all you get is a new odds line. But a better one than before.

What is to be gained is avoiding bad bets and making more good bets.

It comes down to scenarios like this:

You've got a horse at 4/1 on your line (20% probability) but he's 7/1 on the board (10% of the pool). Looks like a good overlay.

Problem is, in reality, when horses are 4/1 on your line and 7/1 on the board, they only win 12% of the time, not 20%, and such bets actually have a slight negative expectation or break-even at best.

[With my adjustment, I would not make the bet, while you would be taking a loss. And if you used Kelly-type betting where the amount of edge you've got is important, you'd be taking a big loss.]

And let's say in the same race there was a horse at 3/1 on your line, and he's also 3/1 on the board. No value there, and you pass on that wager.

Problem is, if you had downgraded your 4/1 horse from a 20% to 12% probability, that means you're also upgrading the chances of the rest of the horses, and your 3/1 horse is now a slight overlay. Only after adjusting the whole field in such a way (based on what has actually happened in the past) do the true underlays & overlays emerge.

So here was a race with an overlay that you missed and didn't bet on while you were betting on an underlay that you thought was an overlay.

The adjustment I am suggesting discovers these relationships between your odds lines & the tote odds and corrects for them. And now you're moving from the red to the black, or from the black to the gold.

Please, look at your own data (if you do in fact make odds lines and all this isn't just an academic exercise) and stratify all your 25% horses by odds groupings, and your 20% horses, and your 15% horses and so on. The rate at which the horses from any specific assigned probability (e.g. the 25% horses) will win will not be constant between odds groups -- I guarantee it. As the odds get higher, the win% will go down. As the odds get lower, the win% will go up.

THIS IS WHAT IS WRONG WITH MOST VALUE LINES. You're looking at an average rate of winning, but in fact if you break it down by odds you can see the the rate of winning is actually moving up & down in lock-step with the odds. Wouldn't it be great if all your 25% horses won 25% of the time regardless of the odds?


So what is gained is that you find out when your 25% horses are really 15% horses and when they are really 25%, or 30%, or whatever.

If you're going to argue that is not useful to know, then you're just being silly.

I have said everything at least three times already. I've got nothing left.

[VetScratch]

Dave,Originally posted by Dave Schwartz
Since GT worked so hard at this, I will take one more small stab.

The goal in producing an odds line is to first reflect each animal's REAL chances in the race. THe second logical step (for most people) is to add a point at which they would bet.

What I believe GT is saying is that the more accurate the line is the better the results will be. (Yes, we realize that there may be no definitive way to test this; that would result in a different argument.)

So, what GT and I do, albeit differently, is to "smooth" the line towards reality.

You can present charts until the cows come home to explain why you think this approach doesn't/con't/won't work. The bottom line is that is does work FOR US. That is not open for debate. (And, no, I won't post 6 months of 4 tracks per day to prove it.)

VS, if you care to gain something from this thread, please do so. Otherwise, I will once again (but this time with more feeling) exit this thread.If Game Theory agrees with your post, then you have supplied the missing part of this puzzle.

Three things instead of two are needed to make blending useful:
(1) Track Odds
(2) Handicapped Odds
(3) A reason to doubt the accuracy of the handicapped odds (as compared to the track odds).

This brings us right back to you defying anyone to produce handicapped odds that are comparable in accuracy to track odds.

If you are right, blending would make sense for obvious reasons.

However, using identical methodology for determining accuracy, what if there was no significant difference in accuracy between track odds and handicapped odds?

Then I submit that blending could still make sense because the results might be characterized by lower coefficients of variation than either of the two "equally" accurate sources.
This is that aspect of "smoothing" that has appeal, but we still have three requirements:
(1) Track Odds.
(2) Handicapped odds with a confirmed degree of accuracy comparable to the track odds.
(3) Confirmation that blending will reduce coefficients of variation without sacrificing accuracy.

The methodology for testing the accuracy of odds lines (actual or handicapped) is no mystery. Anyone with Quirin's 1979 book can start on page 293 and soon see how this may be done, and may easily be refined to produce extended results for more discrete value intervals.

All this stuff has been around so long that many handicappers have undoubtedly tested the accuracy of both track odds and their handicapped odds. "Even I know" three serious handicappers who have done this many times over with large samples of races. More than three, I'm sure, would contest your assertion that handicapped odds as accurate as track odds are absolutely unattainable.

What has me excited about what you and Game have contributed to this thread is persistence. Up until now, I had not realized the potential value of thoroughly examining coefficients of variation. Now that track odds can be captured online, it makes perfect sense to determine if blending will reduce variances without sacrificing accuracy.

If this proves to be true, I will be the first to proclaim your genius!

The implication, of course, is that "smoothing" via "blending" would exceed your goal of moving "towards reality." The result could be likened to a "new reality."

[VetScratch]

FWIW, after I do some testing, I will post my test results. If blending produces an odds-line with lesser variance versus order of finish than the two sources, that should really excite exotic players who make their own lines.

[alysheba88]

If your 4-1 horses are winning 12% of the time then your line is wrong. That's all. Or else the contender selection process is off

[hurrikane]

interesting banter

The goal in producing an odds line is to first reflect each animal's REAL chances in the race.


I know what Dave meant to say was:

The goal in producing an odds line is to first reflect YOUR OPINION OF each animal's REAL chances in the race.

Because of this one variable the tests that Dave, Game, or Vet do will be meaningless to anyone but themselves.
IMHO, in this day of simulcast and playing 10 tracks a day you obviously can't look at every horse in every race, at least not the way I play. The one quantifier you have is the tote. . It makes perfect sense to use the parimutual system to smooth your initial assuptions used to create your odds line.

Game

I'm starting to get that formula_2002 feeling again....

ahhh..don't we all know that feeling. :D

[hurrikane]

I do want to ask one question of JustMissed.

