Betting early vs betting late

[arw629]

I post on this website from time to time and enjoy reading the discussions, but one subject I've come across from time to time has been bugging me.....

Why do some people make their bets before the races start and then just sit back and watch? There is so much information that can be learned from the races as they progress through the card such as biases, rider changes, weather changes, biases, scratches, biases, biases, biases......I like to try to analyze the races beforehand and pending on how the track is playing make my selections....I also like to watch the tote board for information.....I feel like the people who bet early are really at a disadvantage to the player who makes his bets at post time......I'm not calling people out, but if someone can post a rebuttal to this I would like to hear the reasoning in case I am missing something.....I'm always looking to improve my handicapping and would love someone to bounce something off me.....ok i'm rambling now

[davew]

Some people may want to bet on a horse or race, but have other time commitments during pre-race and race.

[arw629]

Some people may want to bet on a horse or race, but have other time commitments during pre-race and race.

What about the people who don't have time commitments?

[Stillriledup]

I post on this website from time to time and enjoy reading the discussions, but one subject I've come across from time to time has been bugging me.....

Why do some people make their bets before the races start and then just sit back and watch? There is so much information that can be learned from the races as they progress through the card such as biases, rider changes, weather changes, biases, scratches, biases, biases, biases......I like to try to analyze the races beforehand and pending on how the track is playing make my selections....I also like to watch the tote board for information.....I feel like the people who bet early are really at a disadvantage to the player who makes his bets at post time......I'm not calling people out, but if someone can post a rebuttal to this I would like to hear the reasoning in case I am missing something.....I'm always looking to improve my handicapping and would love someone to bounce something off me.....ok i'm rambling now

Are these people serious players and borderline pros who are betting hours in advance? Or, are you talking about elite players who are tossing thousands of dollars on their picks as soon as the betting opens and then laying on the beach while checking the results later that night to see if they've won?

[thaskalos]

I post on this website from time to time and enjoy reading the discussions, but one subject I've come across from time to time has been bugging me.....

Why do some people make their bets before the races start and then just sit back and watch? There is so much information that can be learned from the races as they progress through the card such as biases, rider changes, weather changes, biases, scratches, biases, biases, biases......I like to try to analyze the races beforehand and pending on how the track is playing make my selections....I also like to watch the tote board for information.....I feel like the people who bet early are really at a disadvantage to the player who makes his bets at post time......I'm not calling people out, but if someone can post a rebuttal to this I would like to hear the reasoning in case I am missing something.....I'm always looking to improve my handicapping and would love someone to bounce something off me.....ok i'm rambling now

Some players have convinced themselves that their methods of play can win without any regard given to the odds on the tote board. They have supposedly tested their methods on paper from their home, and the methods have proven to be profitable, even when their plays are made all at once...right at the beginning of the racing program. This way, not only do these lucky bettors avoid the nagging second-guessing that the rest of us have to deal with...but they can also choose to leave the often unpleasant atmosphere of the racetrack or the OTB, and go spend their day doing something more pleasurable...while returning later at night to collect their winnings.

Another pleasant byproduct of this style of betting is that you don't get upset over the day's events, and lose your composure...thus making extra wagers which you never originally intended to make. You remain much more disciplined this way.

[Stillriledup]

Some people have convinced themselves that their methods of play can win without any regard given to the odds on the tote board. They have supposedly tested their methods on paper from their home, and the methods have proven to be profitable, even when the plays are made all at once...right at the beginning of the racing program. This way, not only do they avoid the nagging second-guessing that the rest of us have to deal with...but they can also choose to leave the often unpleasant atmosphere of the racetrack or the OTB, and go spend their day doing something more pleasurable...and return later at night to collect their winnings.

Another pleasant byproduct of this style of betting is that you don't get upset over the day's events, and start making extra wagers which would didn't originally plan to make. You remain much more disciplined this way.

They also don't have additional aggravation of watching their winners get cut, sometimes in half, at the 'top of the lane' (top of the lane is in quotes)

Here's the list of horses who magically went down in price 'at the top of the lane' from April 12th:

Races 5 and 7 at Aqu

Races 5 and 6 at SA

Race 5 at GG.

Im sure there are more, listem if you got em!

[ultracapper]

How come nobody ever lists the horses that went up in price during the running of the race? I had one Saturday at Santa Anita. Went from 5/1 to 6/1 half way down the backstretch with a length and a half lead. Finished strong and wired them. It was Glatt/TBaze in the 6th, I think.

HAHA. Only bet I made this week and I can't remember which race it was and too lazy to look. It was Glatt/TBaze and it paid $14 and change. 5/1 out of the gate and down the backstretch.

[thaskalos]

How come nobody ever lists the horses that went up in price during the running of the race? I had one Saturday at Santa Anita. Went from 5/1 to 6/1 half way down the backstretch with a length and a half lead. Finished strong and wired them. It was Glatt/TBaze in the 6th, I think.

HAHA. Only bet I made this week and I can't remember which race it was and too lazy to look. It was Glatt/TBaze and it paid $14 and change. 5/1 out of the gate and down the backstretch.
Maybe because it happens so rarely?

[ultracapper]

They raced 9 at SA on 4/12, another 8 at GG. How many at Aqua? 10? If so, SRU provided 5 examples out of 27 races. Everybody makes it sound like it's a race to race thing. So we have less than 20% of the races where the winner's odds drop during the race. It's a function of data collection, IMHO. I'm sure if you tried equally as hard you could find 5 losers that dropped after the gates opened also.

[ultracapper]

Maybe because it happens so rarely?

IMHO, we don't hear about the ones that go up because it would totally screw up the bitching value of talking about the ones that go down.

[Stillriledup]

Maybe because it happens so rarely?

1 goes up for every 9 that come down.

[Robert Fischer]

Some players have convinced themselves that their methods of play can win without any regard given to the odds on the tote board. They have supposedly tested their methods on paper from their home, and the methods have proven to be profitable, even when their plays are made all at once...right at the beginning of the racing program. This way, not only do these lucky bettors avoid the nagging second-guessing that the rest of us have to deal with...but they can also choose to leave the often unpleasant atmosphere of the racetrack or the OTB, and go spend their day doing something more pleasurable...while returning later at night to collect their winnings.

Another pleasant byproduct of this style of betting is that you don't get upset over the day's events, and lose your composure...thus making extra wagers which you never originally intended to make. You remain much more disciplined this way.

These have to be two of the leading reasons.

I believe the 'discipline' reason to be the most powerful.
If you look back at your losses and you find that you lost your composure following certain triggers (maybe a losing streak), then you are going to plan ahead to deal with and/or avoid those triggers.

You also have to play to your strengths.
If you are a rare bird that has developed the ability to have an actual edge with reading track bias, or with reading the race-by-race market behavior, - then you want to use that edge to the best of your ability.
But if you honestly do not have those edges, it becomes more a matter of what you'd rather be doing - playing along with the races, or another activity.

Multi-race players also have only the lead-in market and doubles to see. So some multi-race players find that it's best to play their multi-race sequences after a cutoff point like "morning Scratches/Changes", which then frees them up to either play the single race wagers or do something else.

[traynor]

I post on this website from time to time and enjoy reading the discussions, but one subject I've come across from time to time has been bugging me.....

Why do some people make their bets before the races start and then just sit back and watch? There is so much information that can be learned from the races as they progress through the card such as biases, rider changes, weather changes, biases, scratches, biases, biases, biases......I like to try to analyze the races beforehand and pending on how the track is playing make my selections....I also like to watch the tote board for information.....I feel like the people who bet early are really at a disadvantage to the player who makes his bets at post time......I'm not calling people out, but if someone can post a rebuttal to this I would like to hear the reasoning in case I am missing something.....I'm always looking to improve my handicapping and would love someone to bounce something off me.....ok i'm rambling now

Basically, it depends on how you view races. If you regard each race as a unique, never-to-be-repeated event, subject to a multitude of real-time factors and influences that must be analyzed, evaluated, and used as part of the decision-making process ... then, yes, it makes sense to only wager at the last minute. It also makes sense to totally ignore any "statistical" evaluations as little more than distractions, because "statistical pressure" on individual events, or small samples of events, is pure myth.

