trolls of the board take heart!!

[formula_2002]

I'm noticing more and more board members being called a "Troll".. As far as I know I was one of the first to be called a troll by the PA..

Confounded by what is ment by the tag I found this gem of a definition on the internet.

Let all board Trolls take comfort..it could be good to be troll!! :D

"There are many stories told about Trolls. For over a thousand years, different countries

have each had their own versions of the ancient Troll legends. Most stories were not

written down on paper, but had been passed down through the generations by telling

these tales at small gatherings of family and friends. So the folk tales would change

over the years and also differ from region to region. Some of these tales were eventually

recorded and the translations can be found in a variety of books on this subject.

There you will find many different types of Trolls described, as well as a good variety of

personal traits and appearances. In some stories, mostly from Norway, Trolls were

sometimes described as ugly, dim-witted and mean. But there are a lot of stories from

other countries such as Sweden and Denmark, which tell about Trolls that were wise,

kind and helpful...especially to animals and deserving humans."

[PaceAdvantage]

Joe, I've met you in person, way before PaceAdvantage.Com came into being. There is no way I would call you a troll. When did this happen? Please refresh my memory.

[formula_2002]

Joe, I've met you in person, way before PaceAdvantage.Com came into being. There is no way I would call you a troll. When did this happen? Please refresh my memory.

No harm done, I think is was back when I was frustrating a lot of people by making my point that it is impossible to beat a 15%+ take-out and that those who believe they ARE beating it, can not statistically prove it...

[GameTheory]

I'm the one who called you it. PA almost came around to agreeing with me since you refused to engage in any debate about your arguments and steadfastly refused to answer even the simplest of questions.

Anyway, this is what is meant by troll:

http://redwing.hutman.net/~mreed/warriorshtm/troller.htm

Nothing to do with fantasy trolls, but trolling, as in fishing -- throwing out bait, trying to get bites.

In any case, I no longer think you are a troll. You are a Ferrous Cranus. This is a much more accurate description for you:

http://redwing.hutman.net/~mreed/warriorshtm/ferouscranus.htm

[Dave Schwartz]

Formula,

>>>No harm done, I think is was back when I was frustrating a lot of people by making my point that it is impossible to beat a 15%+ take-out and that those who believe they ARE beating it, can not statistically prove it...<<<


You are certainly not a troll.

But...

When you make comments like the one above, you make yourself appear foolish.

So, when someone makes a living at the races, as several people that I know of on this board do, if they can't prove it statistically then they aren't actually beating the races?



See, that is shear stupidity at its best and wears on the most of us. It is very tiresome.

How, then, are they making their living? Is it a mind-over-matter thing? Because if it is, I'd really like to learn it.


Dave Schwartz

[Pace Cap'n]

Well, Joe, since its been brought up...

Having read your posts for a couple years or so, I cannot for the life of me comprehend your compulsion to enlighten us poor horseplayers. I can only surmise it is the result of the liberal bent to save everyone from themeselves.

When people visit a discussion board, most likely they do so for the purpose of education and affirmation. When a member continually posits a contrarian view, it is often perceived as trolling, and frequently results in much heated debate (ok, flaming).

It is only due to the derived benefits from your projected personna as a nice guy that no more slings and arrows have been hurled in your direction than you currently perceive.

It might be worth mentioning that the authorities frown upon the promulgation of nefarious tax schemes.

[sq764]

No harm done, I think is was back when I was frustrating a lot of people by making my point that it is impossible to beat a 15%+ take-out and that those who believe they ARE beating it, can not statistically prove it...
Did you ever, in the deepest reaches of the universe, think that POSSIBLY you just aren't a very good handicapper and/or bettor?

I am not trying to be rude at all, I am just pointing out that to think no one beats the takeout in this game is just ludicrous thinking, and slightly arrogant in my opinion.

I could never be an astronaut, but I believe I have seen others accomplish the feat. I am not so delirious to think that if I can't do it, no one can..

[Tom]

Nop computer study can replicate real life betting. In real life, you make descisions that no computer can make. You make bets on certain situations and pass others, you don't bet every single horse to come along.
You will lose money if you bet every single 2-1 shot, but no one bets every single 2-1 shot.
Joe, if you can eplain logically, in words, how you can make that statement, I will listen.

[formula_2002]

Nothing to do with fantasy trolls, but trolling, as in fishing -- throwing out bait, trying to get bites.

http://redwing.hutman.net/~mreed/warriorshtm/ferouscranus.htm

Your reference’s speaks volumes to your character.
A scurrilous attack. Get a new "line" of work

Better you should prove me wrong , in a mathematical way, that the take out cannot be beaten.

Try not to use too many words. Numbers and formulas would do just fine.

I will offer the same to any here.

No name calling, no "I know this guy" kind of stuff, just an elegant, mathematical work, to prove your point.
Other then that, no need to reply. I will understand your displeasure (in me) and or your inability to properly prove your point.

Beyond that, enjoy what you do and go on with it.

[formula_2002]

I could never be an astronaut, but I believe I have seen others accomplish the feat. I am not so delirious to think that if I can't do it, no one can..

Those astronauts walked on the shoulders of many a fine thinker, who themselves, could not go into outer space.

Space travel is made possible by many things, the least of which is based upon math. If the math is not there, then here we stay. Sometimes we may get lucky and do the feat first and the math follows, but the math was always there..

[sq764]

Here's simple math for you. If for every $2 you invest at the track for the year returns >$2, you profit for the entire year. There's your formula, put it on a sticky for yourself.

[formula_2002]

Here's simple math for you. If for every $2 you invest at the track for the year returns >$2, you profit for the entire year. There's your formula, put it on a sticky for yourself.

That is about as deep as it goes with you..
Stop trying so hard...There is no need for you to go any further..I understand completely..thank you so much..

[sq764]

That is about as deep as it goes with you..
Stop trying so hard...There is no need for you to go any further..I understand completely..thank you so much..

No, it's called simplicity. You overthink and you are under the belief that everything must have a mathematical solution. News flash - it does not!

It's solid handicapping coupled with intelligent wagering structure that leads to positive ROI.. It can be done and is done.. How do you think people can make their living off of horseracing? It IS being done right now.

You need to step back and identify your shortcomings and realize that this is being done now, just not by you.

[GameTheory]

Your reference’s speaks volumes to your character.
A scurrilous attack. Get a new "line" of work

Better you should prove me wrong , in a mathematical way, that the take out cannot be beaten.

Try not to use too many words. Numbers and formulas would do just fine.

I will offer the same to any here.

No name calling, no "I know this guy" kind of stuff, just an elegant, mathematical work, to prove your point.
Other then that, no need to reply. I will understand your displeasure (in me) and or your inability to properly prove your point.

Beyond that, enjoy what you do and go on with it.


Don't even try it -- it has been your refusal to re-examine your math (even at the most basic level) that has earned you this rep. I have never tried to argue with anything other than empirical evidence, and certainly haven't done any "name-dropping".

[boxcar]

Sounds to me that Form, being a loser at the track, wants desparately to project his personal results onto everyone else. "Mathematical formula" for this is very simple:

Misery = strong desire for lots of company.

Boxcar

[JustMissed]

I am certainly no expert on data bases, especially horse race results, but would like you to answer this question.

Is it possible to obtain a valid answer when querying a data base for information that is not included in that data base?

Not sure that made sense but let me use two examples.

1. If I asked you which post position resulted in more winners at Delaware Park in 2004, the 2nd or 10th in 6F sprint races. You could give me an answer and I believe it would be correct and probably be useful to me with my handicapping.

2. Now, if I asked you to query your data base and tell me which track has the better popcorn, Tampa or Delaware Park, I don't believe you could give me an answer because popcorn info is not included in your data base.

Now, if I wanted to prove this statement to be 'true or false': "it is impossible to beat a 15%+ take-out and that those who believe they ARE beating it, can not statistically prove it...", you would not be able to come to a conclusion because the info needed is not included in your data base--that being a particular bettor's betting history and/or pattern.

There must be dozens of different betting patterns in use, none of which are included in anyone's data base except for an individuals own records.

I suspect that even if you gave two players the same amount of money to bet the same races and even gave them the order of finish for each of those races-each player would walk out with different amounts of money.

Anyway, I think it is real funny in a sad sort of way that Joe has spent all this time and effort to prove something that cannot be proved simply because the adequate data is not included in his data base.

