Researchers learn how to beat online betting sites, and promptly find their accounts limited
There is one immutable rule of gambling: the house always wins. The subtext to this rule is that you, on aggregate, will always lose out. Researchers from the University of Tokyo found this out the hard way when they discovered a way to consistently beat the system using entirely legal means, and then found their accounts limited or flagged for manual inspection.
Before they reached this unfortunate end, they had made a profit of $957.50 across 265 football matches – a return of 8.5% overall. Their system ensured bets paid off 47.2% of the time – even better than the 44% win-rate they had simulating 479,440 bets on matches played between 2005 and 2015 based on historic odds.
How did the researchers beat the bookies?
Before we get to the researchers’ winning methods, we have to explain a little about how online bookies ensure they don’t make a loss, no matter what. This kind of insurance is relatively trivial for a game like roulette, where there are 38 possible outcomes, giving each gambler odds of 18/38 if they pick red or black. A payout of $2.111 is fair for a $1 stake, but the house pays $2 and keeps the difference – in other words, it’s always guaranteed a profit.
They try to do a similar thing with football, but it throws up its own problems. Bookies – both online and off – go to great lengths to ensure that their odds are as accurate as possible, employing teams of statisticians and sports experts to ensure they’re not being overly generous either way.
But these sophisticated models betray an underlying weakness that the researchers were able to exploit: bookies do a certain amount of hedging their bets to protect against the risk of a huge payout. As any fellow fan of Derby County will understand, sometimes we bet more out of hope than expectation. If people do this in a large enough volume, bookies will worry about the risk of a large payout, and compensate by making the opposite bet seem more appealing with more generous odds. That way, they make some of their money back and buffer against a big loss.
You’re unlikely to spot this with a single bookie, but by comparing odds across bookmakers, the researchers discovered they could easily spot outliers where the odds favoured the punter rather than the house, creating an unusual opportunity.
So, the researchers began with the assumption that the odds given by the bookies on aggregate are a fair reflection of the likely outcome: win, lose or draw. To that end, they developed some code that would crawl the bookies collecting the odds on football matches around the world. They calculated the average odds on each game, spotted the outliers and bet with the overall wisdom of the bookie-crowd.
If the bookies’ odds were right, they should make a killing.
How did they do?
Speak to any scientist, and you’ll know they’re not exactly swimming in money, so they began by running their system on historical data. Namely the odds and outcomes of 479,440 football matches played between 2005 and 2015. They won on 44% of the bets, giving them a yield of 3.5% from 56,435 imaginary $50 bets – a profit of $98,435.
But did they just get lucky? To check this, the team put on 2,000 random bets in another simulation. This time, despite winning 39% of the time, the return was -3.2% – or a loss of $93,000.
“The probability of obtaining a return greater than or equal to $98,865 in 56,435 bets using a random bet strategy is less than one in a billion,” the researchers write. Fairly conclusive, then.
There was a problem, though: these generous odds are typically only available very close to game time, and aren’t always available to regular gamblers. “We decided to conduct a more realistic simulation in which we placed bets at odds available from one to five hours before the beginning of each game,” the authors explain. To test this, they built a bot that collected odds from betting sites across the world from September 2015 to February 2016. This approach gave them a yield of 9.9% – or a $34,932 profit from 6,994 $50 bets (although this time a random strategy also gave a profit of $825).
It was now time to put real money where their virtual mouths were, and they did so over 265 bets of $50, generating a profit of $957.50 over 265 bets. The number of bets taken was lower because they didn’t have someone watching the markets 24 hours per day, and missed some valuable opportunities as a result. Still: a return of 8.5% is not to be sneezed at.
The bookies strike back
The bookies, on the other hand, would like to sneeze at this method. “Although we played according to the sports betting industry rules, a few months after we began to place bets with actual money bookmakers started to severely limit our accounts,” the team writes. They found that their odds would be limited, or insist that their bets undergo a “manual inspection” before they could be accepted.
This doesn’t feel fair to the team. “The sports betting industry has the freedom to publicise and offer odds to their clients, but those clients are expected to lose and, if they are successful, they can be restricted from betting,” they write.
Not only is it not fair, the researchers believe it could also be illegal, writing that “advertising goods or services with intent not to sell them as advertised, or advertising goods or services with no intent to supply reasonably expectable demand but with the intention to lure the client to buy another product (a practice often called ‘bait’ or ‘bait and switch’ advertising), is considered false advertising and carries pecuniary penalties in the U.K., Australia, and the United States of America.”
For now though it turns out that immutable rule of gambling remains true: the house always wins. Even if it did get a scare this time around.