If you regard each race as one of many similar, with more in common with other races in that (however arbitrary) grouping than differences, and--in particular--your decision-making process is based on the result of a large sample of similar races, then it makes sense to wager whenever, minus the last minute agonizing and flip-flopping that seems to stir up the emotional juices of (many) other bettors.

It should come as no great surprise to anyone that people who are/have been/occasionally still are professional blackjack players tend to favor a statistical approach to wagering, minus the last minute agonizing and flip-flopping. They bet on horse races like they bet at blackjack.

Conversely, those who seek (or find) "emotional stimulation" in wagering tend to relish the anxiety and uncertainty as part of their "reward."

I don't want the "emotional stimulation," nor do I consider it of any interest or value. I just want the money.

[Fingal]

Some players have convinced themselves that their methods of play can win without any regard given to the odds on the tote board. They have supposedly tested their methods on paper from their home, and the methods have proven to be profitable, even when their plays are made all at once...right at the beginning of the racing program. This way, not only do these lucky bettors avoid the nagging second-guessing that the rest of us have to deal with...but they can also choose to leave the often unpleasant atmosphere of the racetrack or the OTB, and go spend their day doing something more pleasurable...while returning later at night to collect their winnings.

Another pleasant byproduct of this style of betting is that you don't get upset over the day's events, and lose your composure...thus making extra wagers which you never originally intended to make. You remain much more disciplined this way.

I'll agree with this except for the 1st sentence, odds do play a factor. I don't care who it is (insert name) & how sure a projected winner it is if the ML or my feeling is going to result with a number less than 3/1, pass. If I'm going to risk 100% of my money, I want the opportunity for at least a 300% or better return for taking that risk.

But I'll map out my bets the night before & place them as soon as I walk in the track for a couple reasons. First it's more relaxed since I usually get there about an hour before 1st post & the whole crowd hasn't shown up yet. Second, I find it a lot easier to cancel a bet than to place one, it's put the ticket in the SAM & hit cancel. To place a bet it's more steps & with impatient people around you I'd rather not go through that, let others get stressed out if they'll get action or not.

This way I've got my bets in, I can socialize & if I the tote says the odds are unacceptable it's excuse myself for a couple minutes to find a SAM.

[green80]

1 goes up for every 9 that come down.

That cannot be mathematically correct.

[arw629]

They also don't have additional aggravation of watching their winners get cut, sometimes in half, at the 'top of the lane' (top of the lane is in quotes)

Here's the list of horses who magically went down in price 'at the top of the lane' from April 12th:

Races 5 and 7 at Aqu

Races 5 and 6 at SA

Race 5 at GG.

Im sure there are more, listem if you got em!

Although it didn't change in the lane....april 4th, race 6 at mountaineer I placed a $100 win bet on :2: Ski Pole at 9-5 at post time....the gates sprung and the toteboard changed into the first turn up to 4-1.....Ski Pole won...I was pretty damn happy about that.....someone obviously placed a large bet on :5: at the last possible second to drop to 1-5 only to run third

[Stillriledup]

Although it didn't change in the lane....april 4th, race 6 at mountaineer I placed a $100 win bet on :2: Ski Pole at 9-5 at post time....the gates sprung and the toteboard changed into the first turn up to 4-1.....Ski Pole won...I was pretty damn happy about that.....someone obviously placed a large bet on :5: at the last possible second to drop to 1-5 only to run third

This happened at SA yesterday, someone hammered the 1 in the first race and the winner paid 8 bucks.

These instances are rare though.

[Stillriledup]

That cannot be mathematically correct.

i'm talking about 10 separate races, not 9 horses going down in price (while one goes up) in one specific race.

[arw629]

This happened at SA yesterday, someone hammered the 1 in the first race and the winner paid 8 bucks.

These instances are rare though.

Very rare indeed just sharing an example of the other side of the fence.....

[ultracapper]

SRU, you may want to do a study on this subject. It would at least either add more fuel to your concerns or alleviate them to some degree. It's obviously quite maddening to you when you see price drops after the gate opens. This is an issue you can investigate with readably available factual data. Maybe do it for a day, or a week if you really want to get into it. Watch the replays for Santa Anita next week. Make note of the odds on the favorite and winner as they enter the gate. Compare them to the odds when the race goes official. My guess, from strictly years of observation would be something like 40%drop/40%stay the same/20%go up on favorites. On winners I'd guess 30/50/20. I'd also guess the higher the odds of the winner, the more likely the odds would drop at the expense of both the other two, as big money being placed on a higher priced horse would be more likely to be held as long as possible, at least more likely than on an odds on favorite. I'd guess money held as long as possible on shorter priced horses has more to do with habits and personal preferences than on a fear of screwing up an already short priced horse.

I know somebody will say I should do it myself, but I really don't care enough about it, and I don't feel "after the start betting" is prevalent anywhere in North America. As a matter of fact, aside from intentional fraud, I don't believe it exists at all. Personal opinion.

I'm suggesting this to you because it obviously drives you crazy. If the results fall near the parameters I'm guessing, I'd think that would suggest to a reasonable person that you can be comfortable that it's a by-product of data collection and reporting. Of course it won't confirm anything one way or another, but it might suggest one way or the other.

BTW, did you see the Hollendorfer that won yesterday at SA and paid 12/1. That horse entered the far turn 21/1, and then dropped to 12/1. Somebody nailed it good.

[DeltaLover]

Basically, it depends on how you view races. If you regard each race as a unique, never-to-be-repeated event, subject to a multitude of real-time factors and influences that must be analyzed, evaluated, and used as part of the decision-making process ... then, yes, it makes sense to only wager at the last minute. It also makes sense to totally ignore any "statistical" evaluations as little more than distractions, because "statistical pressure" on individual events, or small samples of events, is pure myth.

If you regard each race as one of many similar, with more in common with other races in that (however arbitrary) grouping than differences, and--in particular--your decision-making process is based on the result of a large sample of similar races, then it makes sense to wager whenever, minus the last minute agonizing and flip-flopping that seems to stir up the emotional juices of (many) other bettors.

It should come as no great surprise to anyone that people who are/have been/occasionally still are professional blackjack players tend to favor a statistical approach to wagering, minus the last minute agonizing and flip-flopping. They bet on horse races like they bet at blackjack.

Conversely, those who seek (or find) "emotional stimulation" in wagering tend to relish the anxiety and uncertainty as part of their "reward."

I don't want the "emotional stimulation," nor do I consider it of any interest or value. I just want the money.

:ThmbUp: :ThmbUp:
good post

[Stillriledup]

SRU, you may want to do a study on this subject. It would at least either add more fuel to your concerns or alleviate them to some degree. It's obviously quite maddening to you when you see price drops after the gate opens. This is an issue you can investigate with readably available factual data. Maybe do it for a day, or a week if you really want to get into it. Watch the replays for Santa Anita next week. Make note of the odds on the favorite and winner as they enter the gate. Compare them to the odds when the race goes official. My guess, from strictly years of observation would be something like 40%drop/40%stay the same/20%go up on favorites. On winners I'd guess 30/50/20. I'd also guess the higher the odds of the winner, the more likely the odds would drop at the expense of both the other two, as big money being placed on a higher priced horse would be more likely to be held as long as possible, at least more likely than on an odds on favorite. I'd guess money held as long as possible on shorter priced horses has more to do with habits and personal preferences than on a fear of screwing up an already short priced horse.

I know somebody will say I should do it myself, but I really don't care enough about it, and I don't feel "after the start betting" is prevalent anywhere in North America. As a matter of fact, aside from intentional fraud, I don't believe it exists at all. Personal opinion.

