Computer groups...how much fat is left on the bone

[teddy]

Interesting story came my way from a reliable source. Rebate agent hires "computer boy to build build a system to find overlaid exotics" Now computer boy who was broke has millions in the bank, all on rebates. Now that makes at least 3 groups I know of that are doing that, scraping the totes and comparing fair payoffs. Running automatic programs that are now giving out the same results. Im assuming they are betting exactas and maybe win bets. This game seems like it is really heading down hill for the player without the computer because we dont know the real fair odds. Most of us play far to fast to calculate exacta odds. So were playing against guys that have that calculated for them. Zero chance of getting an overlaid exacta if the computer groups are playing that track...

[PhantomOnTour]

The amount of actual handicapping that people do is diminishing rapidly. That, or what we used to call handicapping just isn't the same thing anymore.
Databasers play (and look for) long term auto plays, and these guys you are talking about probably never even pick up a form or a set of pp's.

Scary

[maddog42]

Interesting story came my way from a reliable source. Rebate agent hires "computer boy to build build a system to find overlaid exotics" Now computer boy who was broke has millions in the bank, all on rebates. Now that makes at least 3 groups I know of that are doing that, scraping the totes and comparing fair payoffs. Running automatic programs that are now giving out the same results. Im assuming they are betting exactas and maybe win bets. This game seems like it is really heading down hill for the player without the computer because we dont know the real fair odds. Most of us play far to fast to calculate exacta odds. So were playing against guys that have that calculated for them. Zero chance of getting an overlaid exacta if the computer groups are playing that track...

There is no such thing as a perfect odds line. These things are only opinions.
Some opinions are better than others.
" Running automatic programs that are now giving out the same results." This is really the only part of your post I disagreed with. I doubt very seriously that these programs are giving the same results. There would be some overlap.
Do any two programs pick pace/running lines the same way? I doubt it. Do they each weigh the variables the same way? Did they run the same set of races to get the same impact value for trainers off layoff ? Almost impossible.

[Ian Meyers]

Interesting story came my way from a reliable source. Rebate agent hires "computer boy to build build a system to find overlaid exotics" Now computer boy who was broke has millions in the bank, all on rebates. Now that makes at least 3 groups I know of that are doing that, scraping the totes and comparing fair payoffs. Running automatic programs that are now giving out the same results. Im assuming they are betting exactas and maybe win bets. This game seems like it is really heading down hill for the player without the computer because we dont know the real fair odds. Most of us play far to fast to calculate exacta odds. So were playing against guys that have that calculated for them. Zero chance of getting an overlaid exacta if the computer groups are playing that track...

Yeah, it really is that easy. Only took him 3 and 1/2 minutes to come up with the model. He has already won about $4 billion betting. We should all do it. ;)

Come on now...The reality is winning teams have sunk seven figures into research, testing, etc. over the course of many years. ANY endeavor that lucrative requires tons of hard work and significant investment whether it's time or money or both.

[Some_One]

Thanks Ian for the voice of reason, lol...I'm sure there is a survivorship bias at play to, I'm sure a bad run off the start have sunk many teams

[dansan]

there are plenty of decent exacta payouts just need to know where to look and you dont need a puter

[teddy]

Yeah, it really is that easy. Only took him 3 and 1/2 minutes to come up with the model. He has already won about $4 billion betting. We should all do it. ;)

Come on now...The reality is winning teams have sunk seven figures into research, testing, etc. over the course of many years. ANY endeavor that lucrative requires tons of hard work and significant investment whether it's time or money or both.


Actually I heard a previous group paid 300k for their program. This person I spoke of I cant tell you how long it took him to do what he did but he did it. Probably continues to fine tune it. Last I heard hes screaming to just play for the rebate. I know another that did the math in his head... and the guys that took his bets own mansions. He however did not handle success so well. I cant fathom how anyone that owned an adw could do anything but prosper with 3% takeout. Yes I know adw have expenses, but if your players paid for that then you should be a rich player real quick.. Imagine 15% rebate on wps or even 12%....

[teddy]

I would say the computer groups pay less than 1% to the adw if that.

[Stillriledup]

Interesting story came my way from a reliable source. Rebate agent hires "computer boy to build build a system to find overlaid exotics" Now computer boy who was broke has millions in the bank, all on rebates. Now that makes at least 3 groups I know of that are doing that, scraping the totes and comparing fair payoffs. Running automatic programs that are now giving out the same results. Im assuming they are betting exactas and maybe win bets. This game seems like it is really heading down hill for the player without the computer because we dont know the real fair odds. Most of us play far to fast to calculate exacta odds. So were playing against guys that have that calculated for them. Zero chance of getting an overlaid exacta if the computer groups are playing that track...

