Horse racing data: clarity for bettors and bookmakers
A London firm best known inside bookmaker circles for pricing and risk management is pushing further into horse racing analytics, using video-based biomechanics to advise on racehorse purchases while arguing that one of racing’s biggest challenges is making its data less intimidating for new bettors.
In an interview with SiGMA News, Stephen Davison, head of commercial operations at Pythia Sports, described a model that runs predictive and pricing systems across racing jurisdictions. “We price every horse race around the world, whether it’s the UK, Ireland, USA, Australia, or South Africa.”
Racing has always been a data sport, even when it didn’t use that language: punters, trainers and owners have long pored over form lines, pedigrees, times and track conditions in search of an edge. Davison’s point was not that the industry lacks information, but that it can be difficult to translate into decisions, and, for newer audiences, difficult to digest. “There’s a misconception that there isn’t much use of data, but there’s actually a huge amount of it. It’s more about using it correctly,” he said.
Davison’s comments underline the delicate balance the industry faces: sportsbooks want engagement, customers want trust, and the modern racing product is built on data most fans never see.
“It’s not an exact science”: modelling the breeze-ups
Davison suggested that, compared with football’s analytics culture, racing is behind the curve. “It’s been prevalent in football for a long time now,” he said. “Pretty much every single football team will be using some form of data analysis.” In racing, he argues, the ingredients are present (time, stride, movement, pedigree), but the industry has been slower to standardise how they are captured and interpreted.
At the sales ring, he said, the evidence is there to be gathered: how fast a horse runs in a breeze (the timed workout or gallop used to assess a horse’s fitness), how it moves, and how those patterns relate to pedigree and later performance. The company’s approach centres on the breeze-up sales, where young horses are timed in a short workout for prospective buyers.
Pythia’s model, he explained, is built on three strands. “One section of our model looks at time, the time the horse runs,” he said, before moving to gait metrics: “We can also see the stride of the horse: how many times it turns over its legs, the frequency of the stride, and how big the stride is.” The third layer uses video analysis away from the track, in what Davison called a “biomechanical walk model”. “That’s back at the stables, when the horse is walking, just trotting around. We overlay a model, with various points on the horse’s body, whether it’s the fetlock, the knee, the head, whatever it might be.”
The aim, he insisted, is not to promise certainty but to narrow the field. “We then back-test that against previous walks,” Davison said, “to see: this particular horse walked in a certain way from this pedigree and it was a highly rated horse, so this one looks similar.” He added: “It’s not an exact science where we say, ‘this will be a great horse, this will not’. It’s about reducing the noise and helping people when they’re purchasing horses.”
Overwhelming data?
If bloodstock is one frontier, the other is market expansion for Pythia’s managed trading service. Stephen Davison described a product called “Racing One Markets”, sold business-to-business to bookmakers and platforms.
Behind the expansion sits an old tension in bookmaker services: firms like Pythia promise improved pricing and risk management, but racing’s long-term problem is demand. Davison’s answer was a stronger emphasis on making the product easier to use, particularly for younger customers. “We’re looking at how we can make it easier for the consumer to bet,” he said, mentioning “recommended bet prompts”, “a carousel at the top of a bookmaker site”, and “a horse racing bet builder”.
But he also acknowledged the sport’s steep learning curve. “You’ve always got the jockey name, the trainer, how many runs it’s had, where it’s finished, the distance it’s run over, even the weight of the jockey,” Davison said. “There’s quite a lot of information for consumers to use.”
In his view, racing needs to “broaden its appeal… internationally, but also for the next generation of bettors”, partly by “lowering the barrier to entry”. He offered a blunt example of how alien the sport can sound to newcomers: “No one outside of horse racing regular bettors really talks in pounds and furlongs, right?”
For those betting on horse racing, the difficulty is compounded by uncertainty. “The challenge has always been that finding a winning selection isn’t simple, and the industry needs to find smarter ways to lower that barrier to entry for everyday bettors,” he said.
That is where, he said, simplified data could help customers feel more confident without being sold “tips”. “We’re not necessarily selling tips,” Davison said. “We’re just trying to give people a more informed view of the options available to them to bet on.”
Yet, even as he talked about broadening appeal, Davison was clear about the commercial reality on the bookmaker side. “We manage the risk for a lot of these clients, and we need to achieve an acceptable margin each month,” he said. Some operators, he added, effectively outsource the decisions: “You’re the experts, you tell us what’s best to do to maintain a margin.” Others take a more hands-on line, discussing staking and exposure with the supplier. “It becomes a real partnership,” he said. “Both sides have skin in the game, and working toward the same outcome is where you tend to get the best results for everyone involved.”
Africa’s iGaming force is rising, and Cape Town’s shaping the continent’s future. From 03–05 March 2026, SiGMA Africa will bring together 3,000 delegates, 780 operators, and 150+ speakers to the table.
