12/08/2015 3:55PM

Symposium: Sophisticated number-crunching may be future of handicapping


TUCSON, Ariz. – The era of Big Data is likely to usher in the widespread availability of sophisticated handicapping and betting programs that closely resemble the type of high-dollar systems that have operated in the shadows of the racing industry for the past 15 years, according to officials speaking at the University of Arizona Symposium on Racing and Gaming on Tuesday morning.

Whether the programs will be readily embraced by handicappers accustomed to less data-intensive but far more personal methods is entirely unclear. But presentations at the Symposium during a session devoted to Big Data – the term that has been given to current-day efforts to exploit the glut of information being churned out by applications connected to the internet – indicated that the day is soon coming when horseplayers of all budgets and experience will have access to the programs, for better or for worse.

Already, one company, predicteform.com, has developed a program using proprietary algorithms that analyze past-performance information over a range of races, applies that analysis to real-time odds feeds, and then spits out recommended bets based on those analyses. The lead investor in the company, Tom Grossman, appeared during the panel to describe the program, which is currently being marketed commercially.

For 15 years, investor-owned computerized robotic wagering programs, or CRWs, have performed the same dual analyses, but those programs are provided with a direct interface into the bet-processing system so that the programs can see every detail of the betting pools in real time and send thousands of wagers into the pools in a matter of seconds. Those systems are also operated exclusively at off-track betting outlets that reward the operators with rich rebates on their handle, allowing the owners of the systems to post profits at a lower return on investment than the typical bettor in the parimutuel pool.

It is not clear if the new commercially available systems will be as valuable to individual players if they are not also provided with a direct link into the pools and generous rebates. However, the new systems could provide competition to the existing computerized wagering systems, a dynamic that could upset the market in much the same way that the first widespread publication of speed figures decades ago led to a reduction in the advantage held by players who produced their own figures, as pointed out during the session by David Siegel, the president of a handicapping-products division of Equibase.

“We saw that as those products suddenly got into the hands of everyday consumers,” Siegel said.

Much of the discussion of computerized-robotic wagering programs occurred after the session’s formal presentations, largely spurred on by comments made by Todd Bowker, the general manager of Premier Turf Club, an account-wagering company. Bowker defended the use of such programs, which have been the topic of debate in the racing industry for as long as they have existed, even if the most fervent debates have taken place in small corners of the industry.

However, the debate has widened over the past year as more and more racing constituencies have embraced calls for the sport to dramatically expand the amount of free data available to racing fans, including detailed betting data. Supporters of the concept have said that all other sports provide enormous amounts of data free of charge and that racing should do the same in order to keep pace in an increasingly data-centric world. Those calls have dovetailed with the debate over CRWs because of the robotic programs’ seemingly favored access to parimutuel wagering pools and the benefits provided by the real-time look.

Equibase, the racing industry’s official data collector and supplier, has made a sizeable amount of its in-house data available online in the past five years, but it has kept in place restrictions that prevent people from scraping the database for patterns, as some sports allow. At the same time, earlier this year the company forged a partnership with a sports-data marketing company, STATS, that develops proprietary predictive data tools for other sports, largely for the benefit of fantasy-sports players.

Jim Corelis, the senior vice president of STATS, said during his presentation at the symposium that the company “is setting [the algorithm] guys loose” on the Equibase database in an attempt to come up with new, simple predictive tools the company can provide to new and existing racing fans. He also said the company is working on a “dashboard” that can be used by customers to develop their own data-mining tools.

The effort by STATS and the effort by Grossman’s company share a goal. Grossman acknowledged that his program “takes the work out of handicapping,” while Siegel of Equibase said that much of the effort to create simple representations of complex interplays of handicapping factors would work best for new visitors to the racetrack.

"I think that new fan wants to be told what to do, not be bothered with the particular data analysis," Seigel said.

That leads to a larger question as to whether the sport can attract and retain fans if the vast majority of the handicapping is done by computerized entities. And, it should be noted that after the advent of publicized speeds figures, data providers started publishing more data, not less.

Meanwhile, the sport’s most ardent fans usually cite the richness of the sport’s handicapping process as one of its most attractive features. In the future, the question might be: "So who’s yer robot like?"

Correction: An earlier version of this article erroneously linked the account-wagering company Premier Turf Club with computerized robotic wagering. Premier Turf Club does not have any CRW clients.

Ray Sousa More than 1 year ago
At the end of the day if the integrity issues are not addressed all handicapping will be for naught. aqueduct this Thursday is a good example no amount of handicapping would come up with the winner of the 1st and 4th races .it was textbook how to fix. i don't know about the rest of the card i gave up after those results and went home. im not a big fan of systems i dont believe they work in racing.i think the whole algorithm thing does work for sports and definitely the stock market but not horse racing. handicapping horses is an art and a science. science alone just does not work not for big scores anyway. you might irk out a pennies on the dollar roi over a period of time with a lot of discipline but what fun is that. without the rebates that the offshore sites give you i dont believe you can stay afloat. But who knows. i believe what good handicappers need is a fair shot and that comes only when there is no hanky panky.or at least not to much.
Otis Boersma More than 1 year ago
Love the robot stuff. It's fun to play with, can be highly predictive and with broader access will level the playing field data-wise. However, and with some level of irony, I think the ultimate edge for handicappers in the robot age will eventually reside in paddock and post -parade assessment. In other words, things not easily captured in the numbers. Like a how a horse looks and feels that day.
Jack Armstead More than 1 year ago
This will not take over racing in our lifetimes. There are "whales" who are catered to by these "CRW's" who manage to dump hundreds of wagers using their OC45 bandwidth, but; these "whales" lose big too. There will never be a substitute for evaluating how the horse looks and feels that day, no matter how much Data or "Beyers" or "Rag's" are used. You could give me the ability to data mine Equibase's data for the past 25 years and I may come up with different Trainer insights with different owners of different breeds with different equipment.... blah, blah, blah... but I wouldn't bet it unless I had to (because of a Multi-race Wager).
John Rothschild More than 1 year ago
As a horse owner, I completely agree. We know things about our stock, medical procedures and trainer intentions that can't be quantified on a stats sheet. John Rothschild