But if one exacta is overlaid and the computer boys bet it down, that means another exacta goes UP in price. Same thing with win, they stick a few grand to win on someone, other horses go up in price.

The one thing i don't understand is that if you bet on a 3-1 shot in 1980 (before rebates, computer boys, etc) you get 8 dollars and in 2011 if you bet on a 3-1 shot you still get 8 dollars. Nothing has changed as far as YOU are concerned. 3-1 shots still pay 8 bucks regardless of who's betting on them or how they're betting on them.

[classhandicapper]

Some of these advanced stats guys are off the charts bright. So I'm sure the combination of huge databases, their statistics and probability knowledge, plus computers programmed to quickly look for value and exploit pools they are making it tougher on the rest of us. I know I think it's way tougher than it was 20 years ago.

However, when you automate and get very stats oriented, you tend to miss some visual things and misunderstand some subsets that can't be evaluated well without some subjective human analysis.

I'm a huge basketball fan and gamble on that game also. I've spent several years discussing advanced stats for basketball with some absolutely brilliant people. I've learned a ton. But sometimes they spend so much time with their spreadsheets and computers without actually watching the games they miss some very obvious aspects of the game. So for every area of brilliance there is a blind spot.

So perhaps for every area of the game they are making it tougher to find value, they are creating a new area where there is more value. We just have to find it.

[Stillriledup]

Some of these advanced stats guys are off the charts bright. So I'm sure the combination of huge databases, their statistics and probability knowledge, plus computers programmed to quickly look for value and exploit pools they are making it tougher on the rest of us. I know I think it's way tougher than it was 20 years ago.

However, when you automate and get very stats oriented, you tend to miss some visual things and misunderstand some subsets that can't be evaluated well without some subjective human analysis.

I'm a huge basketball fan and gamble on that game also. I've spent several years discussing advanced stats for basketball with some absolutely brilliant people. I've learned a ton. But sometimes they spend so much time with their spreadsheets and computers without actually watching the games they miss some very obvious aspects of the game. So for every area of brilliance there is a blind spot.

So perhaps for every area of the game they are making it tougher to find value, they are creating a new area where there is more value. We just have to find it.

Excellent stuff CH.

[Overlay]

But if one exacta is overlaid and the computer boys bet it down, that means another exacta goes UP in price. Same thing with win, they stick a few grand to win on someone, other horses go up in price.

The one thing i don't understand is that if you bet on a 3-1 shot in 1980 (before rebates, computer boys, etc) you get 8 dollars and in 2011 if you bet on a 3-1 shot you still get 8 dollars. Nothing has changed as far as YOU are concerned. 3-1 shots still pay 8 bucks regardless of who's betting on them or how they're betting on them.
I agree. As long as no horse or combination can be determined in advance to have a 100% guaranteed chance of success, the pari-mutuel system will continue to produce value somewhere. The main challenge in playing against computers is maintaining a sufficiently wide range of visibility to deal with or compensate for the effects of last-second (or later :rolleyes: ) wagers and odds changes.

[Robert Fischer]

But if one exacta is overlaid and the computer boys bet it down, that means another exacta goes UP in price. Same thing with win, they stick a few grand to win on someone, other horses go up in price.


pretty sure they blend tote odds with a power ranking to come up with their best guess at FAIR ODDS

the TOTE ODDS BLEND they use here is implied to be from "inefficiencies" across pools.

now this is is the part pertaining to your comment - they calculate a conservative estimate as to how much they can bet while maintaining an inefficiency.

[Robert Fischer]

Some of these advanced stats guys are off the charts bright. So I'm sure the combination of huge databases, their statistics and probability knowledge, plus computers programmed to quickly look for value and exploit pools they are making it tougher on the rest of us. I know I think it's way tougher than it was 20 years ago.

However, when you automate and get very stats oriented, you tend to miss some visual things and misunderstand some subsets that can't be evaluated well without some subjective human analysis.

I'm a huge basketball fan and gamble on that game also. I've spent several years discussing advanced stats for basketball with some absolutely brilliant people. I've learned a ton. But sometimes they spend so much time with their spreadsheets and computers without actually watching the games they miss some very obvious aspects of the game. So for every area of brilliance there is a blind spot.

So perhaps for every area of the game they are making it tougher to find value, they are creating a new area where there is more value. We just have to find it.

interesting tangent = there's an academic paper done by a bright guy which pretty much sets out to (and most academic circles believes that he accomplishes) disproving "streaks" in Free Throw shooting.
sure it's in google to be found, maybe if i get a free minute / looking unlikely...

anyway it's very scientific and uses probability very well but if i remember correctly in the end it perverts the definition of streaks and is so faithful in belief of the science that it loses touch with the actual model and therefore unintentionally becomes pseudo-science.

