Que
03-17-2001, 02:15 PM
If you haven't had a chance to read Andrew Beyer's column on Computerized Betting (http://www.washingtonpost.com/wp-dyn/sports/columns/beyerandrew/A59744-2001Mar12.html) you should. The article is about the North Dakota off-track betting operation that allowed a syndicate of gamblers to make their bets electronically directly through AmTote. I am outraged that the parimutual industry allowed this to happen, and that the president at Racing Services, where these bets were made believes there's nothing wrong with this practice. Let me explain how wrong this practice really is.
To begin, to rationalize that these bettors are just better handicappers than the rest us, and the large amounts of money they wagered justified this privilege is just pure bs. I believe this syndicate didn't look at a single past performance, and that the majority of their bets where in the place, show, exacta, and according to Beyer, the trifecta pools. I can assure you, it's relatively easy to write a program that calculates the probable percentage that each horse should finish 1st, 2nd, or 3rd. Once you know these percentages then its relatively easy to calculate the probabilities for all exacta and trifecta combinations. All you need to know then is the amount of money in each pool, to calculate the payoffs. Multiply the expected payoffs with the expected percentages and you can determine if the combination is over or undervalued.
I know that its relatively easy to create a program that does this is because I've done it. Several years ago I wrote a program that automatically downloads the Bris Supertote data, parses the information, then calculates the fair value for each place, show, and exacta combination. (Note. I couldn't calculate trifecta combinations because apparently only the betting syndicate in North Dakota is given access to trifecta pools.) My program calculated win percentages using methods described in Dr. Z's, "Beat the Racetrack," and place and show percentages using Asch & Quandt's, "Racetrack Betting." I then used small adjustments based on odds to account for the fact that very low priced horses are often underbet while very high priced are often overbet. I also made small adjustments to horse's place and show probabilities based on an analysis of the odds of practically every race ran in North America since 1998. The only other "handicapping" variable I used was post position to adjust the percentages, especially the place and show percentages. This method never looked at a horse's past performances, and I'll bet neither does the North Dakota's syndicate's program.
I tried using this program with minimal success, only giving up because of all the late price fluctuations. For example, I became tired of betting 30% overlays 1-2 minutes prior to post--only to find that the wager was a 5% underlay after the race. The reason this method didn't work, wasn't because of the methodology, but because of slow execution. Just see how fast you can make 167 bets using YouBet, Philly PhoneBet, or Ladbroke's Call-A-Bet. However, this betting syndicate in North Dakota with direct access to AmTote placed 167 wagers, for a total of $25,569, in a total of three seconds... and this was just seconds before post time. Not to mention the fact that the syndicate was reportedly given access to trifecta pools--something you or I don't have access too. The fact that these bets were made in North Dakota should make you even more alarmed.
Let's say that I was given the same access to AmTote that this syndicate was--here's how I would change my program. As far as the selection methodology discussed above--nothing would change. However, with regards to betting methodology, this is what I would do. Given the ability to get up-to-the-second parimutual data seconds before post, I would program my computer to make place, show, exacta, and trifecta wagers on any combination that was a 10% overlay or more, while adding all my wagers into the tote data for the program's calculations. The program would then optimize the size of the bet to make it pay exactly fair value--you read that correct, bet enough money to pay exactly what it should pay--less track take, breakage, and IRS fees.
Now I'm sure you're asking why bet everything down to it's fair value, and not simply wait for overlays. The answer is because I'm betting in North Dakota, and I'm assuming that I'm getting a healthy rebate. Therefore by increasing the total value of my wagers by betting anything just above fair value, I can be assured of earning a healthy 8-10 percent on my investment--mostly through rebates. The added advantage of this strategy, is that it prevents all the other horseplayers (i.e. my competitors) from using this same strategy, since I'm taking their overlays and betting them down to fair value. Specifically, I'm making my profits from the rebates and anything left from betting the overlays down to fair value is additional money in my pocket.
The bottomline is that these people are taking money from your pockets. Not to mention the fact, that these people don't need to know anything about horseracing or handicapping in order to do it. And the real scary thing is that the majority of the racing industry doesn't think there's anything wrong with this practice. Unless this practice is stopped, then the value you get on your bets (with the exception of your win bets) will be greatly reduced.... i.e. the money these people take is money right out of your own pocket... and they are using your handicapping skills (used to make their computations) to do it. And if the racing industry allows this to continue, I'm either going to pack my bags and move to North Dakota or I'm going to give up this game altogether.