JM

[formula_2002]

To prove a 10% profit at 15% take-out and 1% chance of error in the win pool.



first col is the incremental average odds

second column is the # of playable horses.
(this total decreases as the % profit or the % take-out increases).

third column is the % winners

fourth column is the total # of horses you would have to review to find the playable horse.

fifth column is the accumulated totals


1 95 55.00% 1288 1288
2 190 37.00% 1668 2956
3 270 27.50% 2259 5214
4 350 22.00% 3475 8690
5 435 18.30% 5241 13931
6 525 16.00% 7557 21487
7 600 14.00% 9812 31299
8 700 12.00% 13024 44323
9 750 11.00% 16011 60334
10 850 10.00% 19824 80158
11 900 9.00% 24148 104306
12 980 8.50% 30250 134556
13 1200 7.80% 40341 174898
14 1250 7.30% 49541 224439
15 1250 6.90% 51669 276108
16 1300 6.50% 62269 338376
17 1450 6.10% 78018 416394
18 1500 5.80% 82505 498899
19 1600 5.50% 96700 595599
20 1650 5.25% 207292 802891
17845 802891


So if you wanted to prove you could make a 10% profit at say AQU, between the odds range of 1 and 10, you would only have to handicap about 80,000 horses

Of course if anyone can come up a a more accurate set of NUMBERS, if would certainly review them..

[JustMissed]

Sorry Joe, but you have been busted.

No where in your table or data do you present any information as to a particular player's bets. Without this information no conclusion can be made as to whether one can or cannot beat the takeout.

Nice try Joe but the scam is up.

JM :D

[formula_2002]

Sorry Joe, but you have been busted.

No where in your table or data do you present any information as to a particular player's bets. Without this information no conclusion can be made as to whether one can or cannot beat the takeout.

Nice try Joe but the scam is up.

JM :D


Oh boy.. Mitchell, Fabricand, Quirin and about 100 or so horse racing studies have similar "scams".
If you can meet the conditions you can beat the take-out..go for it.

[formula_2002]

Don't even try it -- it has been your refusal to re-examine your math (even at the most basic level) that has earned you this rep. I have never tried to argue with anything other than empirical evidence, and certainly haven't done any "name-dropping".


Not a meaningful number or formula in your reply.
Typical GameTheory reply.

[formula_2002]

How do you think people can make their living off of horseracing?


My judgement is the same way people make a living off of the lottery..luck.

[sq764]

My judgement is the same way people make a living off of the lottery..luck.
You need to find a new game to play. You obviously have been mentally defeated. You've got one thing right, YOU cannot beat the takeout.

[JustMissed]

Hey Joe, remember the Quayle-Bentsen debate:

"Quayle: "I have as much experience in the Congress as Jack Kennedy did when he sought the presidency." Bentsen: "Senator, I served with Jack Kennedy. I knew Jack Kennedy. Jack Kennedy was a friend of mine. Senator, you're no Jack Kennedy." (Omaha, Nebraska, October 5, 1988)"

Joe, I have not heard back from GameTheory yet but I don't believe Mr. Michell ever presented results based on data and the data wasn't even there.

Maybe you should take up Alchemy for a hobby, hehehe!


JM :D

[Show Me the Wire]

formula_2002:

The formating of your proof post made it difficult for me to read on my screen, too wide. So I apologize in advance if the answer to this question was contained in your posting.

Your refer to the "playable horse". How do you define playable horse in your equation? What a playable horse is, is the threshhold question.

I assume you are using a mathematical equation to define a playable horse to empiracally qualify wagers.

I am looking forward to your response.

Regards,
Show Me the Wire

perception is reality

[formula_2002]

formula_2002:

Your refer to the "playable horse". How do you define playable horse in your equation? What a playable horse is, is the threshhold question.


It would be your selection to win, using any criteria you wish.
The table is just a way of gauging your performance.
The table is telling you what % winners you need to be statistically significant at a 15% take out track while obtaining a 10% profit.

[Dave Schwartz]

Formula,

I broke bread a few months back with a man that offered me a different kind of proof of his winning performance.

First, he showed me his $2m apartment in Manhattan. Then he showed me his cars, his lifestyle and his generosity.

Oh, and then he showed me his tax return from the previous year that listed $2.3m in horse racing income (after losses).

Would that qualify as "proof" for you?

If that man offered to show you how he does what he does (he won't), would you "demand" your form of proof?


Regards,
Dave Schwartz

[Tom]

Joe is not trying to prove you cannot beat the take in ral life. He is proving that you cannot come up with a simulatin of numbers that have absolutley no basis in reality - becaseu they are not the real bets made by real people.
NO OMPUTER can think like a person, thus wtihthe human element left out of Joe's equations, he making a Type II error (or is it Type I???)
Joe is offering apples to disprove oranges. My betting records prove beyond any doubt I have been beating the take for years. I collect more that I bet every year.

[formula_2002]

Formula,

Oh, and then he showed me his tax return from the previous year that listed $2.3m in horse racing income (after losses).

Would that qualify as "proof" for you?

If that man offered to show you how he does what he does (he won't), would you "demand" your form of proof?

Dave Schwartz
Dave I'm very happy for the guy, but it's not the kind proof I'm talking about..

ps, he wouldn't have to show me what he does, just the ( price waterhouse etc.certified,) . incremental win pool odds results...he may be able to make more money for less work selling his book.

[CapperLou]

Joe etal:

I'll just add a few thoughts here----Yes, Joe; it is very hard to be profitable over the course of a year for the many reasons mentioned--but for me the biggest roadblock is the takeout.

In looking over my regular plays based on computer handicapping over last year I too could not "beat the races." The net loss on regular plays was 7.8%.

As Gordon Pine always says---it's the takeout stupid--and I must agree with him on my regular plays.

However, on my trainer and trainer/jockey plays, I had a 23% profit. Now that may sound great, but the reality is that those plays were about 1/3 the total plays.

What I have learned is to play only CA & NY tracks with regular plays and make half or one quarter sized plays at the others I play. Ellis Park ruined me this past year---Bejarano--!! Will not play Ellis again. CA & NY takeouts make a big difference for me---probably most of you too.

This year I'm going to concentrate on emphasizing my trainer plays and de-emphasize my regular plays and try to make most of those at only NY & CA!!

We will see what happens-----again I should mention it's all about the takeout IMO--at least in my case.

Joe, in closing, I would tell you to forget about craps--I have been going to Vegas for decades and have met them all and the way things are set up from a legal standpoint in Nevada--you and no one else is going to be able to play craps or blackjack on a continuing basis if you start winning (they will politely ask you to leave my friend). Everything has changed in recent years as most of you know!!!

All the best,

CapperLou

[sq764]

Dave I'm very happy for the guy, but it's not the kind proof I'm talking about..

ps, he wouldn't have to show me what he does, just the ( price waterhouse etc.certified,) . incremental win pool odds results...he may be able to make more money for less work selling his book.

Sounds more like you dont want proof, you want to know their method for yourself..

[formula_2002]

Sounds more like you dont want proof, you want to know their method for yourself..

If you think that table I posted will give away anyone's handicapping method, then I should be making a fortune...

I have to go to the dentist. It's his turn to drill me!!

[JustRalph]

Formula,

I broke bread a few months back with a man that offered me a different kind of proof of his winning performance.

First, he showed me his $2m apartment in Manhattan. Then he showed me his cars, his lifestyle and his generosity.

Oh, and then he showed me his tax return from the previous year that listed $2.3m in horse racing income (after losses).

Would that qualify as "proof" for you?

If that man offered to show you how he does what he does (he won't), would you "demand" your form of proof?
Regards, Dave Schwartz

That's it Dave!

I will never invite you back to my pad!

[GameTheory]

I am certainly no expert on data bases, especially horse race results, but would like you to answer this question.

Is it possible to obtain a valid answer when querying a data base for information that is not included in that data base?

Not sure that made sense but let me use two examples.

1. If I asked you which post position resulted in more winners at Delaware Park in 2004, the 2nd or 10th in 6F sprint races. You could give me an answer and I believe it would be correct and probably be useful to me with my handicapping.

2. Now, if I asked you to query your data base and tell me which track has the better popcorn, Tampa or Delaware Park, I don't believe you could give me an answer because popcorn info is not included in your data base.

Now, if I wanted to prove this statement to be 'true or false': "it is impossible to beat a 15%+ take-out and that those who believe they ARE beating it, can not statistically prove it...", you would not be able to come to a conclusion because the info needed is not included in your data base--that being a particular bettor's betting history and/or pattern.

There must be dozens of different betting patterns in use, none of which are included in anyone's data base except for an individuals own records.

I suspect that even if you gave two players the same amount of money to bet the same races and even gave them the order of finish for each of those races-each player would walk out with different amounts of money.

Anyway, I think it is real funny in a sad sort of way that Joe has spent all this time and effort to prove something that cannot be proved simply because the adequate data is not included in his data base.

