12/16/2016 12:34PM

Zhou has become The Terminator


Tony Zhou sees the future. At least that’s one explanation for the success he’s had this year on the National Handicapping Championship Tour. Heading into last weekend, Zhou, a live-bankroll specialist, had accrued points in on-site contests at Belmont, Laurel, Woodbine, and Mohegan Sun. His fifth-place finish last weekend at Monmouth puts him in first place on the tour, just ahead of Cheryl McIntyre, Rick Broth, and Gary Gristick.

There is a small subset of contest players who use computer models to aid in their handicapping contest play. Zhou, 33, who works at a hedge fund in New York City doing quantitative analysis, uses the same skill set he applies at work to help him with his hobby: horse racing.

“Picking horses is the exact same thing as picking stocks,” he said, “the only difference is there are way more shenanigans that go on in finance.”

Zhou has spent the last few years developing a computer model that delivers a true odds line for each race. This allows him to essentially outsource the handicapping part of the equation and spend his time and attention on how best to leverage any market inefficiencies.

Quant methods provide a lot of advantages. “A model is unbiased and it gives you a forced discipline,” he said, “and it also gives you the confidence to bet big on an overlay because you know it’s grounded in math.”

Just like in finance, the fundamental factors are typically built into the price. So Zhou’s model goes deeper, looking at many additional variables – alphas in quant speak – and using them to get an edge.

“Even if an alpha has a half-percent edge, if there are five of them that edge can add up to something significant and it’s an edge that won’t get arbitraged away over time,” he said, noting that his best bets come on horses that look terrible on paper.

“Some people think the public is stupid, but the public isn’t stupid at all: 90 to 95 percent of the time their odds are spot-on, within the margin of error the takeout suggests. It’s the other five or 10 percent of the time where you need to press your edge.”

That ability to press his edge is the reason Zhou thrives in live-bank play. “In an online tournament you’re playing every race and you’re betting the same thing every time,” he explained. “In live bankroll I get to pick the race and I get to pick the pool where I have the largest edge.”

Zhou isn’t just a gambler, he’s also a fan. He has been captivated by racing since the 2010 Belmont Stakes. In 2012, he got to see Union Rags, his favorite horse of all time, win New York’s storied race. He loved the excitement of the big racing days and the intellectual challenge of finding winners and competing against the crowd. It was only a matter of time before he discovered tournament play.

“I heard about contests a few years ago from the show 'Horseplayers,' ” he said, “and that convinced me to play in contests. I played in the NHC for the first time last year and I got the bug.”

His original goal for 2016 was to win two seats to the 2017 event. “I didn’t really know about the tour,” he said. “After I got my second seat a buddy told me he thought I had a shot at it and since then I’ve been playing in as many tourneys as I can.”

Zhou, despite being as new school as they come, has tremendous respect for old-school horseplayers. “Old-school players are still around for a reason,” he said. “The best ones have a secret sauce.”

He pointed out that the nature of all markets is to become more efficient over time. Things that provided an edge years ago might not even get you to even today. “The question shouldn’t be, ‘Are you old-school or new-school,' ” he said, “the question should be ‘Are you willing to put in the hard work and adapt?' ”

Garry Kasparov, one of the greatest chess players of all time, has said that while the best computer can now beat the best human at chess, the combination of a very good human and a very good computer can still beat the best computer. This is the future that Tony Zhou sees. “The future is not going to be about man versus machine,” he stated. “It’s going to be about man plus machine.”