[davew]

glad they are doing well


the problem with most valur drivern bets or overlays is the percentage of the handle that shows up after it is too late to do anything about it


I was shocked last week when watching a PRM race last week, and over 80% of the handle showed on the tote after the race started and it was too late to bet


The parimutual system is ideal for taking away from the guys who increase their betting to the point of lowering their odds to unprofitability


how high a rebate can be found, and where? the track is usually taking near 25% of all handle in exactas

[Stillriledup]

glad they are doing well


the problem with most valur drivern bets or overlays is the percentage of the handle that shows up after it is too late to do anything about it


I was shocked last week when watching a PRM race last week, and over 80% of the handle showed on the tote after the race started and it was too late to bet


The parimutual system is ideal for taking away from the guys who increase their betting to the point of lowering their odds to unprofitability


how high a rebate can be found, and where? the track is usually taking near 25% of all handle in exactas


There's gotta be more 'computer' players and if they all decide at the last second that horse X is value, then that horse could be negative value if everyone gets the 'same idea' at the same time.

[davew]

It could have been



it was a 330 yd quarterhorse race with an early bet heavy favorite


I think some of the smaller handle circuits have a larger percentage of simultcast handle


I remember reading years ago that some mid-week Aqueduct races or race days (can't remember) had more money going out to all the ontrack bettors than they had bet (combined)- since they did so much better than all the simultcast money

[Robert Fischer]

There's gotta be more 'computer' players and if they all decide at the last second that horse X is value, then that horse could be negative value if everyone gets the 'same idea' at the same time.

good point this likely to happen at times. :ThmbUp:

[TheGhostOfOscarB]

I'm a little bit skeptical about these stories , but I imagine its possible to scrape the pool in such a way as to guarantee some profit WITH REBATES.

I know more than a little bit about the exacta pool and I know some of these stories don't hold much water.

Without an actual opinion and knowledge about the race outcome (ie horseplaying) simply banging out tickets on supposed overlays is a fast way to the poorhouse.

[Saratoga_Mike]

I'm a little bit skeptical about these stories , but I imagine its possible to scrape the pool in such a way as to guarantee some profit WITH REBATES.

I know more than a little bit about the exacta pool and I know some of these stories don't hold much water.

Without an actual opinion and knowledge about the race outcome (ie horseplaying) simply banging out tickets on supposed overlays is a fast way to the poorhouse.

It's 40 years of experience, right, Oscar?

[teddy]

Whats so hard to believe that people can figure out inefficiencies in the tote and are playing on 4% takeout on superfectas....at some tracks..

[HUSKER55]

what kind of a strike rate and roi do you suspect these guys are getting with their computers?

If I bet 9 races a day and they bey 3000 of course they will make more money. But what about track a compared to track a.

how well do they stack up?

[thaskalos]

what kind of a strike rate and roi do you suspect these guys are getting with their computers?

If I bet 9 races a day and they bey 3000 of course they will make more money. But what about track a compared to track a.

how well do they stack up?
The chief advantage of successful computer wagering is not found in its "strike rate"...or its "ROI".

Computer wagering takes the "emotion" out of the betting process...and sidesteps the many "psychological" traps so prevalent in this game.

It puts an end to the haphazard betting style that sinks the vast majority of the bettors out there...

[teddy]

The chief advantage of successful computer wagering is not found in its "strike rate"...or its "ROI".

Computer wagering takes the "emotion" out of the betting process...and sidesteps the many "psychological" traps so prevalent in this game.

It puts an end to the haphazard betting style that sinks the vast majority of the bettors out there...

I agree totally. Its computed edge. Seems like I heard where they do actually handicapping in one room and those horses are plugged into the program. No one in the group knows what the others are doing. Task specific. Then its a wager or no wager based on if an edge exists.

[Overlay]

The chief advantage of successful computer wagering is not found in its "strike rate"...or its "ROI".

Computer wagering takes the "emotion" out of the betting process...and sidesteps the many "psychological" traps so prevalent in this game.

It puts an end to the haphazard betting style that sinks the vast majority of the bettors out there...
That aspect can also be a two-edged sword that misses qualitative/intuitive observations that can't be reduced to data points. However, I think it would be difficult to incorporate the effect of those nuances on a consistently accurate basis from one race to the next (not to mention the extra handicapping time they consume). All told (and as you point out), I consider their marginal value to be outweighed by the advantages of bulk objectivity.

[teddy]

The person I know that does it manually uses pace numbers and bias... then whittles the fields down to contenders and key boxes tri and super. He gets the same rebate as the computer groups.