JM


Yes, that's right -- Joe has shown over and over you can't make money (in any odds range) by playing all horses. And then he makes the incredible leap of generalizing that result to any possible way of selecting horses. He has never explained or justified why he makes this leap. He also brings the names of Mitchell & Quirin into the discussion as if they would agree with him, when in fact they would be aghast at the tarnishing of their good reputations with this drivel. Joe thinks he is doing good statistics -- he is not. If he showed his work to any statistics professor anywhere, they would immediately point out his errors. That, coupled with the fact that those very same professors are in recent years increasingly abandoning the type of hypothesis testing he is *trying to do* (and failing to do correctly) because it is and always has been a mathematically defective method without solid grounding in theory. (Which is why it needs so many complex tables of "critical values" of statistical significance, etc, instead of just spitting out an answer elegantly -- any professor will also tell you that statistical significance and practical significance have no particular correlation to each other). It has always been somewhat of an "ad-hoc" science that nevertheless took hold (and is now losing hold) because Bayesian theory was not very developed either at the time. Probability theory & statistics are some of our newest sciences, and they are still changing at the very foundation. Joe is standing on a foundation of sand at high-tide. Now, I'm not going to try to convert Joe from his classical statistics to Bayesian statistics -- it would be good enough if he only did the classical statistics correctly as it is plenty enough to debunk his nonsense.

He keeps acting as if he has presented an argument that you can't beat the take, but he never has. His table of incremental odds ranges is irrelevant to the question (by itself). That is the *context* within which we handicappers must work in order to profit, but without any selection method we have nothing to test.

[formula_2002]

Yes, that's right -- Joe has shown over and over you can't make money (in any odds range) by playing all horses. And then he makes the incredible leap of generalizing that result to any possible way of selecting horses. He has never explained or justified why he makes this leap. He also brings the names of Mitchell & Quirin into the discussion as if they would agree with him, when in fact they would be aghast at the tarnishing of their good reputations with this drivel. Joe thinks he is doing good statistics -- he is not. If he showed his work to any statistics professor anywhere, they would immediately point out his errors. That, coupled with the fact that those very same professors are in recent years increasingly abandoning the type of hypothesis testing he is *trying to do* (and failing to do correctly) because it is and always has been a mathematically defective method without solid grounding in theory. (Which is why it needs so many complex tables of "critical values" of statistical significance, etc, instead of just spitting out an answer elegantly -- any professor will also tell you that statistical significance and practical significance have no particular correlation to each other). It has always been somewhat of an "ad-hoc" science that nevertheless took hold (and is now losing hold) because Bayesian theory was not very developed either at the time. Probability theory & statistics are some of our newest sciences, and they are still changing at the very foundation. Joe is standing on a foundation of sand at high-tide. Now, I'm not going to try to convert Joe from his classical statistics to Bayesian statistics -- it would be good enough if he only did the classical statistics correctly as it is plenty enough to debunk his nonsense.

He keeps acting as if he has presented an argument that you can't beat the take, but he never has. His table of incremental odds ranges is irrelevant to the question (by itself). That is the *context* within which we handicappers must work in order to profit, but without any selection method we have nothing to test.


show me the numbers!!

[GameTheory]

show me the numbers!!
See what I mean? All he can say is, "Show me the data", his stock answer when he doesn't want to answer a question. He posts some irrelevant chart and then asks for "data" to refute it.

Let's try this. Joe, could you please post an example of some "numbers" that would fulfill your requirements?

Two can play at your game, you know. Here is some data that proves you can beat the take:

600 / 700 = .857

Now, refute it! I dare you. No questions about my data are allowed. No explanation on my part is necessary. No justification is needed. No theory called for. I will respond to nothing but data. DATA ONLY!

[Dave Schwartz]

Joe,

You obviously are totally convinced that winning is impossible. There is no point in further discussing this with you.

I wish you well.


Regards,
Dave Schwartz

[PaceAdvantage]

Come on Joe. You can do better than that. GameTheory has taken the time to respond to you in a rational and thoughtful manner. You could at least do the same, and continue to make this thread quite interesting.

Can't you give GameTheory more than a typical '46zilzal' three word response?

[CapperLou]

Game Theory: Great Post!!!!

Dave: I agree with you!!!

Joe: I think I made my little point that one can beat "the take"--there are several ways to do it and each person who is able to is not going to give away their method of wagering etc.

You can see that I lost in one area, but percentage wise profited in another so that at the end of the day as they say---the year ended on the plus side.

Nothing else to say here--but if you know you cannot win at the races--don't try your luck at the craps table Joe---it will cost you even more!!!!

All the best,

CapperLou

[GameTheory]

Come on Joe. You can do better than that. GameTheory has taken the time to respond to you in a rational and thoughtful manner. You could at least do the same, and continue to make this thread quite interesting.

Can't you give GameTheory more than a typical '46zilzal' three word response?
Don't give me too much credit about acting mature here -- I lost all hope of having any rational discussion with Joe years ago (literally, years). And I've generally left him alone during that time, but since he started this thread just to draw attention to himself, I couldn't resist the urge to identify him as the dreaded Ferrous Cranus, impervious to reason and totally unmoveable no matter how foolish he looks.

Actually, he is looking more and more like a Target:

http://redwing.hutman.net/~mreed/warriorshtm/target.htm

[JustMissed]

Thanks for taking the time to respond to my question. Very thoughtful response and I appreciate it.

It is as I suspected. You cannot draw a valid conclusion from data that is not presented.

Kinda like those court room shows where the judge instructs the jury that they are not to consider any facts or evidence not put forth in evidence.

Old Joe maybe should be put into 'Horseplayer jail' for putting forth a conclusion about other's ability to be profitable at the track when he doesn't have any evidence or facts to support that conclusion.

Could be that Joe is just pulling everyone's leg or maybe he really is psychotic. I don't know.

JM

[Pace Cap'n]

I don't know either but it sure is getting tiresome.

[Show Me the Wire]

formula_2002:

Absent a defined selection method as Game Theory stated:

"That is the *context* within which we handicappers must work in order to profit, but without any selection method we have nothing to test."

In any terms without a defined population (playble horses) I cannot accept your assumption about beating the take.

If your selection process does not perform at a profit based on your tables, you cannot logically extend your experience to all selection methods.

Best wishes.

Regards,
Show Me the Wire

perception is reality

[DJofSD]

Certain aspects of this thread remind me of the courtroom scene in "The Caine Mutiny" where Bogart is absolutely convinced he can prove mathematically that some one stole the strawberries.

DJofSD

[Dave Schwartz]

DJ!

A voice unheard from in awhile!

How are you? How is life in SD?


Regards,
Dave Schwartz

[formula_2002]

formula_2002:

If your selection process does not perform at a profit based on your tables, you cannot logically extend your experience to all selection methods.


perception is reality
OH yes I can

Lets do it this way;

Would anyone here disagree with me that in order to make a 10% profit playing 3-1 horses at a 15% take-out track (or any track) you must win 27.50% of your bets?

hint(edge=(odds X win%)-(1-win%))


Next, any data base user here can verify the following for themselves;

At a 15% take out track, horses in the odds range of >=2.5-1 through <=3.4 win appox 21% of the time
(my data base of 6555 such horses win 21%, another data base of 201 such horses win 20.39%).

Only those who agree need reply :cool:
I will go on from there for their eyes only. Everyone else is dismissed

GT, since you wrote;
"Let's try this. Joe, could you please post an example of some "numbers" that would fulfill your requirements?"

pay attention

[hurrikane]

oh boy.....here I go being stupid again

ok, joe,
i'll agree. now what?

[hurrikane]

on the other hand. Never mind,

I've been through this a million times with you. You have not found a way select more than 28% winners a 3-1. That's all. the rest of the stuff is just noise. YOu have spent enormous amounts of time and effort to arrive at the obvious.

I just glad you got over the 'go to my website' thing.

For me that seems like all that can be done with you.

good luck and have a happy life.

[Pace Cap'n]

So, what are the results if you only bet 3-1 horses whose jocks are dressed in blue and white?

Or just the gray ones?

Or maybe, just those who are exhibiting a positive form cycle, have run competively against similar company, with good pace & speed figures, on a surface of their liking, going at their best distance, and returning the good jockey who's won with him before?

How do these perform?

[formula_2002]

on the other hand. Never mind,

You have not found a way select more than 28% winners a 3-1.

You are 100% correct. I never said I did

The table is saying, that to prove that the given conditions were significant, you would have to win 28% of the time..
(strawberries not included)

[formula_2002]

How do these perform?

Only "yes" replys please...else if you wish, start a new thread.

[Pace Cap'n]

Sorry, didn't mean to interject logic.

[sq764]

OH yes I can

Lets do it this way;

Would anyone here disagree with me that in order to make a 10% profit playing 3-1 horses at a 15% take-out track (or any track) you must win 27.50% of your bets?

hint(edge=(odds X win%)-(1-win%))


Next, any data base user here can verify the following for themselves;

At a 15% take out track, horses in the odds range of >=2.5-1 through <=3.4 win appox 21% of the time
(my data base of 6555 such horses win 21%, another data base of 201 such horses win 20.39%).