I've never heard anyone use the term 'gimme race' when talking about mountaneer.
I'm curious why this horse went off at 1-1. Why didn't you sell your house and car and dump on him. If I ever find a 'gimme race' that is exactly what I plan to do.

And just a head up friend, if you bet Baird based soley on the information you posted you will not have your house or car very long. Just my opinion, well actually, it's the opinion of my data.

[JustMissed]

Originally posted by hurrikane@HTR
I do want to ask one question of JustMissed.

I've never heard anyone use the term 'gimme race' when talking about mountaneer.
I'm curious why this horse went off at 1-1. Why didn't you sell your house and car and dump on him. If I ever find a 'gimme race' that is exactly what I plan to do.

And just a head up friend, if you bet Baird based soley on the information you posted you will not have your house or car very long. Just my opinion, well actually, it's the opinion of my data.

Hurrikane-Did you listen to the simulcast show? When Mark and Nancy barely mention the 'big horse' just before the post parade break and then delicately point out the negatives when they come back-you better reach deep into your pocket cause you know they are trying their damnest to keep the price up. (They both also had the horse as their third pick-I guess so no one would accuse them of stiffing the simi viewers).

We normaly only make prime bets at the mountain and some nights may only get one or no plays. Tuesday happened to be a good night.

Oblat took the first race. His sire is Funny Cides' sire. Both sire and dam broke maiden first out, sire had 20% 1st start record and with a 19% jock on board this was almost a "gimme race" and paid $5.40.

The sixth was won by Loving Feeling, who won it's maiden race last out but placed third by DQ winning by 4 lengths. He won this, his second maiden race and paid $4. A near "gimme".

The 9th was a $75K stakes race with a field of 7. After two scratches the obvious winner of this field of five was Sonic West, out of Gone West, with a 3-2-1 out of 7 record and a 131.9 Prime Power rating, this also was a "gimmie" and paid $3.80.

As far as Dale Baird, just a few weeks ago he had a huge dropper that the kept in the paddock stall, guarded in front by the grooms, till past the claiming deadline. Damnest thing I've seen. The camera keep showing the horse or what little you could see of him, and Mark keep going on about how Baird didn't want anyone to look at the horse cause they might see he was gimped and so forth. Well, as I expected the horse won going away at a pretty good price because by post time everybody except the wise guys thought the horse was lame-he went unclaimed.

Anyway, we will be back playing the mountain tonight and hopefully we will get a play or two.

Good luck to you,

JustMissed
;)

[hurrikane]

Good luck to you JM.

Mnr is one of my fav tracks.
I've just never heard the term 'gimme' used at Mnr.

[VetScratch]

Game,
Please, look at your own data (if you do in fact make odds lines and all this isn't just an academic exercise) and stratify all your 25% horses by odds groupings, and your 20% horses, and your 15% horses and so on. The rate at which the horses from any specific assigned probability (e.g. the 25% horses) will win will not be constant between odds groups -- I guarantee it. As the odds get higher, the win% will go down. As the odds get lower, the win% will go up.

THIS IS WHAT IS WRONG WITH MOST VALUE LINES. You're looking at an average rate of winning, but in fact if you break it down by odds you can see the the rate of winning is actually moving up & down in lock-step with the odds. Wouldn't it be great if all your 25% horses won 25% of the time regardless of the odds?
Is this a put-on?

The most fundamental premise of handicapping has always been (for centuries) that accurately handicapped events should occur in proportion to their odds-based probabilities.

By definition, the purpose of handicapping is to make risks and rewards proportionate (even if a house percentage makes all risks a losing proposition). Every odds-based gambling gimmick in the world would collapse if this was not the premise.

What school of horseplayers were you addressing, the pickers or handicappers? In all of the threads that I have read here, I have never seen anyone attempt to demonstrate that horses win at a flat rate. Indeed, credible race-result studies show that horses tend to win in accordance with their odds-based probabilities when thousands of races are considered.

If parimutuel horseracing didn't reflect this characteristic, it would have perished a long time ago.

If ANYONE ignores this, they should abandon handicapping and go back to picking winners.

The only confusion that I have ever noticed is that some folks do not know how to convert parimutuel odds back into probabilties. This confusion arises because track odds are diluted by takeout and breakage before they are presented to the public. Thus, for a 33% win probability, the parimutuel odds will be about 1.40/1 and not 2/1.

Small differences occur because takeout and even breakage can vary from track to track, but the same principle applies.

I think you must have been kidding. Did you think you would get any takers? Your "I guarantee it" was good dry humor.

[canuck]

Vet wrote

By definition, the purpose of handicapping is to make risks and rewards proportionate (even if a house percentage makes all risks a losing proposition). Every odds-based gambling gimmick in the world would collapse if this was not the premise.

-----------------------------------------------------
If I have read this correctly-you are comparing a hoss race to a bacarrat game.

I used to play HTR and went with their 5 top contenders based on their value line.

I looked a literally hundreds of races very closely and found one irrefutable fact.

The closer the actual odds were to the projected odds the more they won.

If the computer spit out Horse A was a value bet at 5-1 and ends up being 20-1 there is no way he will win more often than Horse B-who was a value bet at 6-1 and ends up being 9-1.

Why does this happen?

The public collectively knows more than we do.

[steve fierro]

Hi Everyone,

I have taken the Fairplex meet off for my first vacation from racing in 5 years. I am just getting back and was surprised at the thread and the many great responses. I am a big pace advantage fan. Thanks to all for the nice comments. Let me answer a couple of items. Players that know me are well aware I am no man of mystery. I have always been very open about my playing activities as anyone who knows me will verify.

First, one of the key items I believe all players should concentrate on is being a winning player. How much of your income is derived from the windows is not as important as the fact that you are producing a profit and know why.