I'm suggesting this to you because it obviously drives you crazy. If the results fall near the parameters I'm guessing, I'd think that would suggest to a reasonable person that you can be comfortable that it's a by-product of data collection and reporting. Of course it won't confirm anything one way or another, but it might suggest one way or the other.

BTW, did you see the Hollendorfer that won yesterday at SA and paid 12/1. That horse entered the far turn 21/1, and then dropped to 12/1. Somebody nailed it good.

As far as when the money comes in doesn't matter to me or bettors who had the winner.....if you think you're getting one price and you get another price, it really doesn't matter if its past posting or not, all you know is you got less than you thought you were getting.

[Capper Al]

Some people may want to bet on a horse or race, but have other time commitments during pre-race and race.

These will be action bets for me when I don't have the time. I'm a mini-roller, so it doesn't really matter much ($2 vs $10 or $15).

[traynor]

Maybe because it happens so rarely?
RACE 5 - Buffalo Raceway - NY - April 10, 2015
Conditions: HORSES & GELDINGS CLAIMING $5,000
Gait: Pace Purse: $4,800 Class: HG5000CL Distance: 1 mile Track Cond: GD Temp-Allow: 48-1 Off Time: 7:49 PM


HN Horse PP 1/4 1/2 3/4 Str Finish Actual LQ Odds Driver Trainer
8 Diditagain Hanover 8 7/7 6/6 5/4H 4/1H 1/1 1:58.0 28.4 119.25 Ron Beback Jr Michelle Gramza

3 Schooner Or Later 3 4/4 3/2 3/1H 2/H 2/1 1:58.1 29.3 3.15 Shawn McDonough Russell Bratt

4 Precise Accusation 4 5/5 4/3 4/2H 5/2 3/1T 1:58.2 29.3 10.90 Jack Flanigen Alex Giuliani

6 Aberarder Smitty 6 2/1 2/1 2°/H 1/H 4/2H 1:58.2 30.0 2.25 Kevin Cummings Angelo Nappo

7 Lavern`s Art 7 6/6 5XIP/5 7/13 6/12 5/16Q 2:01.1 30.1 78.00 Shawn T Gray Lisa Lederhouse

5 Charlie Beans 5 1/1 1/1 1/H 3X/1 6/17H 2:01.2 33.0 16.40 Lee Dahn Paul Zambito Jr

2 Larrys Mist X2 8/13 7/8 6°/12H 7/15 7/24Q 2:02.4 32.0 41.75 Paul Wright Jr. Courtney Ohol

1 Release The Terror 1 X3/3 X8X/18 X8/DIS 8/DIS 8/DIS *1.00 Drew Monti Neal Russo

Time: 28.3 58.0 (29.2) 1:28.2 (30.2) 1:58.0 (29.3)

HN Horse Driver Win Place Show
8 Diditagain Hanover Ron Beback Jr 240.50 43.00 16.20
3 Schooner Or Later Shawn McDonough 4.20 2.90
4 Precise Accusation Jack Flanigen 5.40

The winner was a solid second choice to the favorite, ML 8 or 9/1. For those jaded few who still understand what advance betting dutch selections can accomplish, it must have been a memorable race, indeed.

Even something for the exotic pool cleaners (as in, "Bet $50, clean the pool.")
Exacta (8-3 $1,242.00) Trifecta (8-3-4 $3,033.00)

[ultracapper]

As far as when the money comes in doesn't matter to me or bettors who had the winner.....if you think you're getting one price and you get another price, it really doesn't matter if its past posting or not, all you know is you got less than you thought you were getting.

I fail to see your point then. So if a horse enters the gate at Even, then in the turn he drops to 4/5, and we somehow know that it was a legitimate, before the bell bet, you have issues with this? Is your issue the fact that money is coming in from all over the country and it frustrates you that it may take an additional 30 seconds to process? Are you suggesting some one should give you that Even because when the race went off it wasn't your fault the process isn't instantaneous? I thought you were troubled about past posting. Now it seems your just whining. Hell, at 3 MTP my horse is 5/1, exactly what I'm looking for. I go to the toilet and then in the next 3 minutes it goes to 7/2. I should be angry because I'm not getting 5/1? What are you talking about now?

I need to get out of this one.

[therussmeister]

SRU, you may want to do a study on this subject. It would at least either add more fuel to your concerns or alleviate them to some degree. It's obviously quite maddening to you when you see price drops after the gate opens. This is an issue you can investigate with readably available factual data. Maybe do it for a day, or a week if you really want to get into it. Watch the replays for Santa Anita next week. Make note of the odds on the favorite and winner as they enter the gate. Compare them to the odds when the race goes official. My guess, from strictly years of observation would be something like 40%drop/40%stay the same/20%go up on favorites. On winners I'd guess 30/50/20. I'd also guess the higher the odds of the winner, the more likely the odds would drop at the expense of both the other two, as big money being placed on a higher priced horse would be more likely to be held as long as possible, at least more likely than on an odds on favorite. I'd guess money held as long as possible on shorter priced horses has more to do with habits and personal preferences than on a fear of screwing up an already short priced horse.

I know somebody will say I should do it myself, but I really don't care enough about it, and I don't feel "after the start betting" is prevalent anywhere in North America. As a matter of fact, aside from intentional fraud, I don't believe it exists at all. Personal opinion.

I'm suggesting this to you because it obviously drives you crazy. If the results fall near the parameters I'm guessing, I'd think that would suggest to a reasonable person that you can be comfortable that it's a by-product of data collection and reporting. Of course it won't confirm anything one way or another, but it might suggest one way or the other.

BTW, did you see the Hollendorfer that won yesterday at SA and paid 12/1. That horse entered the far turn 21/1, and then dropped to 12/1. Somebody nailed it good.

I did a study of this in 2008 I believe. What I found was in any race with at least eight betting interests typically three horses odds dropped after the gate opened. These horses were almost always three of the top four betting favorites. Given late money was favoring three low odds horses in most races, late money was not that accurate. It was almost comical how often that the one horse out of the top four favorites that didn't get late money won.

[Dave Schwartz]

SRU, you may want to do a study on this subject. It would at least either add more fuel to your concerns or alleviate them to some degree. It's obviously quite maddening to you when you see price drops after the gate opens. This is an issue you can investigate with readably available factual data. Maybe do it for a day, or a week if you really want to get into it. Watch the replays for Santa Anita next week. Make note of the odds on the favorite and winner as they enter the gate. Compare them to the odds when the race goes official. My guess, from strictly years of observation would be something like 40%drop/40%stay the same/20%go up on favorites. On winners I'd guess 30/50/20. I'd also guess the higher the odds of the winner, the more likely the odds would drop at the expense of both the other two, as big money being placed on a higher priced horse would be more likely to be held as long as possible, at least more likely than on an odds on favorite. I'd guess money held as long as possible on shorter priced horses has more to do with habits and personal preferences than on a fear of screwing up an already short priced horse.

I did this about 5-6 years and posted results here on several occasions.

Simple rule: If the track has high rebates, then it goes like 75% of the winners will be bet down from min=0 to winner's circle.

Low (or no) rebates, such GG, SA, KEE, CD, etc, it is about 40-20-40 (up, same down).

I have on my agenda to do a more formal study but because it is time consuming I will never do a "complete" study.

[davew]

How come nobody ever lists the horses that went up in price during the running of the race? I had one Saturday at Santa Anita. Went from 5/1 to 6/1 half way down the backstretch with a length and a half lead. Finished strong and wired them. It was Glatt/TBaze in the 6th, I think.

HAHA. Only bet I made this week and I can't remember which race it was and too lazy to look. It was Glatt/TBaze and it paid $14 and change. 5/1 out of the gate and down the backstretch.

About half the horses go up in odds during the running of the race.

If you just look at the winners, about 2/3rds go down and 1/3rd go up. It is more painful when they go down, and remembered more often.