Only those who agree need reply :cool:
I will go on from there for their eyes only. Everyone else is dismissed

GT, since you wrote;
"Let's try this. Joe, could you please post an example of some "numbers" that would fulfill your requirements?"

pay attention

Why do you insist on including examples that rely on the bettor playing EVERY SINGLE horse in that odds range? That is ridiculous..

[formula_2002]

Why do you insist on including examples that rely on the bettor playing EVERY SINGLE horse in that odds range? That is ridiculous..

I'm comparing guys like you who may be able to hit a 3-1 shot 28% of the time to guys like me who can hit it but 21% of the time..I'm not saying that you bet every 3-1 shot, just the ones that allow you to win 28% of the time. If indeed you can hit a 3-1 shot 28% of the time for as many races in the table, that would be significant. And, if you can do meet the significance requirement for the rest of the table, then you are my hero...

Is that clear?

[hurrikane]

ahh.....

joe, you have just defined the differnece between the winners ..and you.

I'm just curious....shouldn't you be changing the name of your website.

globalmediocrepicks or global-losing?


good luck

[formula_2002]

ahh.....

joe, you have just defined the differnece between the winners ..and you.

I'm just curious....shouldn't you be changing the name of your website.

globalmediocrepicks or global-losing?


good luck

I think I'll change it to "Truth In Handicapping"

[sq764]

I'm comparing guys like you who may be able to hit a 3-1 shot 28% of the time to guys like me who can hit it but 21% of the time..I'm not saying that you bet every 3-1 shot, just the ones that allow you to win 28% of the time. If indeed you can hit a 3-1 shot 28% of the time for as many races in the table, that would be significant. And, if you can do meet the significance requirement for the rest of the table, then you are my hero...

Is that clear?

Joe, I think you just found the answer to you puzzle.. You can't run a reasonable test that involves playing EVERY SINGLE HORSE in a certain odds range. It takes some degree of rational play and discretion..

I don't think deep down you believe that NO ONE can beat a 15% takeout.. I think you really are doubting is that there is a flat, mechanical, black box method that can beat a 15% takeout.. If this is the case, you may very well be true..

[formula_2002]

Joe, I think you just found the answer to you puzzle.. You can't run a reasonable test that involves playing EVERY SINGLE HORSE in a certain odds range. It takes some degree of rational play and discretion..



I get it.

The ghost of my third grade teacher sent you to haunt me..

Some body help him out...

[sq764]

Joe, I think the problem is that YOU don't get it.. If you'd put your ego on the shelf for a minute, you would listen to what people are telling you and you might learn a thing or two that would change your defeatest attitude..

The game can be beat, just not in your numerical ruled world.. Numbers are a big part of this game, just not the be all-end all.. Why do you not understand this?

[formula_2002]

Joe, I think the problem is that YOU don't get it.. If you'd put your ego on the shelf for a minute, you would listen to what people are telling you and you might learn a thing or two that would change your defeatest attitude..

The game can be beat, just not in your numerical ruled world.. Numbers are a big part of this game, just not the be all-end all.. Why do you not understand this?

LET ME TRY ONE MORE TIME


We both go to a simulcast.. You did your home work and I did mine.. That day there happend to be 2258 horses that went off at 3-1

You bet 270 of these beast and win 28% of your bets.
I bet 100 and win but 21%. Get it.

I didnt play every one of the 2258 horses.
At the end of the day 21% of the 2258 3-1 horses won.

I did as well as the entire betting public and you did SIGNIFICANTLY better.

[sq764]

LET ME TRY ONE MORE TIME


We both go to a simulcast.. You did your home work and I did mine.. That day there happend to be 2258 horses that went off at 3-1

You bet 270 of these beast and win 28% of your bets.
I bet 100 and win but 21%. Get it.

I didnt play every one of the 2258 horses.
At the end of the day 21% of the 2258 3-1 horses won.

I did as well as the entire betting public and you did SIGNIFICANTLY better.
So your thesis is that if you bet 100 horses at 21% and I bet 270 at 28% that the game can't be beat? So... Because you cannot consistently hit a high enough win % to turn a consistent profit, this game can't be beat?

Wow, I don't even know how to respond to this type of thinking..

[GameTheory]

So your thesis is that if you bet 100 horses at 21% and I bet 270 at 28% that the game can't be beat? So... Because you cannot consistently hit a high enough win % to turn a consistent profit, this game can't be beat?

Wow, I don't even know how to respond to this type of thinking..Didn't you know it is impossible to stratify horses by anything other than odds? Didn't you know it is impossible to select of subset of the 3-1 horses than win at a higher rate than the whole group?

[sq764]

Didn't you know it is impossible to stratify horses by anything other than odds? Didn't you know it is impossible to select of subset of the 3-1 horses than win at a higher rate than the whole group?
As Tim Robbins says to the Warden, "How can I be so obtuse"

I don't know what I was smoking when I tried to rationalize without throwing a formula or some statistical graph in with my response..

[formula_2002]

So your thesis is that if you bet 100 horses at 21% and I bet 270 at 28% that the game can't be beat? So... Because you cannot consistently hit a high enough win % to turn a consistent profit, this game can't be beat?

Wow, I don't even know how to respond to this type of thinking..

Almost there.
3-1 shots represent about 9% of all the horses.

say 8 horse field x 9 races a day x .09=6.48, 3-1 horses per track day.

to find 2259, 3-1 horses you would have to handicap 348 track days.. a year's worth of races fo any single track.

Now , say you like the 10-1 shot. They represent 3.3% of all the horses per track day.
8 x9 x.033= 2.376 horses per track day


to find 19824, 10-1 horses (see the table) you would have to handicap 600,727 track days (that could be a life time)

My thinking is, that if you can indeed beat the races, you could not live long enough to prove it.

[formula_2002]

Didn't you know it is impossible to stratify horses by anything other than odds? Didn't you know it is impossible to select of subset of the 3-1 horses than win at a higher rate than the whole group?


Not according to these guys...They look at horse racing as a game that does not conform to natural odds.

A game where actual winners/expected winners can be >1 to the point where it is significant. Thats not what the data indicates to me.

[sq764]

Almost there.
3-1 shots represent about 9% of all the horses.

say 8 horse field x 9 races a day x .09=6.48, 3-1 horses per track day.

to find 2259, 3-1 horses you would have to handicap 348 track days.. a year's worth of races fo any single track.

Now , say you like the 10-1 shot. They represent 3.3% of all the horses per track day.
8 x9 x.033= 2.376 horses per track day


to find 19824, 10-1 horses (see the table) you would have to handicap 600,727 track days (that could be a life time)

My thinking is, that if you can indeed beat the races, you could not live long enough to prove it.
Ok, ready for this crazy crazy thought?? What if, you were a patient player and focused on demanding 3/1 on your horse.. Let's say you found 1 horse per day to play and you got 3/1 every time. Let's say, just for theory purposes, you hit 30% ofyour plays..

Let's also say you bet $100 to win on that horse each day.. For every 10horse cycle, you would be up $200..

Is this ridiculous to believe this is happening?

Would this be beating the races?

[GameTheory]

Almost there.
3-1 shots represent about 9% of all the horses.

say 8 horse field x 9 races a day x .09=6.48, 3-1 horses per track day.

to find 2259, 3-1 horses you would have to handicap 348 track days.. a year's worth of races fo any single track.

Now , say you like the 10-1 shot. They represent 3.3% of all the horses per track day.
8 x9 x.033= 2.376 horses per track day


to find 19824, 10-1 horses (see the table) you would have to handicap 600,727 track days (that could be a life time)

My thinking is, that if you can indeed beat the races, you could not live long enough to prove it.Ok, I'm going to foolishly act a direct question. Where do the numbers 2259 & 19824 come from? Why do I have to look at 2259 3-1 horses in order to find my 270 "playable" horses? Apparently, I need 76 (28% of 270) winners out of 270 plays to show significance at 3-1, is that right? If I have a sample of 2259 with 21% wins in the whole group, that means I've got a set 474 winners from which I can get my 76, right? So what's the 2259? Let's say I look at only 1000 3-1 horses (at 21% there are 210 winners in the group). Let's say I play 270 of the 1000 and win 76 times (28%). Is that not significant?

By the way, there were 15,700 10-1 horses (rounded off >= 9.5 & < 10.5) just in 2004. It would hardly take me a lifetime to get to 19824. But first I'd like to hear the explanation of why I need the 19824 to begin with.

[Steve 'StatMan']

The only ones that can win at the game will do things like find a subset of the 3/1 shots that can win 28+% of the time.

Those that can't make those distinctions will lose long-term at the this game until they they do.

They may be 3/1 shots that have a speed or pace advantage in a race. They could be 5/2 shots that have a very significant advantage in those same categories. Or perhaps an troubled trip or a trainer angle, so something that comes out of the handicapping analysis. Some factor, or combination of factors, that makes them more than just a typical 3/1 shot.

