Horse Racing: Data Driven models?
Okay, let us say these teams do exist and do prosper. The "bespoke software" (proprietary software) works due in large part to the value of the algorithm and the nature and quantity of the reams and reams of data that is processed.
So what data is there that some millionaire genius in Hong Kong has available but the average punter either there or elsewhere does not have available to him.
Age of a horse? Is precise to the day data more valuable than Two Years Old?
Track Conditions: The Punter hears "Muddy"... does the database guru get a soil hydrology report from the Agriculture Department? Or an actual rainfall over time calculation?
Humidity: Do horses "feel rain in their bones" the way some older people claim to? If so, perhaps the data gurus have humidity levels rather than just Rainy such as a Punter might have available to him?
Pole position history: Heck there aint that many pole positions.
Time it took the horse to cross the line in its non-winning races versus Punters who only have data for and about Winners in prior races?
Sunshine? Jockey's Colors? Cloud cover? Wind speed?
What on earth is it that a data guru would have that an ordinary Punter would not have?
Quote: FleaStiff
What on earth is it that a data guru would have that an ordinary Punter would not have?
Nothing, and these "equations" which take 100s of factors into account have been around since the first computer. As a horseplayer, nothing beats old school capping and seeking out large carryovers*.
*This can be misleading, as there are plenty of tiny tracks that have carryovers in the 2000 range, but since sometimes there's only 1000 bet in every day, it provides a solid return. I've done well on seemingly small carryovers on Pick 5s and Place Pick Alls at places like Freehold, Hawthorne, Los Al, etc.
Past performance data tells you what the horse did and not what it's going to do........handicapping can't be put into a box. Past performance data provides a platform for the "great argument" which is the essence horse race racing.....Successful handicapping requires a certain "feel" that can only be had through years of experience. Yes, there are successful horse players even with takeouts averaging 22-25% across all pools. Rebate shops help but it always about picking winners which machines can't do.....Quote: cclub79Nothing, and these "equations" which take 100s of factors into account have been around since the first computer. As a horseplayer, nothing beats old school capping and seeking out large carryovers*.
*This can be misleading, as there are plenty of tiny tracks that have carryovers in the 2000 range, but since sometimes there's only 1000 bet in every day, it provides a solid return. I've done well on seemingly small carryovers on Pick 5s and Place Pick Alls at places like Freehold, Hawthorne, Los Al, etc.
an algo that could identify "juiced" horses is an interesting idea. Owners/trainers can be very patient when setting up a horse. A good handicapper can identify certain owner/trainer/jockey patterns that signal a horse might be ready. It's conceivable, maybe, that these patterns could be plugged into algo's.....? The algo may have a hard time separating which horse is being juiced the most as most horses in a race are being juiced to one degree or another.Quote: dwheatleyThere was an episode of Numbers where one of these math geniuses with a horse picking algorithm was murdered. The study of his algorithm revealed he had found a way to identify which horses were being juiced, and thus which trainers & stables were probably juicing other horses. Then he would bet on those horses.
Quote: treetopbuddyPast performance data tells you what the horse did and not what it's going to do........handicapping can't be put into a box. Past performance data provides a platform for the "great argument" which is the essence horse race racing.....Successful handicapping requires a certain "feel" that can only be had through years of experience. Yes, there are successful horse players even with takeouts averaging 22-25% across all pools. Rebate shops help but it always about picking winners which machines can't do.....
I think horse handicapping, unlike sports betting, is something that you get better at over time.
And the reason is that the more races you see, the more that you see the favorite doesn't win. And once you realize that, then you can be successful.
I don't think algorithms can predict who will win. But I do think they can say that a horse that is going off at 8-1, should really be going off at 4-1. And if you can do that enough, and look at enough exacta / dd pools to know what you're betting, I do think the takeout can be beaten.
And carryovers help as well.
Don't they test for "juice" in the horse's urine at the end of the race? Though juicing the losers might work better perhaps? Juice the favorites to lose and bet on some longshot nag.Quote: dwheatleyhe had found a way to identify which horses were being juiced.
NOTE: This link is what prompted the thread about bespoke software.
you will go broke trying to beat favorites and on the flip side you will go broke betting on them.....favorites are winning at a higher percentage as a lot of hunch players have left the game for slots, etc. Odds-On favorites win at a 55% clip, still a negative ROI but must be respected when placing a bet.Quote: FinsRuleI think horse handicapping, unlike sports betting, is something that you get better at over time.
And the reason is that the more races you see, the more that you see the favorite doesn't win. And once you realize that, then you can be successful.
I don't think algorithms can predict who will win. But I do think they can say that a horse that is going off at 8-1, should really be going off at 4-1. And if you can do that enough, and look at enough exacta / dd pools to know what you're betting, I do think the takeout can be beaten.
And carryovers help as well.
Interesting take on "juicing" horses FleaStiff. It was said that Al Capone would have racing greyhounds pads sanded before race. Pre-steroid days. All winning horses go to the "test barn" as well as random horses or horses that are viewed as running too good or poorly as deemed by stewards. The motivation to win far outweighs the motivation to catch cheaters.....a la Lance Armstrong. EPO was and to a lesser degree still used. Masking agents are deployed to throw off standard testing. Cheating is part of the horse racing culture.Quote: FleaStiffDon't they test for "juice" in the horse's urine at the end of the race? Though juicing the losers might work better perhaps? Juice the favorites to lose and bet on some longshot nag.
NOTE: This link is what prompted the thread about bespoke software.