[Robert Fischer]

Certain markets have certain characteristic patterns as well.

A simple example is the show pool on many weekday and mid-level tracks.

A few minutes to post, - and the obvious, heavy-Fav is showing $2.60 or even $2.80 on the board.
You throw $10 to take the obvious overlay, and then you actually get paid $2.20 or $2.40.

After the first or second time, you realize that this is indeed a characteristic of this market, and you no longer expect $2.60 or $2.80 on obvious, heavy-Favs, regardless of what the show pools and probables look like a few minutes prior to post.


Same story on a more complicated scale, with a lot of the gripes.

[Stillriledup]

I fail to see your point then. So if a horse enters the gate at Even, then in the turn he drops to 4/5, and we somehow know that it was a legitimate, before the bell bet, you have issues with this? Is your issue the fact that money is coming in from all over the country and it frustrates you that it may take an additional 30 seconds to process? Are you suggesting some one should give you that Even because when the race went off it wasn't your fault the process isn't instantaneous? I thought you were troubled about past posting. Now it seems your just whining. Hell, at 3 MTP my horse is 5/1, exactly what I'm looking for. I go to the toilet and then in the next 3 minutes it goes to 7/2. I should be angry because I'm not getting 5/1? What are you talking about now?

I need to get out of this one.

First of all, not sure why so sensitive, does your dad run the tote companies? You should he happy this issue was being discussed, after all, if you bet horse races from time to time, this affects you too.

all i care about is getting what i see oddswise most of the time. If i see 4-1 walking to the gate, i want to get 4-1 most of the time, but it seems like i'm getting 3-1 or 5-2 most of the time. If i see horses constantly drop multiple ticks and win off, i have to question if these are shenanigans in some fashion.

[thaskalos]

First of all, not sure why so sensitive, does your dad run the tote companies? You should he happy this issue was being discussed, after all, if you bet horse races from time to time, this affects you too.

all i care about is getting what i see oddswise most of the time. If i see 4-1 walking to the gate, i want to get 4-1 most of the time, but it seems like i'm getting 3-1 or 5-2 most of the time. If i see horses constantly drop multiple ticks and win off, i have to question if these are shenanigans in some fashion.
Let it go...SRU. Why waste your time?

[traynor]

First of all, not sure why so sensitive, does your dad run the tote companies? You should he happy this issue was being discussed, after all, if you bet horse races from time to time, this affects you too.

all i care about is getting what i see oddswise most of the time. If i see 4-1 walking to the gate, i want to get 4-1 most of the time, but it seems like i'm getting 3-1 or 5-2 most of the time. If i see horses constantly drop multiple ticks and win off, i have to question if these are shenanigans in some fashion.

Not intended to be facetious. If you keep records, you should know what the particular bet will return--long term, repeated, multiple bets in same scenario--and pretty much ignore the upping and downing of the tote in individual races. That is one of the issues in the early/late dichotomy.

The difficulty is not in the tote movement--it is in confusing the running tally with the final odds available. The latter figure should be generated by your own records, and is (generally) much more reliable an indicator of the long-term potential of a given bet.

[Poindexter]

First of all, not sure why so sensitive, does your dad run the tote companies? You should he happy this issue was being discussed, after all, if you bet horse races from time to time, this affects you too.

all i care about is getting what i see oddswise most of the time. If i see 4-1 walking to the gate, i want to get 4-1 most of the time, but it seems like i'm getting 3-1 or 5-2 most of the time. If i see horses constantly drop multiple ticks and win off, i have to question if these are shenanigans in some fashion.

SRU, when I stated that the effect of whales and big rebated bettors is to drive prices down on value horses(because their motivation is to drive the price down to break even), your argument to me was that, my thinking was flawed and that the 4-1 showing on the board is not the "market price". You elaborated by insinuating that the 4-1 I want, I could never get because the new market is sharper than the old market(we are in the twitter era :lol:) and that the 5/2 or 3-1 is all I can get because that is the market price(as if the market price is the same no matter who is betting). Well guess what, back at you. The same argument you made to be applies to you here. Why, because for the most part they are the same thing. Mostly it is rebated big bettors that are driving your prices down. Why would you expect 4-1 on these horse, when the market clearly has them at 3-1 or 5/2.

Also if you want to provide compelling evidence of past posting, save the pools just prior to off and the final pools. If you suspect past posting, see how the past posted horse was bet. Was he just bet straight, or was he hit in the double pool, or exacta pool(if so with what horses-was it with everyone or just a couple of horses). Claiming that horses dropping from 4-1 to 5/2 after the race isn't going to convince anyone or anything. But show me pools with say the #7 in the 4th race who was 4-1 in the double from the 3rd to 4th and 4-1 in doubles from the 4th to 5th, he is 9-2 as they are loading the gate, breaks 2 clear, drops to 5/2 and doesn't get hit in any other pools.....that would be very interesting.

Late add on. Some horses may be 5/2 on top in the exacta pools but 4-1 in the straight pools. If you think they are magically going to remain at 4-1 dream on. Not going to happen.

[Poindexter]

Not intended to be facetious. If you keep records, you should know what the particular bet will return--long term, repeated, multiple bets in same scenario--and pretty much ignore the upping and downing of the tote in individual races. That is one of the issues in the early/late dichotomy.

The difficulty is not in the tote movement--it is in confusing the running tally with the final odds available. The latter figure should be generated by your own records, and is (generally) much more reliable an indicator of the long-term potential of a given bet.

Your skill set never ceases to amaze me. Not only can you win at will with no regards to odds, but you also know exactly what each horse will go off at. Well that would explain why you are not interested in the odds. Truly gifted.

[Stillriledup]

SRU, when I stated that the effect of whales and big rebated bettors is to drive prices down on value horses(because their motivation is to drive the price down to break even), your argument to me was that, my thinking was flawed and that the 4-1 showing on the board is not the "market price". You elaborated by insinuating that the 4-1 I want, I could never get because the new market is sharper than the old market(we are in the twitter era :lol:) and that the 5/2 or 3-1 is all I can get because that is the market price(as if the market price is the same no matter who is betting). Well guess what, back at you. The same argument you made to be applies to you here. Why, because for the most part they are the same thing. Mostly it is rebated big bettors that are driving your prices down. Why would you expect 4-1 on these horse, when the market clearly has them at 3-1 or 5/2.

Also if you want to provide compelling evidence of past posting, save the pools just prior to off and the final pools. If you suspect past posting, see how the past posted horse was bet. Was he just bet straight, or was he hit in the double pool, or exacta pool(if so with what horses-was it with everyone or just a couple of horses). Claiming that horses dropping from 4-1 to 5/2 after the race isn't going to convince anyone or anything. But show me pools with say the #7 in the 4th race who was 4-1 in the double from the 3rd to 4th and 4-1 in doubles from the 4th to 5th, he is 9-2 as they are loading the gate, breaks 2 clear, drops to 5/2 and doesn't get hit in any other pools.....that would be very interesting.

Late add on. Some horses may be 5/2 on top in the exacta pools but 4-1 in the straight pools. If you think they are magically going to remain at 4-1 dream on. Not going to happen.

No doubt, i believe that the final price is the "market price" and i don't think that has anything to do with rebaters.

It seems that horses who get "great starts" are going down in price, which means people must be refunding horses after they see a few seconds of the race, and that has nothing to do with rebates or rebate bettors.

Here's an example of what is frustrating.

There was a race at Aqu yesterday i believe, there was a horse in a middle race shipping from gulfstream with Eric Cancel riding and that horse was what i thought a "bad bet" at 1-1 ...so, i thought to myself "let me try and beat this "bad bet" and i settled on the horse that ended up winning. But, on the way to the alter, that winner went off the favorite, and halfway around the track, with my "small" win bet, i was like "cmon, really"

Here's an analogy of what's happening now in racing. You know how sometimes there are "Free items" in a bowl with a sign that says "take one"? Now, most of us just take one, right? We don't normally grab the entire bowl of items and empty all those items into our bag and leave, we leave some for other people. But, in racing, it always seems like one big bettor or a team of big bettors take all the items, if a horse is sitting up there at 2-1 and is a fair price at 2-1, these gluttons will empty the entire bowl of candy on this horse and make that horse 6-5...that's one of the bad things about racing in America, the pools are too small to absorb one glutton who decides to try and win all the money.