Sometimes I think of it like 'antique shopping'. You see something at the flea market that they're sellilng as junk, you think it's something of potentially greater value. The better you are at distinguishing between junk and antiques, and reselling your 'antiques' that you bought at 'junk' prices, for more than you paid for the whole lot, the more successful you will be.

Case in point, that I sadly missed this year (was working that day). During the late fall, trainer Steve Asmussen, while having the best year of any trainer, had to serve a suspension (2-3 weeks?) for an offense during the prior year that he'd lost on appeal. During that time, he had his same normal assistant-trainers handling the horses, but now the horses were entered under their assistant's names, instead of Asmussen's name. Asmussen's first time starters have been winning over 20% of the time. At AP, where he sent some good horses, he had 2 first time starters. Instead of the horses debuting as 'Asmusssen' trained (with his great record and 20% stats showing in the PPS), they debuted under the obscure trainer name of Toby Sheets, who had very little history listed. The crowd didn't pick up that these horses might be coming from a successful outfit, from their same day-in, dayout, on-track handler at AP. The pair ran 1st and 2nd, both at odds of over 20/1, and the extacta was a signer.

Betting every 20/1 shot will lose a person a fortune. Finding the gems within that 20/1 group whose chances to win are greater than 5% will help a bettor be successful.

Side story - I took good note of this situation, and about a week later, I upgraded another Toby Sheets/Asmussen horse (a stretch out) using Asmussen's stats, and did catch a > 25/1 shot. (His tote-odds actually kept rising during the race - gotta love it!) That race, plus one other that same day, changed my AP meet from a $1,000 plus loss, to my first $1,000+ profit on an AP meet.

Did I make a profit for all of 2004? Sadly, no. But I know if I can better manage my time, so that I can better balance work + my handicapping (spend enough time looking for those subsets of better horses than the odds, doing more research, keep better records, improve my software and tools), and avoid costly time distractions (like spending too much time arguing politics in off-topic, Thank God the Election is over), then the potential for personal and betting success is there.

[boxcar]

sq764 wrote:

Ok, ready for this crazy crazy thought?? What if, you were a patient player and focused on demanding 3/1 on your horse

Or how' bout this for a really way out, radical idea: Patience + minimum of 3-1 on all your selections = big profits?

Is this ridiculous to believe this is happening?

I've heard of many more ridiculous things.

Boxcar

[boxcar]

Steve'StatMan'BTW wrote:

The only ones that can win at the game will do things like find a subset of the 3/1 shots that can win 28+% of the time.

Those that can't make those distinctions will lose long-term at the this game until they they do.

Great post, Steve! You've nailed it down tight. The "warp 'n' woof " of successful handicapping is precisely being able to make those kinds of critical distinctions -- having the knowledge, the experience, the ability to sift through the Good from the Bad, the Bad from the Ugly.

Handicapping isn't really all that different from most endeavors in life, which makes Steve's analogy involving antique buying very appropriate, for example.

I haven't played the ponies now in about a year and half. (I consider myself "semi-retired" and have taken up other time-consuming pursuits, which won't allow time for any serious handicapping.) But when I did play regularly (as in virtually every day), the primary thing I looked for in a race was to see if the race offered any viable longshot plays in it -- or even more specifically -- if there were any false longshots (> 9-1) in it. False longshots always meant to me that the public was about to make a big mistake, and I wanted to be all over those kinds of mistakes like white on rice. But obviously, in order to do this profitably, one needs to know what constitutes a false (or even vunerable) longshot. If one can't make these kinds of distinctions, then he stands no chance at beating the game.

Boxcar

[formula_2002]

Where do the numbers 2259 & 19824 come from? .

2259 was based upon 3-1 shots representing 11.95% of all the horses that ran (actual figure is closer to 9%)

2259 x.1195=270. 270 represent the min. number of plays one would have to bet, win 28% for a 10% profit at a 15% take-out track in order to obtain a significant # of winners.
----------------------------------------------------------------
19824 was based upon 10-1 shots representing 4.29% of all the horses that ran (actual figure is closer to 3.3%)

19824 x...0429 =850. 850 represent the min. number of plays one would have to bet, win 10% for a 10% profit at a 15% take-out track in order to obtain a significant # of winners.


Apparently, I need 76 (28% of 270) winners out of 270 plays to show significance at 3-1, right?

Yep.

I don’t know if you can take a sub set of 1000 3-1 horses from the 2259 and still have 270 plays. Most likely not, since is the % of plays in the subset would be greater then the % of plays in the sample.

[GameTheory]

2259 was based upon 3-1 shots representing 11.95% of all the horses that ran (actual figure is closer to 9%)

2259 x.1195=270. 270 represent the min. number of plays one would have to bet, win 28% for a 10% profit at a 15% take-out track in order to obtain a significant # of winners.
----------------------------------------------------------------
19824 was based upon 10-1 shots representing 4.29% of all the horses that ran (actual figure is closer to 3.3%)

19824 x...0429 =850. 850 represent the min. number of plays one would have to bet, win 10% for a 10% profit at a 15% take-out track in order to obtain a significant # of winners.


Apparently, I need 76 (28% of 270) winners out of 270 plays to show significance at 3-1, right?

Yep.

I don’t know if you can take a sub set of 1000 3-1 horses from the 2259 and still have 270 plays. Most likely not, since is the % of plays in the subset would be greater then the % of plays in the sample.Ahh... I thought you were saying I'd have to look at 2259 3-1 horses to find 270 to play, but I only need to look at 2259 horses total to find 270 3-1 horses. Which means once again you are talking about playing all 270 of the 3-1 horses that I find. So the 2259 is basically not relevant -- the important thing is that I need to find 270 3-1 horses and win 76 of them to show significance. That's fine.

Now, you still haven't justified your thinking. Namely:

A) Why do we have to break everything down by odds subsets? If we wanted, why can't we consider all horses in all races as one large group (which we know contains appox. 10% winners and will lose the amount of the take). In statistics, you can make the hypothesis anything you want. (Although once you define the hypothesis, you basically forget about it -- that's once of the problems with classical statistics -- you never prove a hypothesis, but attempt rather to accept or reject what is called the null hypothesis which is quite something different.) For some reason you demand that we prove statistical significance in incremental odds ranges, which you can always make smaller & smaller to stay lower than the sample size requirements to show significance. But I could just as easily make the hypothesis, "I can beat the take, period" and use any horse in any odds range as part of my sample. The question is essentially the same -- I am just using a different set of boundaries for my odds ranges than you are. Namely, I am using one big range that covers all odds we see at the track. So if your method is correct on small incremental ranges, then you can't claim it is incorrect just because I've changed the resolution. (The parameters change, but not the method.) Therefore, not breaking things up by odds is just as valid as breaking them up by odds. It is not required, just your preference.

B) This is the big one. You've shown nothing, argued nothing, presented no data that could lead to the conclusion that NO selection method can beat the take. This stuff with the odds is irrelevant to the question. You have told us the win percentage requirements that would be required to show significance given a certain profit margin, but never given any reason why those win percentages are unachievable. You've shown you can't win if you bet using odds as your only factor. Fair enough. And you've mentioned you have found no single factor that can beat the take (across all tracks). Excuse me again, but duh. So far we have gotten to the point the rest of us arrived at approxiately one hour into our handicapping careers. Now, what is the missing link that allows you to jump from point A (you can't win betting only on odds or any other single factor you know of) to point B (you can't win ever, period). Please, please, tell us the reason why those win percentages on your chart are unachievable. Please tell us why the fact that you can't find any single factor that beats the take means that there are no factors that taken together cannot beat the take.

Here's an analogy. I'll pick 5 NBA players, and you pick 1 NBA player. We'll have them play some 5 on 1 basketball. Let's say we can have them play 1 million games real fast (5 guys on one team, 1 guy on the other team). The team with 5 players wins all 1 million games. Big surprise. At this point, I'll let you pick a new single player to replace your 1 guy and we can try again. No matter which guy you pick, he never is able to win even a single game against my 5 man team. You try again and again with different single players, yet still never win any games. Now I say to you: "Ok, you can pick 5 players for your team as well. They can work as... well, a team. Do you think you'll be able to win any games then?"

Most people would say something like: "Yes, it's certainly much more likely. In any case, my experience using 1 man teams isn't really much of a guide to how I'd do with a full 5 man team. Just as I wouldn't expect 1 player to beat 5 in basketball, I wouldn't expect a single factor to be the key to the mint in horse race betting. However, give me 5 players, or factors in combination, and I've obviously got a better chance."

You, however, apparently would jump to the conclusion that since no single player could win any games, then no five players working together could win any either. Now does that REALLY make sense to you?