Second, I have isolated that there are four items that all winning players I have encountered have in common. Handicapping skills is only 1. You next must respect the game as a money venture not a picking winners venture. You must have your price. Then you must formulate a technique that you will follow through on while actually playing. Then finally you must keep records to verify your success and failure rate. The last item will be a real eye opener.

I do derive a strong percentage of my income through the windows. Some months I lose a strong percentage of my income through the windows. Some months the win is very nice. It is no different than any stock investments I have. Racing for me is totally a money game and an investment market.

Bob Harris mentioned a few other opportunities that have come my way. It is exciting to do racing for a living. I have been fortunate throughout my racing career to be asked to participate in racing in many other ways than just betting through the windows. It has given me the opportunity to teach, bring NTRA tournaments to players. Be an ESPN broadcaster. The best part however has been getting to meet so many players from around the country. We are a fraternity onto ourselves. We may disagree on certain items but we still have one common bond. We all love racing. Play your game. Play for money.

Right now I have to get back to my last day at Disneyland. Thanks again Steve Fierro

[alysheba88]

Steve,

Thanks for stopping by. Enjoy Disneyland. Need to hit a trifecta to make it though the day there with wallet intact :)

[keilan]

A couple of comments --

This is perhaps one of the finest 3 threads I have read on this site.

After sleeping on some of the posts I'm convinced that the smoothing of one's "odds line" has astonishing merit.

I have adopted Steve Fierro's method of generating a odds line as outlined in his book but incorporating the tout board- filters/rules would add an new dimension to one's wager strategies.

My thanks goes out to Game Theory and Dave Schwartz for sharing their expertise in this area when they didn't have to -- very generous, also to VS for keeping the discussion alive.

[Dave Schwartz]

Hurrikane,

You said: "The goal in producing an odds line is to first reflect YOUR OPINION OF each animal's REAL chances in the race. "

No, my goal is to reflect the actual chances of each horse. I will not, of course, accurately do that, but it is still my goal.


Dave

[Dave Schwartz]

Steve,

Nice to see you here. Come by more often.

As I have said in print several times, your book started me on a path towards understanding value. I look forward to your next one.


Dave

[VetScratch]

Originally posted by canuck
Vet wrote

By definition, the purpose of handicapping is to make risks and rewards proportionate (even if a house percentage makes all risks a losing proposition). Every odds-based gambling gimmick in the world would collapse if this was not the premise.

-----------------------------------------------------
If I have read this correctly-you are comparing a hoss race to a bacarrat game.

I used to play HTR and went with their 5 top contenders based on their value line.

I looked a literally hundreds of races very closely and found one irrefutable fact.

The closer the actual odds were to the projected odds the more they won.

If the computer spit out Horse A was a value bet at 5-1 and ends up being 20-1 there is no way he will win more often than Horse B-who was a value bet at 6-1 and ends up being 9-1.

Why does this happen?

The public collectively knows more than we do.
"If the computer spit out Horse A was a value bet at 5-1 and ends up being 20-1 there is no way he will win more often than Horse B-who was a value bet at 6-1 and ends up being 9-1. Why does this happen?"

Because horses "tend" to win in proportion to the probabilities that equate to their parimutuel odds.

Confusion about this is a characteristic of horseplaying tools that try to serve too many masters. A good tool should not be marketed outside its proper niche.

Most horseplayers insist that profit is their motive, but they really don't like to play the horses with the rigorous discipline that is required to outperform the betting public by a margin that shows a profit. Performance, of course, is measured by financial results and should not be confused with percentage of winners or other yardsticks.

The marketplace serves the public on the basis of supply and demand. Tools for horseplayers tend to respond to demand. The big national providers of horseplaying tools have found that it is much easier to sell selection-oriented products than pure handicapping products. In some cases they advertise tools as Swiss Army Knifes. A recent example would be advertising for a big-name tout sheet that is described as an odds-line tool for value players but brags about top selections winning at a high percentage. Several PA members criticized this product. My comment on this board was something like "if it didn't work, I hope you at least lost with dignity." What I meant was you can't have it both ways... use it one way or the other, but if you are trying to use it as advertised, you should be too embarrassed to admit it.

Selection-oriented tools provide odds lines because users ask for them. If any of the top-notch selection programs suddenly omitted odds there would be mayhem in the customer service call center. At the same time, handicapping tools appeal to such a small share of the market that providers are sorely tempted to exaggerate (and thereby misrepresent) their purpose or tout results that are meaningless such as "had five of nine winners on top at KEE."

I believe there are very good selection tools and very good handicapping tools, but I don't think any tool that attempts to serve both markets is likely to succeed. This is not to say that independently conceived tools could not be bundled and marketed as one product like Microsoft Office.

[hurrikane]

Good to see you post Steve.


Dave, I understand that is your goal but the odds line is still your opinion and will never match someone elses odds line.

[VetScratch]

Hurrikane,
Originally posted by hurrikane@HTR
Good to see you post Steve.


Dave, I understand that is your goal but the odds line is still your opinion and will never match someone elses odds line. Although Dave and I don't see eye-to-eye on some issues, what he said about his objective was absolutely on target.

The objective of handicapping is to produce "real" probabilities and accurate odds. The results can also be tested. Yes, it is a goal, but that is why tests are run to determine if horses are winning in proportion to their probabilities. When you get results that equal or exceed the accuracy of racetrack parimutuel probabilities, that's as real as it gets (IMHO)!

[GameTheory]

Originally posted by VetScratch
Game,
Is this a put-on?

<snip>

In all of the threads that I have read here, I have never seen anyone attempt to demonstrate that horses win at a flat rate. Indeed, credible race-result studies show that horses tend to win in accordance with their odds-based probabilities when thousands of races are considered.

<snip>

Your "I guarantee it" was good dry humor.


