The Long Bet - Lucky Devils

Kit Chellel spent years embedded with the world's most secretive professional gamblers. What he learned about obsession and process has a lot to say about playing the long game.

The best gamblers in the world aren't really gamblers at all. They're long-term thinkers in disguise: obsessives who've figured out that the path to winning is to stop thinking about any individual bet and start thinking about the next ten thousand.

Superior decision-making, compounded over time, beats luck. The best players know they won't win every pot. They just survive long enough for the math to work in their favor.

Kit Chellel's new book, Lucky Devils, profiles three extraordinary characters who beat the odds not just once, but over and over, across years and decades. Perhaps the most remarkable is Bill Benter, a soft-spoken mathematician who cracked the Hong Kong horse racing market using a proprietary algorithm and, in the process, became a billionaire. Kit wrote a stunning profile of Benter for Bloomberg a few years back; the book expands that world considerably.

The lessons here — about obsession, process, and staying rational when your brain is screaming at you to panic — apply well beyond the casino floor: in business, in investing, in any arena where the long game is the only game worth playing.

Lucky Devils published April 14th. Pick up a copy here — and read my full Q&A with Kit below.

Eric Markowitz: Your subjects — professional gamblers — spent years, sometimes decades, trying to find a fractional edge. What does it actually take to find an advantage that others are missing?

Kit Chellel: I think the primary determining factor is obsession. When you learn how much energy and how many years these guys expend to get a fractional edge, obsession is the only word for it. I've talked to people who've spent a lifetime trying to understand how roulette wheels work. They'll build replica casino roulette wheels in their basement, wire them up to computers, spin those wheels thousands of times a day, and collect data. They just work harder. They want to beat the casino more than the casino is willing to expend effort to keep them out.

I went to the Blackjack Ball a couple of years ago — it's like the big gathering of advantage players in Las Vegas, lots of backslapping and black tie and champagne. I met this guy who said: the reason card counting still works is that we care more. 

His argument was that it's quite easy for casinos to make money from human fallibility — people walk in, put their money on the table, and walk out again. It's much harder to make money in casino games. So the advantage players just put in the research. They go further than any reasonable human would be willing to go, and that's why it still works.

EM: The parallels to investing are, for me, fascinating. You could swap "gambler" for "hedge fund manager" in almost every sentence. What's the core similarity you kept coming back to?

KC:  It's the structure of the market. In both gambling and investing, the majority of participants are losers overall — they will lose more than they win. They're also outgunned. They don't have the right tools, they don't have the information, and they're competing against much more sophisticated players. So you end up with a small percentage of highly successful participants. 

In financial markets, those are your big, consistently profitable hedge funds. In gambling, it's advantage players — and there aren't that many of them. We're talking less than one percent who can sustainably make a good living doing this. It is enormously difficult, just as it is if you're trading options.

There's also an added wrinkle on the gambling side: if you're a successful hedge fund, people want to know your secret, but no one's going to kick you out for doing well. In professional gambling, if you're really good at your job, you risk having your money confiscated. You can be booted off the platform for no reason other than that you're successful. That breeds a very healthy amount of paranoia.

EM: All the advantage players you profiled have exceptional emotional detachment from outcomes. Is that something they learned, or something they were born with?

KC: I think it's both, but the gambling hones it to a razor-sharp point. All the advantage players I've met share a similar mindset: an exceptional ability to detach the emotional side of their brain from decision-making. That's really difficult for most people. The reason we gamble is that we get a neurochemical reward for taking calculated risks. It's evolutionary. We're supposed to explore new territory in search of reward. That's an old, old part of the brain. The part of the brain that deals coldly and rationally with complex problems developed much later, and the two don't necessarily work well simultaneously.

The advantage players I met don't seem to get that neurochemical buzz from winning. They just don't care. Bill Benter spent twenty years developing a horse racing software that could beat the market. For most of that period, he never set foot inside a racecourse. He never watched individual races. He couldn't even name a horse or a jockey. They were numbers on a screen to him. 

Rob Reitzen, my poker player, was developing a statistical system long before anyone else. One of his tests for potential players was to hold their hands out and feel if their fingers were cold. Cold fingers meant the heart was working overtime — stress. He wanted people who could make decisions completely outside the emotional part of the brain. He wanted them to feel nothing.

EM: How did they handle long periods of losses? In investing, we talk about elongating your time horizon — staying in the game long enough for the process to work. Did you see that with these guys?