Que.
To begin, to rationalize that these bettors are just better handicappers than the rest us, and the large amounts of money they wagered justified this privilege is just pure bs. I believe this syndicate didn't look at a single past performance, and that the majority of their bets where in the place, show, exacta, and according to Beyer, the trifecta pools. I can assure you, it's relatively easy to write a program that calculates the probable percentage that each horse should finish 1st, 2nd, or 3rd. Once you know these percentages then its relatively easy to calculate the probabilities for all exacta and trifecta combinations. All you need to know then is the amount of money in each pool, to calculate the payoffs. Multiply the expected payoffs with the expected percentages and you can determine if the combination is over or undervalued.
I know that its relatively easy to create a program that does this is because I've done it. Several years ago I wrote a program that automatically downloads the Bris Supertote data, parses the information, then calculates the fair value for each place, show, and exacta combination. (Note. I couldn't calculate trifecta combinations because apparently only the betting syndicate in North Dakota is given access to trifecta pools.) My program calculated win percentages using methods described in Dr. Z's, "Beat the Racetrack," and place and show percentages using Asch & Quandt's, "Racetrack Betting." I then used small adjustments based on odds to account for the fact that very low priced horses are often underbet while very high priced are often overbet. I also made small adjustments to horse's place and show probabilities based on an analysis of the odds of practically every race ran in North America since 1998. The only other "handicapping" variable I used was post position to adjust the percentages, especially the place and show percentages. This method never looked at a horse's past performances, and I'll bet neither does the North Dakota's syndicate's program.
I tried using this program with minimal success, only giving up because of all the late price fluctuations. For example, I became tired of betting 30% overlays 1-2 minutes prior to post--only to find that the wager was a 5% underlay after the race. The reason this method didn't work, wasn't because of the methodology, but because of slow execution. Just see how fast you can make 167 bets using YouBet, Philly PhoneBet, or Ladbroke's Call-A-Bet. However, this betting syndicate in North Dakota with direct access to AmTote placed 167 wagers, for a total of $25,569, in a total of three seconds... and this was just seconds before post time. Not to mention the fact that the syndicate was reportedly given access to trifecta pools--something you or I don't have access too. The fact that these bets were made in North Dakota should make you even more alarmed.
Let's say that I was given the same access to AmTote that this syndicate was--here's how I would change my program. As far as the selection methodology discussed above--nothing would change. However, with regards to betting methodology, this is what I would do. Given the ability to get up-to-the-second parimutual data seconds before post, I would program my computer to make place, show, exacta, and trifecta wagers on any combination that was a 10% overlay or more, while adding all my wagers into the tote data for the program's calculations. The program would then optimize the size of the bet to make it pay exactly fair value--you read that correct, bet enough money to pay exactly what it should pay--less track take, breakage, and IRS fees.
Now I'm sure you're asking why bet everything down to it's fair value, and not simply wait for overlays. The answer is because I'm betting in North Dakota, and I'm assuming that I'm getting a healthy rebate. Therefore by increasing the total value of my wagers by betting anything just above fair value, I can be assured of earning a healthy 8-10 percent on my investment--mostly through rebates. The added advantage of this strategy, is that it prevents all the other horseplayers (i.e. my competitors) from using this same strategy, since I'm taking their overlays and betting them down to fair value. Specifically, I'm making my profits from the rebates and anything left from betting the overlays down to fair value is additional money in my pocket.
The bottomline is that these people are taking money from your pockets. Not to mention the fact, that these people don't need to know anything about horseracing or handicapping in order to do it. And the real scary thing is that the majority of the racing industry doesn't think there's anything wrong with this practice. Unless this practice is stopped, then the value you get on your bets (with the exception of your win bets) will be greatly reduced.... i.e. the money these people take is money right out of your own pocket... and they are using your handicapping skills (used to make their computations) to do it. And if the racing industry allows this to continue, I'm either going to pack my bags and move to North Dakota or I'm going to give up this game altogether.
Que.