No, it is not a put-on, and my guarantee is not a joke. Remember, I'm talking about looking at a group of horses where you have assigned the identical probability to each of them. So they *should* win at a flat rate if your probability is correct, right? If I break them up by any of the handicapping factors that you DID CONSIDER, then they probably would win at a constant rate between groups. Right? But here we have the odds, probably the most powerful factor of all, and you want to ignore it.

I've said this part over and over. Look at all your 3-1 (value line) horses *ONLY*. Do they all win at the same rate when those horses *ONLY* are stratified by odds? NO. Shouldn't they? (That is what you keep saying -- your goal is to assign a probability that is correct.)

Another way of saying this is: if you can find any factor easily available to you whereby if you break down your set of horses into subsets using this factor only, and you find that the rates of winning vary significantly between subsets, isn't that a factor you want to be using?

Here's yet of another way of saying it: You can use the odds as a handicapping factor WITHOUT destroying value (despite the inituition that it would) -- in fact, it can enhance value. That is, if you use the odds in the right way.


alysheba88 said:

"If your 4-1 horses are winning 12% of the time then your line is wrong."

Again, we're only talking about a subset of the 4-1 horses here. I'm telling you it is possible to identify WHICH subset of your 4-1 horses will win at a lower rate and to make them 12% horses like they should be.

I am also making the assertion that anyone's line *WILL* "be wrong" (or can be improved upon with a simple adjustment so it will be "less wrong", anyway) if they do not consider the toteboard odds.

[Dave Schwartz]

GT,

Well said.


When answering the question, "Am I creating an odds line that represents reality?" the answer is not "no" or "yes," or zero or one. Aristotlian truth does not apply here.

The answer, as you have alluded to in your previous post, lies between 0.00 and 1.00. Our goal (at least yours and mine <G>) is to move our "reality truth score" towards "1.00" and away from "0.00."


You and I should take this show on the road. <G>


Dave

PS: For those who might be interested, GT and I had a very long, late night telephone conversation last night. It is very rare for me to find someone to talk with that knows as much or more about A.I. than I do. To say that GT is well read and knowledgeable, would be a gross understatement.

We go about things in a different way; he is more statistically oriented than I, while my leanings are more towards physical science. Yet I found we had a lot in common. I think we both got a lot out of the conversation. (And I forgot that I called him back; I got the phone bill. <G>)

[sjk]

I make odds lines for the races I play and had never thought about using a hybrid odds line until I read about it on the board.

It is also my expreience that for any fixed price on my line, the chances of winning are greater for the lower priced horse on the board than for the higher priced horse on the board.

However, I think it is possible to use a hybrid line without knowing it.

When comparing my odds line with the odds offered on the tote board, I look for a premium (I use 75%). That is I look for

tote odds>1.75*my odds

If I form a line which is half my odds and half the tote odds, the above is the same as requiring

tote odds>1.27*half/half odds

By changing the 1.75, I am using the board odds to a greater or lesser extent. If I use a variety of bits of information about the race, horse and odds to modify the 1.75 factor, it has the same effect as creating a non-linear hybrid line.

[VetScratch]

Game,
No, it is not a put-on, and my guarantee is not a joke. Remember, I'm talking about looking at a group of horses where you have assigned the identical probability to each of them. So they *should* win at a flat rate if your probability is correct, right? If I break them up by any of the handicapping factors that you DID CONSIDER, then they probably would win at a constant rate between groups. Right? But here we have the odds, probably the most powerful factor of all, and you want to ignore it.You persist in assuming handicappers didn't start where you are concluding your arguments. The goal all along has been to produce odds that correlate with actual performance. You keep insisting that we are missing something, and I keep insisting that our initial objective and only means of measuring success is what you claim we are missing.

By "flat-rate" I interpreted you to mean the common stuff we see advertised, along the lines of "55% winners from top two selections." This has nothing to do with true handicapping or how success should be measured.

Who are these "handicappers" who have been measuring the accuracy of handicapped odds-lines by a different method than is used for actual parimutuel odds?

This is why I thought your guarantee was a joke. How could anyone challenge it? "Did you get any takers?" was my joke.

[GameTheory]

Originally posted by VetScratch
Game,
You persist in assuming handicappers didn't start where you are concluding your arguments. The goal all along has been to produce odds that correlate with actual performance. You keep insisting that we are missing something, and I keep insisting that our initial objective and only means of measuring success is what you claim we are missing.
We agree about the goal, and you (you means *you, VetScratch* -- many others here are getting it fine) are missing something.

Way of rephrasing this #128:

If you ignore real-time odds in making probabilities, then I can make more money with YOUR oddslines than you can.

[VetScratch]

Game,

I won't refute that, because you would be starting out with the best ingredients. :)

I am starting to look at your blending assertion for about 60% of all my database races for a multi-year span. Right now, I'm trying to find the takeout histories (that I have somewhere) so I won't have to calculate them to determine when they changed.

For obvious reasons, I do hope you have been right all along.

[GameTheory]

Takeout histories?

Just take 1/(odds+1) for each horse in a race, and renormalize so they sum to 1.0 to get public probabilities. Precise enough for the task...

[VetScratch]

Why not do it right? All my handicapped odds include T & B as do all the results odds. I'm going to be writing a new database so I can use existing programs against it. I want to see how I would have played the races (with new odds), as well as test the probabilities.

[VetScratch]

I see what you meant, but since I wasn't blending real-time odds, there are many races that include program change scratches. Since blending real-time odds in the future would imply catching all scratches, I am going to get rid of them in the new database.
Then in two steps:
(1) Re-handicap with no scratches left in any race. Run tests.
(2) Blend odds. Run tests again.

The handicapping part expects current takeout to be accessible for each track. The scratches exist where the original races were handicapped hours before program changes were available from any of the services.