KC: I think about it in terms of faith. Their religion is science. They believe in hard data and probability the way many people believe in a deity. Bill Benter, when he first went card counting, lost his bankroll almost immediately. But he was actually encouraged rather than discouraged, because he knew the math was right. He knew it worked. He may have lost on that particular occasion, but he walked out of that casino more confident than ever.

Later, he took a huge risk to move to Hong Kong — having never been there — with about $100,000. The first year was a losing year. Again, he didn't take it as discouraging, because he believed in his process. He believed he was doing something no one else in the market was doing. That unshakable belief in the rigor of your process is incredibly important, and actually quite rare. 

Most people, even if they believe they're right, get shaken. When you keep losing, when the dice rolls are against you, it's so tempting to abandon what you've learned. My three subjects just didn't. They had this unshakable faith.

EM: How do you think about survivorship bias in choosing those three? There must be vastly more gamblers who had total conviction in a process that turned out to be wrong.

KC: I chose these three for a specific reason: they did what they were doing for thirty years and made money in multiple different ways, in multiple different places. That suggested they were doing something independent of good luck or bad luck. It's not difficult to go into a casino and make money once. My personal "strategy" when I visit a casino is to walk to the roulette table, put everything on black, and if I win, I take the money and leave. If I lose, I go back to my hotel. That's essentially a coin toss. Does it make me clever when it lands on black? No, I just got lucky. Anyone can do that.

It's when your decision-making pays off over thousands and thousands of iterations that there's something real there. And even within the world of advantage gambling, that's quite rare. A lot of card counters burn out. They have a period of bad luck, lose their bankroll, and do something else. In the hedge fund space, the same is true. You look at the funds that actually matter over twenty years. I often see stories in the financial media saying "top-performing hedge fund up 27% year-on-year," and I think: brilliant. How are they going to do next year? Those are the wrong timeframes to look at.

EM: The gamblers in your book focused on the quality of decisions rather than individual outcomes. How do you think that framework transfers outside of gambling?

KC: By not focusing on individual outcomes, they focus on making lots of good decisions. That's how they make money. They don't really believe in luck, they don't rely on it, they don't need to get lucky. They make their own luck. And the way you make your own luck is by taking lots and lots of very good gambles over time.

That's a concept I've taken into my own life. Don't put all your eggs in one basket. Manage your risks. Take lots of good risks. Don't expect any single wager to pay off. Make lots of good decisions, and eventually you'll see the benefit. The key is managing your resources so you don't go bust before the expected outcome arrives. That, even someone like me — by no means a statistician — can take on board.

Gambling is also brutally clarifying in a way that most professions aren't. If your ideas aren't logical, if you're not making evidence-based decisions, you bump up against reality very, very quickly. You make money or you don't. You get quite quickly disabused of any ridiculous notions you might have. You also learn how to be wrong. You have to learn when to pack it in. You have to learn how to let the evidence of what's happening in the real world guide what you're doing.

EM: Where does AI fit into all of this? Is it changing the game for advantage players?

Kit Chellel: Absolutely. AI is the next big frontier in advantage gambling, and it's already transforming gambling markets, although people may not realize it's happening. The last time I visited Bill Benter, we talked about his recent years in horse racing. He started building his model in 1984. He still bets horses. His firm still operates out of Pittsburgh, still one of the largest syndicates in the space. In the last couple of years, he told me, he's made more progress than he had in the previous ten. And the reason is that he's adopted AI modeling into his system.

In the early days, he had to guess which factors would be significant — the temperature, the jockey's record, the trainer — plucking things out of the air and testing them one by one. With AI, you don't need to do that. You need lots of data. And Bill has forty years of it. He feeds those data sets into an AI model, and with sophisticated guidance, it interprets the factors and weights each one accordingly.

He said it's been some of the most exciting work of his career. If Bill Benter is doing this in horse racing, you can guarantee it's happening in sports betting and prediction markets too. AI is ready-made for this kind of task.

EM: After spending all that time with these guys, did you catch the bug? Did you want to participate?

KC:  Quite the opposite. They made me realize how inadequately equipped I am to do what they do. I don't really bet horses anymore, and when I bet on sports — which is infrequently — I have no realistic expectation that I'm going to do well, because I know what I'm betting against. I'm betting against hedge fund-level technology. I'm betting against a room full of MIT graduates who do nothing but this. And I'm sitting on my sofa with a can of beer watching Premiership football. What chance do I have? Very, very small.

That said, I'd give Bill Benter $1,000 to manage if I could. He's been going such a long time, and he's adaptable. The fact that he's still at the forefront all these years later means something. Gambling markets are enormous now, and they're for sure less efficient than financial markets. 

If you know somebody who has a genuine edge — yeah, why not?