[hugh]

There is a group messing with the Exacta Pool most races at Golden Gate. They hit the pool right away... It's remarkable how many they hit. I just follow along... Pretty much the payments that seem too low...
It hurts every time but money is money.

[hugh]

Race 5 for example.....
The 3 horse is 4-1.... the 6 is 7-1... and Baze's 7 is 9/5 .....
You'd think the 3-7 exacta box would pay less than the 6-7 box right?
Wrong... 3-7 pays 28$ and 6-7 pays 27$ .............
So for me the obvious play is the 6-7 even though the payout is sh*t.
The odds will change of course but noticing this kind of action early when the pools open shows where the big money is coming in...

[Robert Fischer]

you have to think that the guys who do this are using some sort of "predictive" ranking as well to come up with a safe estimate for "fair odds"

as a standalone, pool inneficiencies can be a winner in spots, but combined with fair-odds they are more efficient themselves.

a Beyer system could be used, or if they wanted to get complex they could do the "hong kong thing" and attempt to create their own fair-odds smoothed to the publics odds.

[Robert Fischer]

Race 5 for example.....
The 3 horse is 4-1.... the 6 is 7-1... and Baze's 7 is 9/5 .....
You'd think the 3-7 exacta box would pay less than the 6-7 box right?
Wrong... 3-7 pays 28$ and 6-7 pays 27$ .............
So for me the obvious play is the 6-7 even though the payout is sh*t.
The odds will change of course but noticing this kind of action early when the pools open shows where the big money is coming in...

this is interesting

if I understand correctly, you are talking about Exactas that pay less than they "should" as being tip-offs that those horses are actually BETTER bets than their win pool monies would suggest??

I think that at times that view has some merrit.

What the "computer groups" would do - is actually bet more on the 3-7 EXACTAS(BOX OR NOT PROBABLY MORE OF A RATIO) = so in your scenario (if i get it), the computerGroupers would lose on thoese specific plays.

It's harder for a computer to do what you are talking about, but I agree, that there are times when you see money on a non-favoritehorse and it may just be in an exotic pool, and in fact it is a tip that the horse is live.

[therussmeister]

Race 5 for example.....
The 3 horse is 4-1.... the 6 is 7-1... and Baze's 7 is 9/5 .....
You'd think the 3-7 exacta box would pay less than the 6-7 box right?
Wrong... 3-7 pays 28$ and 6-7 pays 27$ .............
So for me the obvious play is the 6-7 even though the payout is sh*t.
The odds will change of course but noticing this kind of action early when the pools open shows where the big money is coming in...

I don't know why you assume this is "big money". I notice this at all the tracks I watch, but usually at the first flash there is only about $250 in the entire pool, so it's not exactly big money. Smart money, yes. Big money, no.

[point given]

I'm curious on your collective thoughts on whether these computer groups are into the horizontal p3/p4 bets. We all use the DD willpays to get the 2nd leg info, but the othet legs are blind. Can they project into those races and R they doing it . I am a horizontal bettor. Also do they need A level tracks for large pools only and/or scale back for smaller tracks ?

[Spiderman]

This may be all about nothing: I noticed something abnormal in the 7th race exacta pool at Saratoga, today. I usually use Brisnet's Supertote to follow exacta prices, but the reporting change has been slow or delayed and also use the Toteboard at the NYRA website.

Supertote was not showing changes and I looked at Toteboard, made my bet, 17 exacta box. Then, it was 5 minutes after post time, both Supertote and Toteboard display prices until race is official. I noticed that the total exacta pool on NYRA Toteboard was $395K and looked at Brisnet Supertote and noted that it reported $465K exacta pool.

I really do not know if this has anything to do with topic or that there was an inefficiency in reporting the actual pool at NYRA or the $70k difference arrived late. The possible payoff for the 17/71 went from 53/46 to 40/36. The result was 16 paying $48. Before the reported changes, I recall that 16 was paying 53 and do not remember the 61 willpay.

Race 7 had the second highest exacta pool of the day, topped only by the Hopeful. $70k seems to be a fairly high amount to be coming in late.

[pondman]

If you are storing and organizing observations, and utilize a computer to manipulate the data, you are a computer gang...

And have a distinct advantage over people who don't. I don't think it takes a great deal of brilliance to make money at the track. Most 6th graders have enough observational skills.

The majority of people who bet on the races do not record anything. They throw their data away at the end of the day.

I think the theory of bots crawling through the mutual pools, buying and selling their interest in a horse, looking for a few cents is a fabrication.

There are a few people who have access to the last cue, which does give the person an advantage, and that should be an issue.