I have been meaning to do step #1 anyway, before Karl started this damn thread. :D

[GameTheory]

The difference between the estimated probabilities after renormalization and the actual pool percentages will be very tiny -- it doesn't make a difference.

A random race pulled from my db:

Col#1 = Estimated pool percentage after converting odds to probabilities via 1/(odds+1) and renormalizing

Col#2 = Actual exact percentage of the pool bet

Col#3 = Difference


0.04721951028 0.04768405197 -0.0004645416897
0.298228486 0.2990844474 -0.0008559614077
0.08585365505 0.08651854641 -0.0006648913624
0.1349128865 0.1359560346 -0.001043148137
0.03845932964 0.0387756543 -0.0003163246643
0.3953261326 0.3919812653 0.003344867261



How are you going to do the blending? We didn't even really get into that since we were stuck on whether or not it was even a good idea. As I said, a simple linear adjustment probably will not work...

[VetScratch]

That's why I was only going to do it with about 60% of the races (excluding maidens, turf, marathons, etc.). Why wouldn't linear blending now be a good test since I am only dealing with races where the handicapped odds are just as predictive than the track odds?

Both odds set are 100% lines.

I couldn't find all the old takeouts in my archives since things have moved from system to system so often in recent years. I can compare the oldest archive with the current takeouts. This should show where changes have occurred since 1998.

[VetScratch]

BTW, for testing I have always exactly duplicated the tests described by Quirin (1979) because almost everyone has this reference in their library (or has a friend that does). My tests present results as illustrated in Quirin's Table 1 and Table 2 for Appendix A. If you have a problem with that, let me know.

[GameTheory]

Originally posted by VetScratch
That's why I was only going to do it with about 60% of the races (excluding maidens, turf, marathons, etc.). Why wouldn't linear blending now be a good test since I am only dealing with races where the handicapped odds are just as predictive than the track odds?


You've lost me again.

I don't have confidence that you have any idea what I'm suggesting you take a look at, but you good ahead with whatever it is you're doing and we'll see...

[VetScratch]

Originally posted by GameTheory
You've lost me again.

I don't have confidence that you have any idea what I'm suggesting you take a look at, but you good ahead with whatever it is you're doing and we'll see... What I mean is that for roughly 60% of the races, I already know that the handicapped odds test just as good against expectations as the track odds do. By eliminating the race types (roughly 40%) where this is not true, I was intending to "simply" blend accurate odds lines with track odds.

How would you suggest going about it?

Do you agree that Quirin's test methods are acceptable since they are almost universally understood?

Most of what I was starting to do was already justifed by my desire to have a scratch-free historical database. I want to re-handicap after eliminating the scratches, but what I am looking for there is not directly related to blending, but rather to adding automatic/online scratching and re-handicapping during a normal raceday. Right now I have to manually flag the program change scratches and select the races that need re-handicapping. Back when most of my database was created, I wasn't even playing online and had no way to do this at the track because I never trusted leaving a laptop laying around. Folks say they will watch it while you are away from your table, but you can't ask them to really be accountable when they have just as many reasons to go places as you do.

(Aside: I was intrigued by Storm Cadet reporting that he was getting a wireless laptop connection at SAR. Are reliable wireless laptop connections beginning to appear at tracks and OTBs? Often, it becomes a logistical nightmare just to find an electrical outlet, and that always inhibited me too. I tried lugging backup battery packs around in my bra, but was soon too sore to stay for the whole day anyway. :))

[sjk]

I took a look at my odds lines in light of the discussion. At first, it was disconcerting to find that to accurately predict win percentages, it was necessary to use a line which was 25% my line and 75% the public line.

On further reflection, the condition that the price beat this hybrid line by 20% (takeout) is mathematically equivalent to requiring that the price beat my original line by 75% which is exactly what I do.

I did notice that for horses which I give less than a 15% chance to win, the proper line combination was more like 20/80. So it makes sense to require a 100% edge on these in the future.

Also need to take a look at my bet size calculation. I am probably overbetting some races where my actual chance to win is pretty slim (hence the 30 race losing streaks).

Thank for the good ideas.

[JustMissed]

QUOTED FROM JOE'S IMMUTABLE LAWS:

THE IMMUTABLE LAWS OF HANDICAPPING--PART 5
(So far in this series we’ve looked at weight, layoff horses, first-time starters, asking a horse to do something new and repeaters.)

6---ALWAYS BET OVERLAYS AND AVOID UNDERLAYS.

I can’t begin to tell you how many articles and books I’ve read over the past 40 years urging us one way or the other to always get a fair mutuel return for every one of our $2 investments. Some refer to this as “value”. Most writers inevitably end up at one point in their discourse by asking their reader a question. It usually goes something like this: “Is the horse that you are about to bet mutuelly fair in return and worth a bet, is it overlayed and really, really worth a bet, or is it underlayed and to be avoided at all costs”?

And how are you supposed to answer a question such as this? These same writers ask you yet another question. If goes like this: “If this same race were run 100 times, how many times would your selection win”? They continue with: “If your horse is going off at even money, he’d have to win 51 of them out of 100 before you could make any money”.

Some authors additionally offer some kind of a table or graph with odds from 1-5 all the way up to 50-1, that shows you how many wins out of 100 same races that you’d need to keep your bankroll in the black. In plain English, if you think that your next bet would win at least 34 “like” races out of 100 “like” races, then you can accept 2-1 or better today, but nothing less!

Should you think that he would only win 26 “same or like” races out of a similar 100 race sample, then you would have to get at least 3-1 or better, but nothing less----and so on until the odds hit 50-1 on their chart.

There is one MAJOR flaw in their approach and one that they can’t “dance around”! They lead their readers to believe that the “same race” can actually be run 100 times sometime in the mythical future.