[formula_2002]

GT writes;

"Why do we have to break everything down by odds subsets? If we wanted, why can't we consider all horses in all races as one large group (which we know contains appox. 10% winners and will lose the amount of the take). "

If you believe that, I have no case to make..and neither do the 100's of published analysis, many of which you can find in "Efficency Of Racetrack Betting Markets" and "Horse Sense" By B.Fabricand.


I rest my case on those fellows..

You can take it up with them..

[GameTheory]

GT writes;

"Why do we have to break everything down by odds subsets? If we wanted, why can't we consider all horses in all races as one large group (which we know contains appox. 10% winners and will lose the amount of the take). "

If you believe that, I have no case to make..and neither do the 100's of published analysis, many of which you can find in "Efficency Of Racetrack Betting Markets" and "Horse Sense" By B.Fabricand.


I rest my case on those fellows..

You can take it up with them..

I didn't say what I believed -- I asked a question which you avoided as usual. I didn't say breaking down by odds might not be useful, I asked why it was *REQUIRED*.

You keep resting your case without ever making it first. Have you ever made your case publicly? I've never seen it. You just give us your unfounded conclusions with no reasons given whatsoever. Not only that, but you keep citing authors who disagree with you -- some of whom make lots of money betting horses. You don't think they can beat the take?

Of course you ignored the whole second part of my message. If you are standing on such a rock-solid foundation of science and data, why is it you are afraid to even ACKNOWLEDGE the most minor questioning of your conclusions? Ever heard of peer-review?

So again, how do you get from point A to point B? There must be some math, some data, some formula, that explains it, right? Surely, you have some reason for making your conclusions.

What's the big secret? Just tell us the reason. What is the reason? WHAT IS THE REASON????

[hurrikane]

ok, so tell me this,

A horse i have been following is in a race today at my favorite track in Maryland ( you can already tell this is not a true story).
i have been watching this horse and today he is listed at 10-1, i get to the track and he is still 10-1 I dump a lot of money on him and he goes off at 3-1 and wins.
There's one of your 3-1 horses.

same story only this time on the way to the track I have an accident. I don't get the bet down and the horse goes off at 10-1 and wins. There's oen of your 10-1 shots.

to me they are the same horse that i had figured at 3-1
to you they are different

same exact situation but the horses are going off at 2 different odds. Same chance of winning, only difference is I didn't get there to make the bet.

I had the horse running with the same chance of winning. I just didn't get the bet down. You on the other hand have made all kinds of assumptions about the results without having the knowledge that I didn't make a bet and the horse would have been 3-1.

where does your theory land now.
in the dumpster

The horses have no idea what the frigging odds are. They dont' even know when they have won, placed or ran 3rd.
They have never heard of B Fabricand and don't know a damn thing about their chances or winning at 3-1 or 10-1
All they know for sure is everyone is running and it's probably because they are getting hit by the same damn thing that keeps smacking their ass and it hurts.

you are making conclusions about your results that are not only simplistic and elementary but of no betting value whatsoever.

Come on Joe, stretch the envelope just a little.

take your data and find horses that should be 3-1 but are 10-1. Bet them down to 3-1 and make some money.

[Tom]

My last 10 bets:

2-1 -
9-2 W
7-2 W
8-1 W
5-1 -
3-1 -
4-1 -
7-1 -
9-2 W
5-2 -

Explain to me how I lost money, please.
:confused:

[formula_2002]

clarification
per the table I posted

"fourth column is the total # of horses you would have to review to find the playable horse."

futher
"Originally Posted by formula_2002
2259 was based upon 3-1 shots representing 11.95% of all the horses that ran (actual figure is closer to 9%)

2259 x.1195=270. 270 represent the min. number of plays one would have to bet, win 28% for a 10% profit at a 15% take-out track in order to obtain a significant # of winners. "

2259 represents ALL the horses on a card, of which 11.95% will be 3-1 shots.

So if one were to play ALL 270 3-1 shots and win 28% of the time his play would be significant.

Say only 1 in 10 3-1 shots can be handicapped to return a 10% profit.

So you would have to look at 2700 3-1 shots.
to find 270 playable 3-1 shots

(2700/.1195)= 22,594 horses

22594 horses /(8 horses per race x 9 races)=313 race days

[formula_2002]

My last 10 bets:

2-1 -
9-2 W
7-2 W
8-1 W
5-1 -
3-1 -
4-1 -
7-1 -
9-2 W
5-2 -

Explain to me how I lost money, please.
:confused:

Could you tell me what % of 3-1 shots you play(>=2.5 through<=3.4).. It will be helpful to know for my previous note.

Thanks

[formula_2002]

[QUOTE=GameTheory]I didn't say what I believed -- I asked a question which you avoided as usual. I didn't say breaking down by odds might not be useful, I asked why it was *REQUIRED*.

/QUOTE]

And I gave you my answer. I refered you to the work of others.. It's not my intention to reinvent the wheel.
Can you accept that ?

If you do not accept the basis of incremental odds analysis, anything I may say will be meaningless to you.
Do you accept incremental odds analysis ? yes or no.

I'll only acknowledge one word from you YES or NO.

[hurrikane]

no

[Kreed]

Game Theory vs Formula ... the winner is GT. Incremental odds analysis is
an interesting descriptive stat, but let's face it, just how often would any
useful method pick a 15:1 on TOP? Yes, yes, sometimes the Gods bless you
with a windfall pick, but most usually we see top picks return between 8:5 to 4:1. (Just a week ago at AQU, a horse named SCRAPPY T won wire-2-wire and
paid ~$22. It was a total non-brainer; Scrappy was the sole speed on a track
kind to speed & getting kinder with every passing race & was in top shape etc.)
I really think that winners (as a group) that pay MORE than ~$10 form a VERY
different group of winners than say those that pay $6. And lastly, I don't think
anyone has to "beat the take" to win. Develop a GOOD method that selects
top picks paying from 8:5 to 4:1 to win; get NEAR 4% of actually winning;
place bets with Pinnacle or some other rebate shop that take your -4% into
plus territory. IOW, develop a GOOD consistent near-winning method THEN
enhance it with REBATE betting + a good MATH approach that tells you HOW
MUCH to bet & sorta HOW to bet the pick.

[formula_2002]

Game Theory vs Formula ... the winner is GT. horse named SCRAPPY T won wire-2-wire and
paid ~$22. It was a total non-brainer;

Give me an idea of what % winners are no brainers, or are all winners no brainers.

[chickenhead]

I don't belong in this discussion but I a few honest questions.

1) I've seen the term "statistically significant" a few times in this thread. Joe, can you tell me what level of confidence your #'s provide? 80%? 99%? 99.999%?

2) Since there is such a negative expectation (-15%) I am surprised it takes so many samples. Would the sample size required change if you were trying to prove a 0% or -4% return as Kreed suggests (with an eye towards rebates)? Is it smaller? By how much?

3) Also, if all of your incremental odds ranges showed a much better than -15% return, even if each of them individually did not meet the sample size criteria, is that not itself statistically significant to some degree? It seems you could model that and be just as accurate, with a smaller total sample.

I am not "trolling" here, just interested. Thanks.

[GameTheory]

[QUOTE=GameTheory]I didn't say what I believed -- I asked a question which you avoided as usual. I didn't say breaking down by odds might not be useful, I asked why it was *REQUIRED*.

/QUOTE]

And I gave you my answer. I refered you to the work of others.. It's not my intention to reinvent the wheel.
Can you accept that ?

If you do not accept the basis of incremental odds analysis, anything I may say will be meaningless to you.
Do you accept incremental odds analysis ? yes or no.

I'll only acknowledge one word from you YES or NO.

Of course that's all you'll acknowledge, because you can't defend your point. You may refer to the work of others, but you might as well refer to Shakespeare, because the defense for your point will not be found in Efficiency of Racetrack Betting Markets just as it won't be found in Hamlet or MacBeth.

Let's review the definition of the Ferrous Cranous:

"Ferrous Cranus is utterly impervious to reason, persuasion and new ideas, and when engaged in battle he will not yield an inch in his position regardless of its hopelessness. Though his thrusts are decisively repulsed, his arguments crushed in every detail and his defenses demolished beyond repair he will remount the same attack again and again with only the slightest variation in tactics."

Yep, that's about right.


But let's say I do accept the incremental odds analysis -- after all, I DID accept it if you read my message carefully, only pointed out that you could set the increments to anything you wanted, including one large "increment" that covered all horses. You could also use increments even smaller than you do, down to .1 of a point. Or did the voice of God tell you that you MUST round off to the nearest whole number or insightful analysis is impossible?

Anyway, that is all tangential. You still haven't given any reasons for your main point that statistical significance is UNACHIEVABLE is the real world. Please tell us why this is so. If you are going to claim that "others" have given these reasons in their work, then please give a title and a page number, preferably with a direct quote. I would be very much interesting in knowing the part I missed in the "work of others" where they claim that it is impossible to perform at a level to show significance.