[arw629]

No doubt, i believe that the final price is the "market price" and i don't think that has anything to do with rebaters.

It seems that horses who get "great starts" are going down in price, which means people must be refunding horses after they see a few seconds of the race, and that has nothing to do with rebates or rebate bettors.

Here's an example of what is frustrating.

There was a race at Aqu yesterday i believe, there was a horse in a middle race shipping from gulfstream with Eric Cancel riding and that horse was what i thought a "bad bet" at 1-1 ...so, i thought to myself "let me try and beat this "bad bet" and i settled on the horse that ended up winning. But, on the way to the alter, that winner went off the favorite, and halfway around the track, with my "small" win bet, i was like "cmon, really"

Here's an analogy of what's happening now in racing. You know how sometimes there are "Free items" in a bowl with a sign that says "take one"? Now, most of us just take one, right? We don't normally grab the entire bowl of items and empty all those items into our bag and leave, we leave some for other people. But, in racing, it always seems like one big bettor or a team of big bettors take all the items, if a horse is sitting up there at 2-1 and is a fair price at 2-1, these gluttons will empty the entire bowl of candy on this horse and make that horse 6-5...that's one of the bad things about racing in America, the pools are too small to absorb one glutton who decides to try and win all the money.

I agree with SRU....I have come to accept it as the norm but he makes a good point....the rest of you are lying trolls if you don't get at least a little angry inside when you see your horse's odds drop during a race

[traynor]

Your skill set never ceases to amaze me. Not only can you win at will with no regards to odds, but you also know exactly what each horse will go off at. Well that would explain why you are not interested in the odds. Truly gifted.
No different than anyone else who uses databases to build models. I know what a particular model should return--on average--just like anyone else who builds models knows what those models should return--on average. The ups and downs of individual races are only secondarily related to those averages (of which they are comprised)--NOT the individual race decisions to bet or not bet due to odds at some point in time during the leadup to a particular race.

Nothing snarky or "gifted" at all. Just simple number crunching. If my models indicate I should get X return on Y bets over Z samples, I bet that model--regardless of the odds available at or near post time on that particular selection or selections. It has nothing whatsoever to do with a "skill set." I bet that way because it is the most profitable to me to do so.

Within the context of the thread topic, that is why (for a substantial percentage of the wagers I make) I wager early. I have never found it profitable to agonize over individual races, individual race odds, real or imaginary "biases," or whatever other information others rely on as the basis for making decisions.

The essence of wagering models is that they only becomes useful to the extent that one uses the information. Because the models are based on (relatively) large samples of races, their "advantage" only exists in large samples. Picking and choosing which of those modeled races to bet or not bet based on some ever-changing and purely arbitrary set of "new data" minutes before each race seems foolish in the extreme.

[traynor]

I agree with SRU....I have come to accept it as the norm but he makes a good point....the rest of you are lying trolls if you don't get at least a little angry inside when you see your horse's odds drop during a race

There is an alternative. I don't get a little angry inside because I (usually) bet too many races to watch each one individually, or to see (or care about) last minute odds going up or down in individual races. Some might say that takes all the fun out of wagering. I say it takes most (if not all) of the anxiety and other negative emotion out of wagering, while leaving more than enough profit to keep me (relatively) happy.

[Poindexter]

No different than anyone else who uses databases to build models. I know what a particular model should return--on average--just like anyone else who builds models knows what those models should return--on average. The ups and downs of individual races are only secondarily related to those averages (of which they are comprised)--NOT the individual race decisions to bet or not bet due to odds at some point in time during the leadup to a particular race.

Nothing snarky or "gifted" at all. Just simple number crunching. If my models indicate I should get X return on Y bets over Z samples, I bet that model--regardless of the odds available at or near post time on that particular selection or selections. It has nothing whatsoever to do with a "skill set." I bet that way because it is the most profitable to me to do so.

Within the context of the thread topic, that is why (for a substantial percentage of the wagers I make) I wager early. I have never found it profitable to agonize over individual races, individual race odds, real or imaginary "biases," or whatever other information others rely on as the basis for making decisions.

The essence of wagering models is that they only becomes useful to the extent that one uses the information. Because the models are based on (relatively) large samples of races, their "advantage" only exists in large samples. Picking and choosing which of those modeled races to bet or not bet based on some ever-changing and purely arbitrary set of "new data" minutes before each race seems foolish in the extreme.

Can you give me a hypothetical model you might use to declare a horse a wager. I cannot respond because I have no idea what you consider a "model". Off the top of my head I think you mean something like a horse who is sprinting and unable to make the lead and faltering in the lane, stretching out to 2 turns for the first time might show a positive roi. But maybe you mean something totally different-like your computer reads every variable of data and through your programming is able to come up with the winner and show a positive roi over thousands of races. Whatever the case, I still do not know how you can ignore final odds in your determination of whether something is a bet or not. No matter how positive a situation, there is always a chance the horse can be overbet for many other reasons(hot jockey, highly touted, whales like him.....) and your edge is gone.

If we were betting on the roll of dice and not knowing the mechanics of the dice, your modeling would know that 7 is the most likely outcome, but that does not mean that you could make a profit on the 7 unless your are getting above the fair odds.

[thaskalos]

No different than anyone else who uses databases to build models. I know what a particular model should return--on average--just like anyone else who builds models knows what those models should return--on average. The ups and downs of individual races are only secondarily related to those averages (of which they are comprised)--NOT the individual race decisions to bet or not bet due to odds at some point in time during the leadup to a particular race.

Nothing snarky or "gifted" at all. Just simple number crunching. If my models indicate I should get X return on Y bets over Z samples, I bet that model--regardless of the odds available at or near post time on that particular selection or selections. It has nothing whatsoever to do with a "skill set." I bet that way because it is the most profitable to me to do so.

Within the context of the thread topic, that is why (for a substantial percentage of the wagers I make) I wager early. I have never found it profitable to agonize over individual races, individual race odds, real or imaginary "biases," or whatever other information others rely on as the basis for making decisions.

The essence of wagering models is that they only becomes useful to the extent that one uses the information. Because the models are based on (relatively) large samples of races, their "advantage" only exists in large samples. Picking and choosing which of those modeled races to bet or not bet based on some ever-changing and purely arbitrary set of "new data" minutes before each race seems foolish in the extreme.

Honest question, which I hope you won't perceive as an antagonizing probe:

You have used your databases to build a model which has proven to be very successful based on a fairly large sample of races...and you are ready to bet serious money on this model's performance going forward. Since you acknowledge from the start that the "advantage" of these models only exists in large samples of races...how persistent are you willing to be in your betting until you are finally forced to admit that the promising model's profitable performance will no longer materialize in the future?

I know from prior comments of yours that you have a disdain for handicapping ideas which are backed up by insufficient proof of their effectiveness. Do you have the same disdain for quickly pulling the plug on a promising handicapping idea, going forward?

[Poindexter]

I agree with SRU....I have come to accept it as the norm but he makes a good point....the rest of you are lying trolls if you don't get at least a little angry inside when you see your horse's odds drop during a race

We all get a little angry when our horse drops after we bet him late, but there is nothing we can do about it. The main question is why is this happening. It is always going to happen to some degree because the odds will correct(become more efficient) as more money comes into the pools. If this is happening because of past posting this is a MAJOR concern. I have no idea whether it is happening or not, but did provide a few things you can look at to figure out whether it might be going on. I believe for the most part this is happening because rebated big bettors are pounding horses down as far as they can and earning off of the rebates. Thus the reason why I constantly say that rebates are the death of this game. However, there is no way to know for sure that there is going to be no past posting(or cancelling) unless they close the pools earlier and do not start the race until after the final monies are shown in the pools. This could be done no problem, bettors will adjust, but racetracks and horseman know this will cost them handle so they will never agree to do it, unless horseplayers make a strong push for it. Do not think that enough horseplayers are concerned enough(or think it is a legitimate problem) for that to ever happen.