This “value-thinking” and similar concepts are so far from the “reality” that you actually encounter on track everyday, that it’s almost laughable to think of any races being the “same”. Not only can’t 100 exact same races be mythically run in the future, not so much as one “same race” can be run in the future! That’s right, I said to forget about a mythical 100 “same race” sample, because it is just that--totally mythical in every sense of the word and never to happen!

No 2 races are ever exactly alike, nor will they be in the future!

Every race is not only different, it is very unique in its composition. Various class levels increase and decrease, post positions change, horses are added, horses are deleted, jockeys move from one horse to another, the composition of the dirt or turf surface changes due to moisture and/or normal maintenance, the horses themselves physically look different from race to race etc. I could go on and on, but what’s the point? The above examples would convince any “doubting Thomas” that there are more than enough changing scenarios to render each race clearly unto itself.

What develops the first jump out of the gate until the very last when crossing the wire in no way ever exactly duplicates another race. Even if you actually had the same exact participants from another race and all the post positions were the same with the same jockeys and weights the same etc., would the race itself actually be run the 100 times the “same” way. No, in fact I doubt you could so much as duplicate 1 out of 100, let alone every one of those 100 “same” races that you needed for that money-management system.

That makes their “value-thinking” of getting this or that price for any wager based on 100 “same races” totally irrelevant in today’s marketplace as well as 40 years ago. This “immutable law” was total nonsense then and it is still total nonsense. Races simply aren’t run exactly the same way 100 times-----period!

Every race doesn’t have to have “alleged value” or a “value” horse. With tenure in this great game, there are times when you positively know that the price they are offering on a specific horse is not only generous or overlayed, you feel like you are stealing before you buy your tickets, as well as in the post race when you enjoy your personal “gloat time”.

Allow me to ask you a question. If you literally thought that the only way a certain horse could lose this upcoming race was if his jockey fell off him, would walk away from 4-5? And before you answer, hear the question once more stated a bit differently. Would you take 4-5 on a horse who literally could only lose the upcoming race by losing his jockey?

If you answered that you wouldn’t take 4-5 in that exact scenario or never take 4-5 in any other scenario, then you’ve probably never had a winning year in your entire horse playing life. That 4-5 situation occurs more than you might think with today’s anemic fields and if you’re not betting these low-priced overlays, not only are you missing the boat, you were never close to getting aboard to begin with.

Here’s a case in point---a bit simplistic, but I know that everyone will understand plain English. We all know the winning ways of the “Kid” Dollase trained El Cielo, at least those who play Southern California circuit. Place this monster on the Santa Anita “downhill” 6 ½ furlong course and he’s virtually invincible, as evidenced by his life time record down that tricky course where runners have to make 4 lead changes.

It’s 11/05/01 and we’re at Santa Anita where he’s entered to run down the hillside 6 ½ furlong course in the Grade 3 Morvich---a race that he had won last year. He looks “flawless” in the paddock. Here’s his “hillside” record when running at 6 ½ turf panels:

Starts:11 Wins:8 Seconds:2 Thirds:0

I might mention that the only time that he wasn’t 1st or 2nd in his prior 11 lifetime attempts down the hillside was the time his jockey actually did fall off him!

How’s a high 6-5 ($4.60) look for this unbelievable runner? That’s what he paid when positing his 9th win in 12 attempts! Oh, incidentally, he set the track record that afternoon, so he does in fact literally “own the course”! It’s a shame I didn’t have my car title with me. I would have literally tried to put that thru the mutuel windows as well.

Based on one of those odds tables we spoke about, 2-5 would have been acceptable to show a profit. But you don’t need some “Mickey Mouse” odds/value-table. Taking a high 6-5 on El Cielo wasn’t a hard decision on 11/05/01---it was the only decision! Stay focused every day. You’ll know when 4-5, even money, or 6-5 is “right” and you won’t need an odds table!

HOW ABOUT THAT LAST PARAGRAPH,

JUSTMISSED:)

[BillW]

Wonder if ole Joe had his car title on Azeri today? :D

[JustMissed]

Originally posted by BillW
Wonder if ole Joe had his car title on Azeri today? :D

Naw, I think ole Joe had Bobby Frankel's Perfect World in the 5th and decided to quit for the day.

You know ole Joe always likes those early turf races!
;) ;) ;) ;)

JustMissed

[Dave Schwartz]

JM,

>>Naw, I think ole Joe had Bobby Frankel's Perfect World in the 5th and decided to quit for the day. <<

And Karl has been teaching him his strategy for getting the money and going home. Could be a life changer for Joe. <G>


Dave

[Hosshead]

Just Missed - Absolutely no two races are run alike!

All horses that Win, have "Value". I can spot play 1/1 horses and hit more than 50% winners. Most of these are bet down below ML, as they should be. In fact if they're not, then I worry!

The name of the game is Profit. Any way you can get it. If you can bet longshots and come up with a 30% (or whatever) ROI, fine. If you can get the same profit betting even money horses, fine. For me, I'd rather get the same profit from the lower odds group. That way if you miss betting a race, or go on vac., you won't miss that longshot that'll take you 2 weeks to make up!

In your example with El Cielo: Even if he doesn't win, I know that I can still hit over 50% of this odds plays. As long as you know approx. what % winners you can hit, your OK. Also, when I like a horse,- that means I think he's going to win. I don't say he looks like he has a 2/1 chance, so I won't bet him unless he's over 2/1. If I really like him then he always looks at least 1/1 to me! So a 100% profit on that race, looks like "Value" to me. Likewise, so does a 20%-30+% ROI overall. However I'm talking about flat betting, Not the farm.

[JustMissed]

Hey Bill:

One fellow on this thread posted that he didn't want to sound arrogant but he liked 6-1 horses that were value plays.

Another fellow on another thread posted that he threw out horses that were 6-1 or less, mechanically no less.

I guess if they were Cal players they didn't get a whole lot of action today as SA had only one horse that paid better than 5-1.