[formula_2002]

I don't belong in this discussion but I a few honest questions.

1) I've seen the term "statistically significant" a few times in this thread. Joe, can you tell me what level of confidence your #'s provide? 80%? 99%? 99.999%?

2) Since there is such a negative expectation (-15%) I am surprised it takes so many samples. Would the sample size required change if you were trying to prove a 0% or -4% return as Kreed suggests (with an eye towards rebates)? Is it smaller? By how much?

3) Also, if all of your incremental odds ranges showed a much better than -15% return, even if each of them individually did not meet the sample size criteria, is that not itself statistically significant to some degree? It seems you could model that and be just as accurate, with a smaller total sample.

I am not "trolling" here, just interested. Thanks.

1. 99%
2 . Sample size is a function of actual winner and expected winners ( See Quirin's "Winning At the Races" page 297)
Actually as the roi increases the sample size decreases.

3. if the sample size is large enough, you can demonstrate significance and still show a negative return.

[formula_2002]

Additionally, as the track take-out increase the sample size can decrease, simply because the numbers of expected winners decrease.

I have a interactive significance test on my web page.
To be on the safe side the test assumes no take-out.

I'll be updating is soon to include an interactive variable for take-out.

[NoDayJob]

:D Jeez, all this trolling. I'm completely trolled out! :D

NDJ

[chickenhead]

Could you explain to me the proper application of this 99% confidence?

This means you are 99% certain that the results were not due to randomness? Does this also then mean (theoretically) that if repeated, you are 99% certain to achieve a (roughly) 10% ROI again?

I have a follow on question, or a more detailed version of question #2, but I want to make sure I understand #1 first.

[formula_2002]

Could you explain to me the proper application of this 99% confidence?

This means you are 99% certain that the results were not due to randomness? Does this also then mean (theoretically) that if repeated, you are 99% certain to achieve a (roughly) 10% ROI again?



I'll quote from Quirin's book, pg 297
" Usually-99 percent of the time- this number will be somewhere between +2.5 and -2.5. When it is not, we have found something unusual and significant- a group of horsesthat, by their very nature, win more (a value larger than +2 1/2 of less(a value smaller the-2.5) than expected...and figure to continue doing so..."

The number he is talking about is the number generated by the "standard normal" test.

Take a look at the "Test for significance" file on my web page. Be aware that the formula is not adjusted for track take-out.
I'll update the file shortly.

[GameTheory]

Could you explain to me the proper application of this 99% confidence?

This means you are 99% certain that the results were not due to randomness? Does this also then mean (theoretically) that if repeated, you are 99% certain to achieve a (roughly) 10% ROI again?

I have a follow on question, or a more detailed version of question #2, but I want to make sure I understand #1 first.

It is important to understand that in this type of hypothesis testing, you only can reject the "null hypothesis". You assume the null hypothesis to be true, and then try to find enough evidence to reject it.

So let's say we our group of 270 3-1 horses that we've bet on, and we won 26% of them and made a tiny profit. All 3-1 horses win at a rate of 21%. So now we want to know if our 26% win rate was due to chance or due to good handicapping. The null hypothesis in this case would be that the true winning rate of both samples is 21%. So now we attempt to reject that hypothesis. At the 1% significance level (which is 99% confidence) with 270 horses, we need 28% winners to reject the null hypothesis. If we reject the null hypothesis, and we have set up our experiment properly, then we have no choice but to accept the alternative hypothesis -- that our winning at a higher rate was not due to chance. What it really means is that ASSUMING THE NULL HYPOTHESIS IS TRUE, then no more than 1% of the time will we see results as extreme or more extreme (winning 28% or better) if we draw 270 horses at random from the sample of all 3-1 horses.

However, failing to reject the null hypothesis is not the same as accepting it!

So we WOULD NOT then conclude that our 26% winning rate was due to chance -- according to the theory we don't have enough evidence to believe anything conclusively. So just because you can't reject the null hypothesis doesn't mean you are seeing random results -- it just means you MIGHT be. Classical statistics has no mechanism to answer the question you probably really want to know, that is, just how likely are these results to be random? It has no answer for that. Bayesian statistics, on the other hand, will tell you exactly how likely your results were compared to random; and there are some cases where it is upwards of 95% likely that your results are not random and normal hypotheis testing will give you no clue about it.

[Tom]

Could you tell me what % of 3-1 shots you play(>=2.5 through<=3.4).. It will be helpful to know for my previous note.

Thanks

What is your fixation on 3-1 shots? I do not, nor does anyone else (except maybe YOU) limit my play to 3-1 shots. I probably lsoe when you isolate all my plays on 3-1 shots. So what. That is so irrelevant. Imake money at the track. My track income exceeds my betting. I BEAT THE TAKE.
Anything else if fun with numbers and meaningless.

[GameTheory]

Sorry, lost a paragraph (yes, there is more).

Just because you can't show something is statisically significant (like beating the races) doesn't mean that if you ARE making a profit that you aren't "really" beating the races (i.e. will continue to make a profit). It just means you can't prove it. That part of Joe's argument we can accept -- that if you can't prove it, it remains unproven. (We will leave aside the very important fact that racing is a moving target and that even if you could prove it that it wouldn't really mean anything practically.) However, the conclusion that because we can't prove it now that we will we NEVER be able to prove has has no basis whatsoever -- we may just need to maintain our performance over a larger sample. There is also no basis to the conclusion that if your performance is unproven at X level of significance, then your performance is really just random. No basis at all. Take a look at any "intro to statistics" book or webpage and it will explain that. So Joe's assertion that the races can't be beaten has no foundation in classical statistics, and in fact he has no reason for making such an assertion other than just personal conjecture. Since he refuses to talk about the reason that should tell you all you need to know...

[Tom]

Using Joe's ideas, I doubt you could prove you are alive.
Yopu DO NOT prove reality with statistics. You prove reality with facts, in this case, money. NOTHING else is relevant.

[Kreed]

i AGREE that Joe's ideas are just his but he's right in one sense: making $$
at betting horses is very very hard. is it the take? yes, partly. the nature
of the horse? yes, partly. the many possible ways to bet your pick? partly
i guess. (Yeah, the EX bets made sense but those Tri's just wasted money.)
you can lose money in lots of ways. i can't imagine betting for a livelihood.

[nomadpat]

Could you guys imagine Joe in a relationship?

"Sorry dear, but the flowers, phone calls, kind words, presents, etc you gave me doesn't prove statistically significant that you love me..." :p

[Dave Schwartz]

GT,

Good post. You are right on the money.

The problem is that Ole Joe wants to say that not being able to prove it means you aren't doing it. That is what I take issue with.


Regards,
Dave Schwartz

[formula_2002]

GT,

Good post. You are right on the money.

The problem is that Ole Joe wants to say that not being able to prove it means you aren't doing it. That is what I take issue with.


Regards,
Dave Schwartz

It appears to me, the only people who disagree with me are those that feel pari-mutual betting operates on some laws outside of mathematics and statistics...

On another post I worked out that one would have to review 12207 horse to find 270 of those 3-1 shots.
-----------------------------------------------
I'm going to assume 2.5 overlays (at any odds) per race card.

say 8 horses x 9 races =72 horses

2.5/72 = 3.5% overlays per race card

Very rough figure.


3-1 horses represent 8.85% of al the horses that run.

say only 3.5% of those horses are overlays.

Therefore;

270/.035 x .0885=81716

Which means I would have to handicap 81716 horses to find 270 3-1 "system" picks.

81716/( 8 horses per race x 9 raced per card)=
1134 track card days, or perhaps 3 years worth of racing.
---------------------------------------------------

[Pace Cap'n]

It appears to me, the only people who disagree with me are those that feel pari-mutual betting operates on some laws outside of mathematics and statistics...

By Jove, I think you've got it!

[sq764]

It appears to me, the only people who disagree with me are those that feel pari-mutual betting operates on some laws outside of mathematics and statistics...

---------------------------------------------------
Um, I would have to say that includes most of the population. I haven't seen one person agree with you on this ridiculous thesis..

Do you even use algorithms to decide whether you're have a bagel or an english muffin in the morning? Do we rule out that you have a laptop on the kitchen counter?

[chickenhead]

There are two issues, can you prove it was done, and can it be done. I don't see how you can answer NO to either of those, your own math supports that it can be proven (to the confidence level of your choosing), and people on this board support the fact that it can be done. Now, if you don't beleive anyone is beating the take, OK. But your own math supports that it is possible, and that it is provable.

So what exactly is the beef?

[DJofSD]

Mathematics by itself does not prove anything.

According to mathematics, I can travel faster than the speed of light. Physics says otherwise.

Mathematically, you can beat the track take and make a profit. Real world experience of many people demonstrate that it can be done.