[thaskalos]

We all get a little angry when our horse drops after we bet him late, but there is nothing we can do about it. The main question is why is this happening. It is always going to happen to some degree because the odds will correct(become more efficient) as more money comes into the pools. If this is happening because of past posting this is a MAJOR concern. I have no idea whether it is happening or not, but did provide a few things you can look at to figure out whether it might be going on. I believe for the most part this is happening because rebated big bettors are pounding horses down as far as they can and earning off of the rebates. Thus the reason why I constantly say that rebates are the death of this game. However, there is no way to know for sure that there is going to be no past posting(or cancelling) unless they close the pools earlier and do not start the race until after the final monies are shown in the pools. This could be done no problem, bettors will adjust, but racetracks and horseman know this will cost them handle so they will never agree to do it, unless horseplayers make a strong push for it. Do not think that enough horseplayers are concerned enough(or think it is a legitimate problem) for that to ever happen.

Have you any idea when this "odds-dropping-after-the bell" started becoming as prevalent as it now seems to be? Do you recall seeing it, say...15-20 years ago?

[ultracapper]

Naturally, you'd think it would have started when tracks began collecting monies into their pools from a thousand different points. The processing time, if bettors are allowed to bet right up to the start, will lead to final pool collections after the start of the race.

Didn't Dave say he did a study covering a couple seasons of data that showed a 40% drop, 20% no change, and 40% rise in prices of winners some time in the past 15 years? It sounds like he did pretty much what I had suggested SRU take a look at.

We're always going to notice the droppers, particularly if we have bet them, and it's going to, at minimum, disappoint us, and in some cases as we've seen here, infuriate us. But IMHO, I don't think it's anything nefarious, and I don't believe, again IMHO, that it's necessarily as one sided as the more angry of us are making it out to be. That's what Dave's results would indicate anyway.

[thaskalos]

Naturally, you'd think it would have started when tracks began collecting monies into their pools from a thousand different points. The processing time, if bettors are allowed to bet right up to the start, will lead to final pool collections after the start of the race.

Didn't Dave say he did a study covering a couple seasons of data that showed a 40% drop, 20% no change, and 40% rise in prices of winners some time in the past 15 years? It sounds like he did pretty much what I had suggested SRU take a look at.

We're always going to notice the droppers, particularly if we have bet them, and it's going to, at minimum, disappoint us, and in some cases as we've seen here, infuriate us. But IMHO, I don't think it's anything nefarious, and I don't believe, again IMHO, that it's necessarily as one sided as the more angry of us are making it out to be. That's what Dave's results would indicate anyway.

You would think that...but that's not when it started. I remember the early days of full-card simulcasting...and no one talked about those infuriating odds drop-downs then.

By the way...I think you read Dave's report all wrong.

[traynor]

Can you give me a hypothetical model you might use to declare a horse a wager. I cannot respond because I have no idea what you consider a "model". Off the top of my head I think you mean something like a horse who is sprinting and unable to make the lead and faltering in the lane, stretching out to 2 turns for the first time might show a positive roi. But maybe you mean something totally different-like your computer reads every variable of data and through your programming is able to come up with the winner and show a positive roi over thousands of races. Whatever the case, I still do not know how you can ignore final odds in your determination of whether something is a bet or not. No matter how positive a situation, there is always a chance the horse can be overbet for many other reasons(hot jockey, highly touted, whales like him.....) and your edge is gone.

If we were betting on the roll of dice and not knowing the mechanics of the dice, your modeling would know that 7 is the most likely outcome, but that does not mean that you could make a profit on the 7 unless your are getting above the fair odds.

Substitute "spot play" for "model" and it will be close. Not that simple, but it is a better simile than most other terms (familiar to race bettors). Corrected for outliers (not reverse engineered from outliers), the "average mutuel" over some large group of races will be such and such. Some up, some down, but in a (relatively) long run, such and such. That is the "expected vaue" figure I use, rather than the odds in individual races--which may or may not be accurate at the time wagers must be made.

[traynor]

Honest question, which I hope you won't perceive as an antagonizing probe:

You have used your databases to build a model which has proven to be very successful based on a fairly large sample of races...and you are ready to bet serious money on this model's performance going forward. Since you acknowledge from the start that the "advantage" of these models only exists in large samples of races...how persistent are you willing to be in your betting until you are finally forced to admit that the promising model's profitable performance will no longer materialize in the future?

I know from prior comments of yours that you have a disdain for handicapping ideas which are backed up by insufficient proof of their effectiveness. Do you have the same disdain for quickly pulling the plug on a promising handicapping idea, going forward?

You might also recall that I have mentioned previously (several times, at length) that I bet in two fundamentally different ways. "Serious" bets are usually onsite, where I can observe the entries in the flesh, and are contingent on both observation and trip notes (my own and those of others)--in addition to computer models.

The high-volume, bread-and-butter races are lesser amounts, spread over many races. The models used in that type of wagering are coded to self-monitor, and flag anomalies. In plain language, that means the models continually "compare themselves" to expected outcomes, and respond to pre-programmed value ranges. You may recall from our previous exchanges that I said if a week went by and the model(s) in use at the time did NOT show a profit, I would stop betting (that model or those models ) until I discovered whether or not a problem existed with it (them). "WeeK" is a simplification--each model has a different range of acceptable values, and possibly a different set of ranges.

[thaskalos]

You might also recall that I have mentioned previously (several times, at length) that I bet in two fundamentally different ways. "Serious" bets are usually onsite, where I can observe the entries in the flesh, and are contingent on both observation and trip notes (my own and those of others)--in addition to computer models.

The high-volume, bread-and-butter races are lesser amounts, spread over many races. The models used in that type of wagering are coded to self-monitor, and flag anomalies. In plain language, that means the models continually "compare themselves" to expected outcomes, and respond to pre-programmed value ranges. You may recall from our previous exchanges that I said if a week went by and the model(s) in use at the time did NOT show a profit, I would stop betting (that model or those models ) until I discovered whether or not a problem existed with it (them). "WeeK" is a simplification--each model has a different range of acceptable values, and possibly a different set of ranges.
I remember our prior exchanges...that's why I asked the question. Is a week what might be construed as a "fairly large sample size" to determine failure?

[Poindexter]

Have you any idea when this "odds-dropping-after-the bell" started becoming as prevalent as it now seems to be? Do you recall seeing it, say...15-20 years ago?\

Remembering a week ago is challenge for me, so remembering details about 20 and 30 years ago is quite a tall task. That being said, when I was younger, I usually lined races and bet value and on occasion I would be disappointed that a horse I bet would be pounded late (and I bet very close to off time), but it was never a big concern of mine(so that should answer the question to some degree). Back then I spend most of my time betting the harness races(mostly Los Alamitos). But after Harness left Southern California I did transition to Thoroughbreds. Back then Holywood Park especially on weekdays would have some late money crushes, but once again I do not recall it ever become overly frustrating. But it still comes down to rebates. Back then the large bettor had to leave himself some room for profit. Today, the profits are his rebates. As public handles dip, rebated bettors consist of a higher percentage of the pools and with no incentive to leave a margin for profit up on the board, why would anyone expect anyone making an income doing this to leave any value up on the board. It just will not happen. In general the horseplayers best chance in this game is fighting the other squirrels for the nuts(live horses the big bettors miss) or hitting the super exotic pools where the big bettors typically will not be and getting lucky or exceeding your game to the level that you are sharper than the very sharp(good luck with that-seems like ultracapper might be in that category).