Maybe these ole boys took the day off and played the football pools, or maybe someboby ought to open up the door around here and muck out the BS.

Laughing my ass off,

JustMissed

[JustMissed]

Originally posted by Dave Schwartz
JM,

>>Naw, I think ole Joe had Bobby Frankel's Perfect World in the 5th and decided to quit for the day. <<

And Karl has been teaching him his strategy for getting the money and going home. Could be a life changer for Joe. <G>


Dave

Hey Dave, I didn't say he went home, I said he quit betting for the day. I'm sure Joe hung around and made trip notes on the rest of the races for future reference.

Who knows but maybe he noticed a little gimp gate in old Azeri.

JustMissed

[Dave Schwartz]

JM,

Right you are. My mistake.

Of course, that probably means he is working too hard. <G>



Dave

[karlskorner]

The racing of horses isn't a "game" you have to stay to the end to see who wins.

Stay "alive" at 9/5, take the money and go home.

[VetScratch]

JustMissed.
Originally posted by JustMissed
Hey Bill:

One fellow on this thread posted that he didn't want to sound arrogant but he liked 6-1 horses that were value plays.

Another fellow on another thread posted that he threw out horses that were 6-1 or less, mechanically no less.

I guess if they were Cal players they didn't get a whole lot of action today as SA had only one horse that paid better than 5-1.

Maybe these ole boys took the day off and played the football pools, or maybe someboby ought to open up the door around here and muck out the BS.

Laughing my ass off,

JustMissed Many of these personal screening methods may sound contradictory but might make sense if we knew more about exactly where, when, how, and why they were being used. Most posters simply fail to explain why they play as they do.

If we grant that a player actually does a good job of assessing value (finding plausible overlays), then screening becomes a wagering issue rather than a handicapping issue. Most screeners seem to follow rules that apply to the high and/or low end(s) of the probabilities/odds spectrum. Their purpose is usually to avoid losing streaks.

Some screening methods make no sense in the long run, but they may stave off losing streaks that would bust players who employ overly aggressive wagering schemes.

True overlays at the high end of the odds spectrum are often bypassed to raise overall win percentage (i.e., avoid prolonged losing streaks) even though nothing may refute their long-term profitability on a statistical basis.

Most low odds screens are applied at the race level instead of the horse level. The idea is to pass certain races.

Assume that there is a plausible overlay going off at 8/1.

Many overlay players will pass any race that has a horse going off at less than a "losing proposition" threshhold, which they usually assert is somewhere between 7/5 and 2/1. Probably 3/2 or 2/1 are the most commonly used threshold values. Their logic is based on parimutuel studies which reveal that the population of horses going off between roughly 3/5 and 2/1 is predominated by "losing propositions."

What are "losing propositions?" Horses that win more often than their probabilities predict but still fail to return a profit (on account of takeout and breakage). In other words, these are horses you cannot bet "on" or "against" because they statistically outperform their odds but not to the extent that they beat takeout and breakage to show a profit.

The logical assertion goes something like this, "I have an overlay candidate, but the race also contains one of those "losing propositions" that is statistically liable to outperform parimutuel expectations while failing long-term tests for profitability. As a result, I will pass the race because I can't bet "on" the losing-proposition chalk but neither does it make sense to bet my overlay "against" the losing-proposition chalk.

Most "expert" proponents of value play have recommended some variation of low-odds race screening as a way of optimizing value play wagering tactics.

[karlskorner]

Charles Carroll ( Handicapping Speed ) a quote:

"There is clearly no shame in 6/5 odds --- or any odds---that you
WIN. The concern with low odds is there is also a calcuated risk. and many horse players, whether they are handicappers or odds players, have set odds that they will not bet below. In general, handicappers seem to set these higher than odds players, who for the most part, are looking for modest but virtually risk-free "investments"" end of quote.

First and foremost you have to "WIN" and then look at the tote board to see how much you won. I have never met an unhappy handicapper who wagered on a 4/1 horse that after the race only payed 3/1.

[alysheba88]

And i have yet to meet a person who is profitable getting back 3-1 on a 25% probable event. :)

[JustMissed]

I nicked this from the Turfpedia.com site. I believe this is Jim Quinn's stuff but will stand corrected. My comments follow the quote:

YOUR SKILL
If you know your win percentage and the average odds on your winners, you can estimate your winning edge. Your edge is equal to your win percentage multiplied by your average odds on winners, and that product minus your loss percentage.

The amount of money you can expect to win is equal to the amount you bet during the year multiplied by your edge. So if you have a 22% edge on the game and you bet $100,000 during the year, you can expect to win $22,000. Which is a nice profit from your hobby and an exciting game!

Here are the attainable goals for skillful handicappers:

Win% Aver. Odds Edge
30% 5-2 .05
33% 5-2 .15
35% 5-2 .22
40% 5-2 .40
The average odds are held constant at 5-2 because in reality they do not vary much on key horses to win.

The edge is equal to your return on investment (ROI). If your edge is 20%, that means for every dollar you bet, you can expect to get back $1.20. Investors, savers, and business men and women would be pleased to do as well.

Notice that handicappers who improve their skill by 10%(30% to 32%) have tripled their profits (.05 to .15). How nice! It pays to improve your handicapping.

THE AMOUNT OF MONEY YOU BET
Nobody can make a decent profit at the races by betting $2, $5, or $10 on key horses to win. Bet size should vary between $20 and $200, beginning at the smaller amount. Handicappers can increase the size as experience, proficiency, and confidence allow.
Because newcomers and casual handicappers invariably are curious to know, most leading handicappers and successful players bet $200 or thereabouts on the one, two, or three key bets to win they find on a day at the races. Some bet significantly more, and others bet less than half as much.