Read the link. I'm a little skeptical. The numbers thrown out don't seem right. Would have to look at handles at these tracks. Many of the computer guys have gone broke. Having been a horse player for 30 years, I've observed that the Oriental players don't read the Race Form or any other PP's for that matter. "Charlie Don't Surf". They are "board readers". Whether this plays as an advantage to the computer guys is unclear to me. Having 1200 horses on the grounds for the meet is a very quantifiable thus a disadvantage to the machines? None or very few shippers to the track for the meet. Thinking out loud here. It's possible to exploit the other players with leverage. The average horse player is not real affluent. In most races there are several contenders. Average Joe would love to bet that extra horse or two but doesn't have the money. Maybe an advantage is gained there. I knew a couple of lawyers that were doing well with "lug up horses" in the superfecta. They would single out a horse or two that they believed would be in the mix but probably would not win the race. Put the horse 2-3-4 position. They would put 500-600 in the race and hit a lot of "signers". Getting "value" on horses is only as good as the handicapping methods. One guys "value" is another guys underlay. A player/syndicate can power their way to wins.Quote: FleaStiffWell, absent things like sandpaper, steroids and Dutching, what data is available for such data-driven-models. I can't see how these Hong Kong betting syndicates are supposedly bring in millions of dollars based on the amount of jasmine wafting in the air during the race. What more is know by these bespoke software programs than is known by the average punter who buys a copy of the racing form or morning newspaper?
Orientals are "board watchers" not readers...my bad.Quote: treetopbuddyRead the link. I'm a little skeptical. The numbers thrown out don't seem right. Would have to look at handles at these tracks. Many of the computer guys have gone broke. Having been a horse player for 30 years, I've observed that the Oriental players don't read the Race Form or any other PP's for that matter. "Charlie Don't Surf". They are "board readers". Whether this plays as an advantage to the computer guys is unclear to me. Having 1200 horses on the grounds for the meet is a very quantifiable thus a disadvantage to the machines? None or very few shippers to the track for the meet. Thinking out loud here. It's possible to exploit the other players with leverage. The average horse player is not real affluent. In most races there are several contenders. Average Joe would love to bet that extra horse or two but doesn't have the money. Maybe an advantage is gained there. I knew a couple of lawyers that were doing well with "lug up horses" in the superfecta. They would single out a horse or two that they believed would be in the mix but probably would not win the race. Put the horse 2-3-4 position. They would put 500-600 in the race and hit a lot of "signers". Getting "value" on horses is only as good as the handicapping methods. One guys "value" is another guys underlay. A player/syndicate can power their way to wins.Quote: FleaStiffWell, absent things like sandpaper, steroids and Dutching, what data is available for such data-driven-models. I can't see how these Hong Kong betting syndicates are supposedly bring in millions of dollars based on the amount of jasmine wafting in the air during the race. What more is know by these bespoke software programs than is known by the average punter who buys a copy of the racing form or morning newspaper?
Quote: teliotI recently made a post about advantage horse racing on my blog.
It seems becoming an advantage horse bettor might be expensive: "... a program containing a 13-year database on horses from the Jockey Club, is at the disposal of any gambler willing to pay $425 up front, plus a $425 annual fee. "
Atleast this article gave some indication of what the "data" is that drives these programs that identify suitable horses though it seems the margins are narrow and other program-models may converge on the same opportunities thus driving the odds into unprofitable territory.
Your right FleaStiff....I'm guessing most programs land on the same horses. The pools are so thin that any significant betting becomes problematic. I'll take the hardened 40 year horse player over a machine any day.Quote: FleaStiffQuote: teliotI recently made a post about advantage horse racing on my blog.
It seems becoming an advantage horse bettor might be expensive: "... a program containing a 13-year database on horses from the Jockey Club, is at the disposal of any gambler willing to pay $425 up front, plus a $425 annual fee. "
Atleast this article gave some indication of what the "data" is that drives these programs that identify suitable horses though it seems the margins are narrow and other program-models may converge on the same opportunities thus driving the odds into unprofitable territory.
Quote: KeyserYou should stick to casino games. Even on slots, you'll lose less.
It depends. If you are at a track and only bet the races at that track, making one bet every half hour is great for bankroll preservation as opposed to a spin of the slot machine every 3 seconds. Now if you are at a book or at the track and you are betting races at Aqueduct, Calder, Arlington Park, Churchill Downs, Hollywood Park and other tracks, then it could add up very quickly.
The edge in the Parimutuel system is hard to overcome however. I am not sure what it is, but I recall hearing something like 15% or thereabouts. I think someone who is truly in the know and an expert on the issue might be able to overcome them. The only think I have seemed to learn about betting on horses is to avoid ones with cutesy names like Sunday Kisses or something like that because inevitably a bunch of wives will ask their husband to put money on that horse and horses with names like that just never seem to have good value.
Quote: Gabes22Going to the track and playing the ponies is a fun day. I live about 10 minutes from Arlington Park and get down there like 2-3 times a year and make a day of it with my wife.
We are neighbors.
I'm at AP once a month for a full day of races.
I'll also stop by after work for the late double occasionally if it's a nice day.
exercising in front of me.
Now I am remembering when the Arlington Million was first run, 31 + years ago.
John Henry, ridden by the Shoe. SIGH
Quote: BuzzardThanks for nothing, Finn. I was feeling young today in weight training class with all those little hotties, pony tails and all,
exercising in front of me.
Now I am remembering when the Arlington Million was first run, 31 + years ago.
John Henry, ridden by the Shoe. SIGH
What a race! I wasn't alive for it... Well, depends on when you consider life starting...
Was it John Henry, or was it The Bart? No one really knows....
Speaking of being older than dirt, it was 53 years ago I actually had a horse change my bet. I had just walked to the rail when Bald Eagle came by in the post parade. One look at him and I went back inside and bet all I had on him to win.
He won the International that year and the next as well.