[traynor]

I remember our prior exchanges...that's why I asked the question. Is a week what might be construed as a "fairly large sample size" to determine failure?

Models predict trends. Trends should be repetitive. Understand that we are talking about win bets, and cold exactas. I don't use this type of model for more arcane exotic wagers. You may also recall me mentioning that certain models generated (consistently) win rates exceeding 50%, some (unfortunately, all too few) up in the 70% range.

It is definitely NOT a case of shovelling money into the pools to chase once-a-week (or whatever) hits. An example might be a 50% win rate and an ROI of howevermuch. Either might fall off and diminish the other, but I am still getting numerous hits. My "deficient" model may not be producing as expected, but there are no long strings of losses that I have to eat when I decide to stop betting that model.

Again, it is more like blackjack. Returns may be less than expected over some given group of outcomes, but not break-the-piggybank less. More on the order of 90-95% (or higher) return, rather than the expected profit.

[thaskalos]

Back then the large bettor had to leave himself some room for profit. Today, the profits are his rebates. As public handles dip, rebated bettors consist of a higher percentage of the pools and with no incentive to leave a margin for profit up on the board, why would anyone expect anyone making an income doing this to leave any value up on the board. It just will not happen. In general the horseplayers best chance in this game is fighting the other squirrels for the nuts(live horses the big bettors miss) or hitting the super exotic pools where the big bettors typically will not be and getting lucky or exceeding your game to the level that you are sharper than the very sharp(good luck with that-seems like ultracapper might be in that category).
I agree with everything that you say here. :ThmbUp:

The competent rebated player is gradually increasing his position within the game as the less successful non-rebated players are dropping off the map...and this trend is sure to continue into the future. Pretty soon...the non-rebated player will be extinct.

[thaskalos]

Models predict trends. Trends should be repetitive. Understand that we are talking about win bets, and cold exactas. I don't use this type of model for more arcane exotic wagers. You may also recall me mentioning that certain models generated (consistently) win rates exceeding 50%, some (unfortunately, all too few) up in the 70% range.

It is definitely NOT a case of shovelling money into the pools to chase once-a-week (or whatever) hits. An example might be a 50% win rate and an ROI of howevermuch. Either might fall off and diminish the other, but I am still getting numerous hits. My "deficient" model may not be producing as expected, but there are no long strings of losses that I have to eat when I decide to stop betting that model.

Again, it is more like blackjack. Returns may be less than expected over some given group of outcomes, but not break-the-piggybank less. More on the order of 90-95% (or higher) return, rather than the expected profit.
I think I understand what you say...but I am still confused by some of it. You are a stickler for "large samples", and you insist that the advantage of your models can only reveal itself when the sample size is "fairly large", and yet, you pull the plug and stop wagering when your models register a losing week at the windows. Is that what gambling "success" means to you? Winning every single week?

I remember in one of our previous exchanges, where you declared that Ian Andersen's entire gambling career may be statistically insignificant because it didn't encompass millions and millions of blackjack hands...and I agreed with you. But how can a blackjack career which spans years be considered "statistically insignificant"...when we attach statistical significance to a week of losses at the track?

[traynor]

I think I understand what you say...but I am still confused by some of it. You are a stickler for "large samples", and you insist that the advantage of your models can only reveal itself when the sample size is "fairly large", and yet, you pull the plug and stop wagering when your models register a losing week at the windows. Is that what gambling "success" means to you? Winning every single week?

I remember in one of our previous exchanges, where you declared that Ian Andersen's entire gambling career may be statistically insignificant because it didn't encompass millions and millions of blackjack hands...and I agreed with you. But how can a blackjack career which spans years be considered "statistically insignificant"...when we attach statistical significance to a week of losses at the track?

That is an oversimplification. "Large samples" do not necessarily mean thousands of races. Models are not simple. Each contains details of what "reality" should look like. If that reality differs significantly from what it is supposed to look like, it is prudent to ask why it should be so. If a model indicates a relatively normal distribution, a 50% win rate, and a healthy ROI, a string of 30 losses in a row might well be considered cause for alarm. For an essentially similar model with a different distribution and win rate, a string of 30 losses might be perfectly normal. It all depends on the model, the distribution, and the patterns involved. That is all part of the modeling process.

I think you may be confusing modeling and "statistical significance." I make predictive models (what will happen), based on descriptive models (of what has already happened). Distribution of outcomes is a key component in predictive models, and it is deviation from expected distribution--rather than "statically significant sample size"--that determines whether a model is performing correctly.

[traynor]

All my models have a missout counter. It is incremented 1 each time a race matches certain criteria, and, if that number is greater than a total counter, the missout counter value is transferred to the total counter. Each win resets the missout counter to zero.

With a couple of simple lines of code, a value is generated that indicates the maximum number of consecutive losses when that pattern was applied to a "statistically significant" sample of races. If I have a model that kicks out 50% plus winners with a useful number of matches over a sample of many races, with a maximum string of consecutive losses never greater than 6, when I see a string of 20-30 losses, I want to know why.

[thaskalos]

This does not explain why an unprofitable week should be a concern to you...especially when your models have as their base a large sample of races...which they must have, given your apparent disdain for small samples. I know from our prior exchanges that you "really, really, really hate to lose"...but a week's time cannot be considered to be anything but the scantest of short terms...as far as gambling is concerned. A week's profits mean nothing to the long-term bettor...and neither do a week's losses.

[traynor]

This does not explain why an unprofitable week should be a concern to you...especially when your models have as their base a large sample of races...which they must have, given your apparent disdain for small samples. I know from our prior exchanges that you "really, really, really hate to lose"...but a week's time cannot be considered to be anything but the scantest of short terms...as far as gambling is concerned. A week's profits mean nothing to the long-term bettor...and neither do a week's losses.

Perhaps, according to the way you bet. Not at all according to the way I bet.

[thaskalos]

Perhaps, according to the way you bet. Not at all according to the way I bet.

What...a week's results are not "short-term" to you?

[Harvhorse]

I`m a value player, I back my own opinions, If I make a horse say 4/1 and he`s 10/1 on the board I bet him . Early or late I back my opinon.

[thaskalos]

I`m a value player, I back my own opinions, If I make a horse say 4/1 and he`s 10/1 on the board I bet him . Early or late I back my opinon.

What if the horse looks like an obvious 9/5...and it's sitting there at 6/1?

[DeltaLover]

I`m a value player, I back my own opinions, If I make a horse say 4/1 and he`s 10/1 on the board I bet him . Early or late I back my opinon.

Unfortunately, in the vast majority of the cases you are referring here, you are not having an overlay but a miserable horse who will have hard time to just finish the race...

[Stillriledup]

We all get a little angry when our horse drops after we bet him late, but there is nothing we can do about it. The main question is why is this happening. It is always going to happen to some degree because the odds will correct(become more efficient) as more money comes into the pools. If this is happening because of past posting this is a MAJOR concern. I have no idea whether it is happening or not, but did provide a few things you can look at to figure out whether it might be going on. I believe for the most part this is happening because rebated big bettors are pounding horses down as far as they can and earning off of the rebates. Thus the reason why I constantly say that rebates are the death of this game. However, there is no way to know for sure that there is going to be no past posting(or cancelling) unless they close the pools earlier and do not start the race until after the final monies are shown in the pools. This could be done no problem, bettors will adjust, but racetracks and horseman know this will cost them handle so they will never agree to do it, unless horseplayers make a strong push for it. Do not think that enough horseplayers are concerned enough(or think it is a legitimate problem) for that to ever happen.