As a rule, handicappers can expect to find one, two, or three key bets to win each day, an average of two. If regular handicappers play 200 days in a year, the number of key bets to win will be 400 and the amount wagered will be between $8,000 ($20) and $80,000 ($200).
Now would-be handicappers have a basis for estimating how much money they might make when betting to win in any calendar year. Imagine yourself a successful horse player and plug yourself in at one or more of the handicapping and betting levels in the grid below.

$20 WIN BETTOR
WIN% ODDS EDGE BET BETS INVESTMENT PROFIT
25% 5-2 (.12) $20 400 $8,000 ($1000)
30% 5-2 .05 $20 400 $8,000 $400
33% 5-2 .15 $20 400 $8,000 $1,200
35% 5-2 .22 $20 400 $8,000 $1,760
40% 5-2 .40 $20 400 $8,000 $3,200

$200 WIN BETTOR
WIN% ODDS EDGE BET BETS INVESTMENT PROFIT
33% 5-2 .15 $20 400 $80,000 $12,000
35% 5-2 .22 $20 400 $80,000 $17,600
40% 5-2 .40 $20 400 $80,000 $32,000

In the age of full-card simulcasting, successful handicappers are playing multiple tracks each day, or more precisely, one-two-three bets at each of multiple tracks. Profits that formerly accumulated just at the local track can now accumulate at two, three, or four tracks each year. So dedicated handicappers, by all means, can multiply the potential profits resulting at one track from their skill by two, three, or four.

None of this has any relevance to exotic wagering. Successful handicappers also play the daily doubles, exactas, trifectas, and serial bets such as the Pick 3 and Pick 6 every day. The edge on any exotic wager depends upon the value expected from that specific betting situation, and is relatively unpredictable from year to year, or even from month to month, week to week, and day to day.

Effective strategies for playing the various exotic wagers are important, and they will be described in detail in this CD-ROM, but they are not related in any systemic way to the profits talented handicappers can expect year after year.

PRIME BETS TO WIN
A standard of excellence in handicapping is the ability to earn a 20 percent rate of profit on prime bets to win. A prime bet means a high-probability contender that has a clear decisive edge on the field and has been offered at a fair price. Numerous handicappers are talented enough to satisfy the 20 percent standard. But few do.

Handicappers who win 30 percent of their prime bets at average odds of 5-2 share a 5 percent edge on the game. For every dollar wagered to win, they receive $1.05 in return, a 5 percent profit.

A handicapper who can win as frequently as the crowd, 33 percent, but at 5-2 odds (the public averages 8-5), boasts a splendid 15 percent edge on the game. As Dick Mitchell has pointed out, by improving handicapping proficiency 10 percent, profits triple. This is the attainable goal all recreational handicappers should set sails to achieve 33 percent winners, average odds of 5-2, a 15 percent edge and profit margin. The reality for the casual majority is closer to 30 percent winners at odds of 5-2, a mere 5 percent profit margin.

The "edge" is important. A handicapper's true edge on the game suggests the optimal bet-size. It's the edge divided by the odds to $1.00. If the handicapper has a 5 percent and the horse is 5-2, the optimal bet is 5/2.5, or 2 percent.

To bet more is to bet too much in relation to actual proficiency, and guarantees a long-term loss. Overly aggressive betting is ruinous in horse racing. To bet less is to win at a rate of profit less than actual proficiency allows, a sad but not desperate circumstance.

VALUE BETS TO WIN
Value bets to win involve horses having a handicapper's reasonably good chance, but at conspicuously generous odds. The odds will be twice as great as expected, or longer than that. Instead of fair value, the handicapper perceives excellent value, perhaps outrageous value.

The horses do not figure convincingly, and many will not be first choices. Betting decisions remain equivocal, but the price exerts an undeniable pull
In certain situations, as in contentious races, the horses may be mild first choices offered upward of 8-1. That's value, as the horses can be expected to win at least once in eight attempts, probably more than once.

Or, the value bets may be unexpected overlays at 10-1 or better in contentious fields where lukewarm first and second choices have been overbet.
In predictable races, first choices may be severely overbet, leaving the handicapper's second or third alternative as an outstanding board attraction.

Or false favorites may be ridiculously overbet, such that a pair of reasonable alternatives go to the post at very attractive prices.

Or an angle horse in a relatively unpredictable race may be offered at irresistible odds.
Value bets are characterized by lower probabilities of winning than prime bets, so losing runs can be extended. If a horse has roughly a 20 percent chance (4-1), actual odds should be 6-1. A horse that has a 15 percent chance (6-1), should be 10-1 on the board, or better. A horse having a 10 percent chance (9-1) should be 15-1, minimum.
Bet-size should be a portion of the prime bet to win, not an equivalent amount. Stronger opinions can receive 50 percent of prime, weaker opinions 10 to 20 percent.
How will recreational handicappers know whether value bets to win are tossing real winnings, or losses?

It's strictly an empirical question. Handicappers must keep a record of the value bets and study the results. A year's worth of results on 8-1 shots, or greater, reveals the trend. Two year's data tell a fairly accurate story.END OF QUOTE.

How the heck can someone make a value judgement about what are or are not acceptable odds when they don't know squat about the bet size or win percentage?

Most know it is pretty easy to hit 50% or more, if you are damn selective about the races you play. Even at 7-5 average odds(which is damn low) a guy could hit 60 out of 100 races which would give him a 44% edge. At 20 $200 prime bets a week for 48 weeks a fellow could do pretty good for himself.

Before we get into a pissing contest, I hope you will write Mr. Quinn's formula inside your notebook and keep in mind that profitability for the horseplayer is dependent on three things:

X: Win percentage
Y: Bet Amout
Z: Payout

Good luck to you,

JustMissed




;)

[alysheba88]

Barry Meadow's win chart is another invaluable tool to use for win betting. Goes over same exact issues. What amounts to bet based actual odds vs your odds line.