Sportsbooks Vs. Academics: One Wins The Battle, The Other Wins The War

The post Sportsbooks Vs. Academics: One Wins The Battle, The Other Wins The War appeared first on SportsHandle.

Hope springs eternal — and so do academics who think they might’ve cracked the sports betting code. In fact, two recent papers purport to be able to beat the sportsbooks at their own game.

And while both papers “prove” it’s possible to beat the books (assuming one can understand and master Will Hunting-levels of math), both papers, in the end, run into the same problem faced by many a would-be sharp: The books have a habit of protecting their bottom line by making sure winners aren’t exactly welcome. 

The first paper, via Princeton University’s Robert Axelsen, is MLB Moneylines as Investment Assets. From the abstract: “In this paper, I apply prediction market theory to Major League Baseball (MLB) moneyline pricing. Applying various machine learning models to a comprehensive data set of past game and player data, I calibrate probability estimates of teams’ chances to win games. With these probability estimates, I backtest profitable investment strategies using modified versions of the Kelly criterion staking strategy. Finally, I implement a profitable real-world betting strategy using the techniques developed herein over the first three months of the 2021 MLB season.”

Naturally.

While emails to Axelsen (as well as the authors of the paper to be discussed later) went unreturned, trying to break down exactly what is happening here proved to be … difficult. 

But speaking in general terms, Axelsen’s system was more or less as follows: Bet underdogs.

OK, fine, that’s wildly oversimplifying the 57-page paper that includes phrases like, “To visualize binary cross-entropy, consider (without loss of generality) …” and equations like this:

By page 53, Axelsen starts talking about “average bet characteristics,” and that meant, “{T}he strategy heavily favored underdogs, never betting on a team favored to win. In fact, the stake size weighted average odds for a bet taken was +164 ….”

Of course, Axelsen wasn’t simply mashing buttons on dogs; he was using machine learning and historical models and all manner of mathematics to get to his picks, all the while using a modified form of the Kelly criterion to place wagers.

In the end, Axelsen posted a 131% return on investment, according to the paper. 

But …

Axelsen recognizes the big problem with his method, and it’s not dolts like yours truly who don’t understand math beyond basic algebra. Instead, it’s “the scalablity of such an investment strategy is probably limited; perhaps exclusion/bet size limits from sportsbooks are prohibitively obstructive or perhaps any substantial level of success will eventually lead to market impact; any competent sportsbook is sure to notice a consistently profitable bettor.”

Well, that was fun while it lasted.

Rigged game?

The second paper that’s beating the books is titled, Beating the Bookies With Their Own Numbers — and How the Online Sports Betting Market is Rigged, by Lisandro Kaunitz of the University of Tokyo; Shenjun Zhong of Monash University in Australia; and Javier Kreiner of the Data Science department, CargoX, Sao Paulo, Brazil.

That title of the paper is a whopper, for sure. First up, the “beating the bookies” part — and know when they say “football,” they mean “soccer.”

“We designed a strategy to beat football bookmakers with their own numbers,” the three researchers wrote. “Instead of building a forecasting model to compete with bookmakers predictions, we exploited the probability information implicit in the odds publicly available in the marketplace to find bets with mispriced odds.”

And they succeeded. And they started making money. And then …

“A few weeks after we started trading with actual money, some bookmakers began to severely limit our accounts, forcing us to stop our betting strategy,” they wrote.

And that — the limiting of winners — was the second part of the title, the rigged part.

“Our study sets a precedent of the discriminatory practices against successful bettors in the online sports gambling industry: The online football market is rigged because bookmakers discriminate against successful clients,” they wrote.

Guess you won’t be going back to school

In the end, none of this is surprising to Capt. Jack Andrews, a seasoned veteran of the sportsbook world.

“Yes, I’m sure both of those papers are profitable approaches. There’s probably thousands of papers on how to beat sports betting and I’d guess most of them work,” Andrews said. “You just won’t make enough money to recoup the time you spend reading overly complex explanations of simple concepts.”

Andrews also notes what everyone in the industry pretty much knows: Limits are real, and unspectacular.

“Sportsbooks you can beat won’t take your action. Sportsbooks you can’t beat will,” he said.  “Anyone with a pulse can see how to win at sports betting — pick off promos, use sharp books to identify where square books have bad pricing, etc. — and that’s the science of sports betting. However, getting the money down is where the art of sports betting comes in. The academics never get their heads out of their collective classes to see how it all really works.”

One academic with his head firmly out of his class is John Holden of Oklahoma State University, who has authored or co-authored more than three dozen papers on the intersection of sports and academics, with many having to do with sports betting. And Holden, who’s read the Kaunitz (et al.) paper, is on the same page as Andrews.

“The books simply aren’t likely to let someone beat them for millions and millions of dollars,” Holden said. “We have certainly heard lots of examples of books limiting players who look like they know what they are doing. I would expect any player with a viable model to get the reins put on them pretty quickly.”

So, yeah: The hopes and dreams of academics modeling sportsbook-beating systems seem destined to remain hopes and dreams. And while that’s bad enough, Holden came up with a bit of a doomsday scenario for the nickel-and-dimers among us.

“One other aspect I think to watch is for books to increasingly look to gain the same types of edges — to the extent they don’t already — that these academics are exploring,” he said. “I think if some of these companies ever achieve profitability — a big if — we could see money going into research divisions.”

Oy. R&D at the sportsbooks. The academic-sportsbook complex. Didn’t Eisenhower warn us about this or something?

The post Sportsbooks Vs. Academics: One Wins The Battle, The Other Wins The War appeared first on SportsHandle.

Read More

Related Articles

Responses