Here's the problem with the idea that rebated bettors are pounding horses down and earning off rebates. I'm pretty sure i went over the math on this, but here goes again.
The only way your theory could work is if the big bettor is finding short priced horses who are "Destined" to go off at overlays and the rebate bettor is the only rebate bettor who knows this and he bets enough to win where the "overlay" juice is all sucked out and the horse ends up going off at exactly market value. The rebate player isn't looking to make large bets on horses who are already at "market value" because the rebate on a win bet is somewhere in the 5% range. That means for every 100 bucks this guy bets, he gets 5 bucks in rebate. That's not enough to offset knocking horses to be "bad bets".

If the bettor sticks 6k to win on a horse and that horse pays 3.80 to win, when that horse's proper price was 4.00 (assuming for arguments sake that 6k to win will only knock down a 1-1 to 4-5), the guy costs himself 600 bucks in "knockdown" but only makes up 300 bucks in rebate. The math doesn't work here. The horse would have to pay 3.90 just for this guy to "break even" on his large bet and we know that horses only pay 3.90 in NY, and even at 3.90, the guy isn't gaining anything in profit and we know that a 6k win bet would make a 4 dollar horse go much lower than 3.90 in over 90% of the races run in this country.

[traynor]

What...a week's results are not "short-term" to you?

Again, I think you may be confusing modeling and "database handicapping"--two very, very different things. The criteria for a model include how accurately it represents reality--a specific set of definitions. If the model does NOT accurately represent reality, it should not take an extended period of losses to determine that.

Anomalies are always possible, meaning the problem may be in variations within a new set of outcomes that did not exist in the sample used for modeling. Bingo! The model does NOT represent reality--it only represents the limited subset of reality that was comprised of the outcomes in the sample used.

It has little or nothing to do with any conventional (or unconventional) notions of what wagering should or should not be. It is only concerned with whether or not the model accurately represents reality, which in turn is based on how normal the distribution of outcomes was in the sample used to create the model.

[Poindexter]

Here's the problem with the idea that rebated bettors are pounding horses down and earning off rebates. I'm pretty sure i went over the math on this, but here goes again.
The only way your theory could work is if the big bettor is finding short priced horses who are "Destined" to go off at overlays and the rebate bettor is the only rebate bettor who knows this and he bets enough to win where the "overlay" juice is all sucked out and the horse ends up going off at exactly market value. The rebate player isn't looking to make large bets on horses who are already at "market value" because the rebate on a win bet is somewhere in the 5% range. That means for every 100 bucks this guy bets, he gets 5 bucks in rebate. That's not enough to offset knocking horses to be "bad bets".

If the bettor sticks 6k to win on a horse and that horse pays 3.80 to win, when that horse's proper price was 4.00 (assuming for arguments sake that 6k to win will only knock down a 1-1 to 4-5), the guy costs himself 600 bucks in "knockdown" but only makes up 300 bucks in rebate. The math doesn't work here. The horse would have to pay 3.90 just for this guy to "break even" on his large bet and we know that horses only pay 3.90 in NY, and even at 3.90, the guy isn't gaining anything in profit and we know that a 6k win bet would make a 4 dollar horse go much lower than 3.90 in over 90% of the races run in this country.

Let me clarify. When I say pounding a horse down as far as they can, I mean betting as much as they can on a horse without knocking him below a certain price. So hypothetically, I am Poindexter the whale and I like the 3 in the 4th at SRU downs. I consider him a true 2-1 shot(fair value). Off my long history of races bet at SRU downs, my computers have determined that it is best to bet the precise amount that will knock him down to 2.32-1(because other bettors will likely bet him as well). How they estimate this figure I will leave to those a lot more knowledgable in this subject than I am. So I see the same horse SRU does at 4-1 when the horses are loading the gate, and I bet just enough to bet him down to 2.32-1, SRU and the rest of the crowd bet as well and as the horses are turning for home the horse dives from 4-1 down to 2-1. He is clear by 3 1/8th into the race and after the race SRU says to himself another past post on my easy lead winner.

I do not know how many big bettors are betting below fair value(it likely depends on the size of their rebates). I believe there are some double digit rebates given in some pools(usually the very high take exotic pools) at some tracks, so the big bettors in these pools probably are betting their combos below fair value. I have no idea what the top rebate for win bettors are. But whatever it is the larger the rebate the more likely they are betting to below fair value.

Also, if they are able to read bets in transit, than they can bet down even further and with more precision. So while you see 4-1 on the board, there computer knows as the horses are loading the gate that he is currently 3.12-1 and they can bet appropriately off that knowledge.

[Sapio]

Again, I think you may be confusing modeling and "database handicapping"--two very, very different things. The criteria for a model include how accurately it represents reality--a specific set of definitions. If the model does NOT accurately represent reality, it should not take an extended period of losses to determine that. On this subject, is the inverse true? Does a small sample of wins represent reality?

Anomalies are always possible, meaning the problem may be in variations within a new set of outcomes that did not exist in the sample used for modeling. Bingo! The model does NOT represent reality--it only represents the limited subset of reality that was comprised of the outcomes in the sample used.

It has little or nothing to do with any conventional (or unconventional) notions of what wagering should or should not be. It is only concerned with whether or not the model accurately represents reality, which in turn is based on how normal the distribution of outcomes was in the sample used to create the model. Can you explain this further?

Thomas Sapio

[traynor]

Thomas Sapio

Because of the way I build the models, I expect certain components (# of positives, win%, others) to be relatively stable. Because I use cleaned data, truncated outliers, etc., the ROI should be better than anticipated, but rarely less. As to your specific question, a cluster of wins might be an anomaly (a nice one), but I also monitor that frequency. Meaning the distribution of wins and losses should be "normal" for that specific model. An unusual number (or distribution) of wins would be considered as much an aberration as an unusual number (or distribution) of losses.

If I have a "large" sample of races, the individual outcomes in that sample form a pattern. If that pattern is consistent, it should be replicated (approximately) in (randomly) selected subsets of the original data, in a different set of data, and (if predictive) in future sets of (the same generic type of) data.

[traynor]

A good intro to the whys and wherefores of data mining and model building.

https://www.youtube.com/watch?v=m7kpIBGEdkI

Not exactly the equivalent of a graduate degree in business analytics, but cheaper and a lot less time consuming.

[Sapio]

Because of the way I build the models, I expect certain components (# of positives, win%, others) to be relatively stable. Because I use cleaned data, truncated outliers, etc., the ROI should be better than anticipated, but rarely less. As to your specific question, a cluster of wins might be an anomaly (a nice one), but I also monitor that frequency. Meaning the distribution of wins and losses should be "normal" for that specific model. An unusual number (or distribution) of wins would be considered as much an aberration as an unusual number (or distribution) of losses.

If I have a "large" sample of races, the individual outcomes in that sample form a pattern. If that pattern is consistent, it should be replicated (approximately) in (randomly) selected subsets of the original data, in a different set of data, and (if predictive) in future sets of (the same generic type of) data.

Hi Traynor,

Thanks for the explanation.

Thomas Sapio

[ReplayRandall]

Because of the way I build the models, I expect certain components (# of positives, win%, others) to be relatively stable. Because I use cleaned data, truncated outliers, etc., the ROI should be better than anticipated, but rarely less. As to your specific question, a cluster of wins might be an anomaly (a nice one), but I also monitor that frequency. Meaning the distribution of wins and losses should be "normal" for that specific model. An unusual number (or distribution) of wins would be considered as much an aberration as an unusual number (or distribution) of losses.

If I have a "large" sample of races, the individual outcomes in that sample form a pattern. If that pattern is consistent, it should be replicated (approximately) in (randomly) selected subsets of the original data, in a different set of data, and (if predictive) in future sets of (the same generic type of) data.
I say this to you Traynor with all sincerity, this is one your FINEST posts.. :ThmbUp:

[seanm]

A good intro to the whys and wherefores of data mining and model building.

https://www.youtube.com/watch?v=m7kpIBGEdkI

Not exactly the equivalent of a graduate degree in business analytics, but cheaper and a lot less time consuming.

:ThmbUp: :ThmbUp: for this and your previous posts.

wonderful insights

sm