Data Sharing, Analytics Talk Remains Prevalent at HBPA Convention


Steven Vickner | Denis Blake/National HBPA


CLEARWATER BEACH, Fla.–The theme of improving horse racing through greater sharing of data and revved-up analytics continued into Thursday’s second day of programming at the National Horsemen’s Benevolent & Protective Association’s annual convention at the Sheraton Sand Key Resort.

Dr. Steven Vickner, an associate professor in the University of Louisville College of Business’s equine industry program, discussed research projects designed to let the horse industry make smarter business decisions. Vickner portrayed the beneficial ways to “tease out” factors involved in when, why, how and on what people bet as limited only by the access to data. The presentation dovetailed nicely with the Thoroughbred Idea Foundation’s new white paper entitled “Embracing a Future with Free Racing Data,” a subject of Wednesday’s keynote panel that advocated providing raw data files free to non-commercial entities as well as a form of basic free past performances as ways to encourage new horseplayers and to spur innovation.

The white paper also recommended partnering “with universities to study racing data, developing new and advanced metrics for the betterment of the sport.” That’s music to the ears of Vickner, who said U of L’s goal is to regain its position as the world’s academic thought leader on the equine industry.

Vickner discussed a study model he constructed with his former University of Kentucky student Steve Koch (now the executive director of the Safety & Integrity Alliance of the National Thoroughbred Racing Association) that studied 1,515 thoroughbred races over 165 days in 2011 at Woodbine Racetrack in Toronto to identify factors influencing handle. Perhaps surprising no one, the leading determinant of per-race handle was field size, albeit with diminishing marginal returns for each additional horse. Restricted race conditions, such as for Ontario-sired horses, had a negative impact on handle, though larger average field sizes in those races offset the effect. The study also looked at the impact of weather, race distance, surface, off-the-turf races and day of the week and month.

The Woodbine study also is a textbook on why such research is not more extensive. Missing data includes who the horses, jockeys and trainers are in each race. Vickner said it’s not simply a matter of looking up the race charts on Equibase for data. In fact, the Woodbine study was possible because Koch was working there at the time, he said.

“Steve had to work with his IT department to get 1,515 rows in that spreadsheet, and literally there are 100 columns,” Vickner said in a conversation after his presentation. “When I say data, it’s per-race handle for the win pool, place pool, show pool, Pick 3, Pick 4, every pool. Plus on track vs. simulcasting, plus the weather at each point. So imagine for each and every race you’re collecting that level of detail of data. That’s hard to get an IT department to create those kinds of special requests.Then what’s the temperature, wind and precipitation for each race?”

The Woodbine study is only the tip of the iceberg as to what can be achieved, Vickner said. For instance, while takeout — the amount taken off the top of every dollar wagered — was known in the Woodbine study, it takes a multi-year project to assess impact.

Impediments to getting data not only is the massive time and ensuing expense involved but some tracks are reluctant because of concern of revealing proprietary secrets or that something might “shine a negative light on a company,” he said, adding that tracks are more likely to feel comfortable sharing old data.

Vickner said that fear should be assuaged once multiple tracks become part of studies, where conclusions can be broad-based and individual operations can’t be pin-pointed.

He said his research team will be making requests of Equibase, The Jockey Club Information Systems and racetracks for data but wants to do it for upcoming racing rather than looking at a past year, to make the conclusions as relevant as possible. One such project is with Monmouth Park for the New Jersey track’s 2019 racing season that will include data from the Monmouth’s new sports book.

“Now you will have a wonderful set of data that everybody is interested in,” Vickner said. “Then you go to Equibase and say, ‘Here’s my ask,’ rather than expending that on 2011 data. Because that’s a big ask.”

Another important variable: Who is betting how and when, information that can reliably only come from online advance-deposit wagering platforms. One ADW that is pouncing at the opportunity is Louisville-based AmWager, with founder Nelson Clemmens and simulcasting director Papo Morales (a U of L grad) telling Vickner at the convention that they will open up all their data to a research project that can advance the sport.

Vickner said U of L ultimately will conduct economic-impact studies on every Kentucky racetrack every year for free as part of the mission of working for the commonwealth’s horse industry.

“The idea is then to show this product to each and every racetrack in America and say, ‘Wouldn’t you want one of these, too?'” he said.

Not a subscriber? Click here to sign up for the daily PDF or alerts.