If your analysis using statistics, data bases, phases of the moon, or whatever, is suppose to prove the assertion it can not be done then as far as I'm concerned, the onus is one you to prove you assertion is valid ***and*** to explain how some people can do what you say can not be done. The fact that people do consistenly make money at the track, minimally, shows your theory is flawed.

DJofSD

[GameTheory]

The thing is, he is not even supported by his statistics. In technical terms, he seems to be saying that the null hypothesis is proven because he hasn't seen a proof for the alternative hypothesis. But anyone learns on day one of statistics class that you can NEVER prove the null hypothesis. Math & statistics are not on his side. Nothing is on his side. The strongest evidence for his case is anecdotal, but since he rejects that type of evidence that means he has no evidence at all.

[NoDayJob]

:) After reading all of the mathematical and hypothetical claptrap, the only sound conclusion one can come to is how much money you will bank, after all expenses. That is what this business is about; id est, you consider this a business and not a hobby. Hmmmmmmmmmmmmmm? :)

[Tom]

I baked a cake at 350 degrees for 30 minutes, so if I raise the temperature to 700 derees, can I bake one in 15 minutes?

******

Monday, I drank gin and water and I got drunk.
Tuesday, I drank whiskey and water and I got drunk.
Wednesday, I drank Scotch and water and I got drunk.
Science would suggest......that damn water is making me drunk.

[Buckeye]

I say let's buy some more ALL-WAYS files and then . . . :eek:

You prove I can't win and then I do.

You prove I must win and then I don't.

Nice proof.

[formula_2002]

There are two issues, can you prove it was done, and can it be done. I don't see how you can answer NO to either of those, your own math supports that it can be proven (to the confidence level of your choosing), and people on this board support the fact that it can be done. Now, if you don't beleive anyone is beating the take, OK. But your own math supports that it is possible, and that it is provable.

So what exactly is the beef?

Well you are right, my own math says it can be done.
In the case of 3-1 odds all you have to do is review
12207 horses of all odds to fin the 270 playable horses.

For 4-1 odd its 15432 horses
for 5-1 odds its 18182 horses.

for odds of 1-1 through 12-1 it adds up to 685242 horses.
I would think there is some redundency in that figure, so say we cut it almost in half..to 300,000.. thats only 4166 track days. Say 3 tracks a day. Then all one would have to do is bet for about 4 years.. to prove a 10% profit at a 15% take-out track (actually, my figures average out for a 17% take-out track)

You can even cut that number way down if you do the math for a 20% profit at a 20% tak-out track.

You can really cut down the number of plays if you want to prove 30% profit.

So there you have it..

By the way, anyone reading Mitchell's book "commonsense betting" can go to page 40. He indicates that you have to check 2865 races to make a 20% profit at average odds of 3.80-1. I only require you to check 270 races.
IF HE IS RIGHT IT COULD TAKE 40 YEARS.

I'm sure all this will encourage many people to set up shop
and make their fortune :rolleyes:

[Buckeye]

Mark Twain

"There are three kinds of lies: lies, damned lies and statistics."

That's assuming even correct application!

Rule one: it is impossible to prove a negative.

Has that changed?

[formula_2002]

Mark Twain

"There are three kinds of lies: lies, damned lies and statistics."

Would that incluse "past performance data"

That's assuming even correct application!


I'm sure he would exclude the horse racing betting public.

The sum of their thinking is most accurate. And they thrive on stats.

[Buckeye]

I'd be interested in purchasing some "future performance" data, got any of that?

If not, just give me the statistical breakdown.

Hope you'll remember to include all the possibities. :cool:

Consider if you will that "luck" by itself cannot entirely explain the good results some of us will achieve in the future.

[formula_2002]

I'd be interested in purchasing some "future performance" data, got any of that?

If not, just give me the statistical breakdown.

Hope you'll remember to include all the possibities. :cool:

Consider if you will that "luck" by itself cannot entirely explain the good results some of us will achieve in the future.


yep.. just watch the final odds.

[Buckeye]

Dream On

Aerosmith


Every time that I look in the mirror
All these lines on my face gettin' clearer
The past has gone
It went by like dusk to dawn
Isn't that the way
Everybody's got their dues in life to pay


Yeah, I know nobody knows
Where it comes and where it goes
I know it's everybody's sin
You got to lose to know how to win


Half my life's in books written pages
Livin' and learnin' from fools and from sages
You know it's true
All the things come back to you


Sing with me, sing for the year
Sing for the laughter and sing for the tear
Sing with me, if it's just for today
Maybe tomorrow the good Lord will take you away


Well...sing with me, sing for the year
Sing for the laughter
And sing for the tear
Sing with me, if it's just for today
Maybe tomorrow the good Lord will take you away


Dream on, dream on, dream on
Dream yourself a dream come true
Dream on, dream on, dream on
And dream until your dream comes true


Dream on, dream on, dream on, dream on
Dream on, dream on, dream on...aaahhh


Sing with me, sing for the year
Sing for the laughter
And sing for the tear
Sing with me, if it's just for today
Maybe tomorrow good Lord will take you away


Sing with me, sing for the year
Sing for the laughter
And sing for the tear
Sing with me, if it's just for today
Maybe tomorrow the good Lord will take you away.

[formula_2002]

There are two issues, can you prove it was done, and can it be done. I don't see how you can answer NO to either of those, your own math supports that it can be proven (to the confidence level of your choosing), and people on this board support the fact that it can be done. Now, if you don't beleive anyone is beating the take, OK. But your own math supports that it is possible, and that it is provable.

So what exactly is the beef?


I just thought of another mathematical proof for you..

buy low sell high. ;)

[chickenhead]

Well you are right, my own math says it can be done.

Then all one would have to do is bet for about 4 years.. to prove a 10% profit at a 15% take-out track (actually, my figures average out for a 17% take-out track)

I'm sure all this will encourage many people to set up shop
and make their fortune :rolleyes:

I fail to see the doom and gloom in all of this....when I buy a stock, and it goes up or down -- how can I prove whether I was smart, stupid, lucky, or unlucky? Basically, by measuring my performance over a long period of time, I gain confidence (or lose confidence). How is this different? Why would it be different?

Investments always carry risk, and as the little small print at the bottom of everything talking about any stock says -- past performance is no indication of future performance anyway.

[VicD]

This is interesting, and I believe both sides have valid points....
The guy with the stats never takes into account a persons ability and dicipline, and these two things probably can have a statistical or numeric "value" attached to them....
I also smile when someone says that "many" people can make a living betting horses, because in 20 years of serious playing in several different states I have not met "many" of those guys..
One guy I do know that is known to be a pro player and is very well respected in NJ is not strictly speaking a handicapper.. He has working relationships with stables and he pays handsomely for information relating to the animals fitness,
health, and "intent" of the trainer. I am NOT talking about tips here, but real black and white information....
The question is this: Is this guy really a handicapper, or is this akin to insider trading in the stock market??
I have also know guys who claim to be professional gamblers, have big, beautiful homes, and drive expensive cars.... Some of these guys are bookmakers, and last time I checked the book has a mathematical advantage that seems hard to beat, as evidenced by his lifestyle....
I am not saying that consistent winners do not exist because they do...
I do question when someone say to me "I make a living betting horses".
What constitutes a living for one person may not be a living for another...
I do not consider making $500/week a living... I would like to observe someone who claims to make a six figure income/year betting horses....
I don't want to intrude on his style/system, just would like to examine the results over time.... I really do not care how or why he made the bets....
Like many of you, I have a betting style, and have a decent bankroll to implement it,and the results are satisfactory to me, but I would not quit my day job....
I am new to this site and find it to be entertaining.....

[Tom]

Your study of 3-1 odds horses is flawed - with late odds chages, no one CAN bet all 3-1 shots, therefore, your study is of a real life occurrance that cannot happen.

[DJofSD]

Just finished watching this new CBS program.

Very good demonstation why numbers and mathematics are only tools -- not the end-all, be-all that some make them out to be.

DJofSd

[formula_2002]

Mark Twain

"There are three kinds of lies: lies, damned lies and statistics."

That's assuming even correct application!

Rule one: it is impossible to prove a negative.

Has that changed?

I guess so, if it has anything to do with money.
It's very simple to prove a track take out.
Bookies have been doing it for years.

[formula_2002]

Just finished watching this new CBS program.

Very good demonstation why numbers and mathematics are only tools

DJofSd

Thats no small thing.

After watching football for 7 hours, there was not enough gas in the tank to watch that program.

[Buckeye]

It's very simple to prove a track take out.
Bookies have been doing it for years.

They've been beating the public for years, which is not the same thing as beating everybody . . . In fact, some players are simply refused or limited.

The pari-mutuel system does impose limits on winning players but it's just silly to suppose that ALL players lose a percentage equal to the track take.