How To Win At Roulette

One thing I realized quickly when I first worked at a casino, is that people want to believe there is a system they can easily use to beat the "house." The other thing I soon realized, is that there are ways to get the statistical edge over the casino. Unfortunately, they are not easy.

Card counting, for example, really can give you the edge at blackjack. The problem is that it is often a small advantage, and there is a lot of work and self discipline required. I remember one "regular" who I think made money consistently at the blackjack tables, but even he wasn't getting rich.

You see, you may only get a 1/2 percent edge over the casino. If your average bet is just twenty dollars, and the dealer deals out sixty hands per hour, you will make about $6 per hour in the "long run." In the short run you will lose or win hundreds on different days. That's with perfect play and $20 bets.

Do you have the stomach for $100 bets, meaning on some days you'll lose thousands of dollars? By the way, if your eyes glazed over from this little bit of math, you don't have what it takes to beat the house. It's all about the math.

Biased Wheels

Okay, there are great books out there if you want to count cards. This, however, is a story about roulette. One of my favorite players used to win regularly. When I ran the table, we would talk philosophy while he patiently made his simple bets all night. I knew he was winning, but not how much. After I quit the job, I met him for coffee and discovered that he had made over $80,000 in sixteen months of part-time play. That made it more interesting.

In case you aren't familiar with roulette, here's how it works. The dealer or "croupier" spins the wheel in one direction, and the ball in the other. There are 38 "pockets" on an American wheel (1 through 36, plus 0 and 00). Players place their bets, and when the ball falls, they are paid according to the number of that pocket. We'll ignore all the various bets and concentrate just on the "straight up" bets. These are when you bet on one number. You get paid 35 to 1 if your number comes up.

There are 38 pockets, and you get get just $35 for each dollar bet, plus you keep the bet. You can see that the house has an edge (5.6 percent, to be precise), but what if certain numbers came up more often than they should - more often than 1-in-38 spins? That's when you can make money.

Let's suppose, for example, that number 5 is coming up an average of once every 29 spins. If you bet ten dollars on it every time, you would lose 28 times, or $280, every 29 spins, but win once which would pay you $350. In other words, in the long run, you would be making $70 for each 29 spins. ($350 minus $280) When there weren't many customers, I sometimes did 60 or more spins per hour, so you can see that this could be very lucrative.

So why would that number or any other come up more often? The short answer is Who Cares! The longer explanation has to do with the nature of the wheels. The pockets could be manufactured imperfectly, with one or more slightly larger than the others, therefore catching the ball more often. One or more of the dividers between the pockets could be loose, meaning it absorbs the force of the ball instead of bouncing the ball away. The ball would therefore tend to drop into that pocket more often.

There are other reasons, including more temporary ones, like a drop of sticky pop in one of the pockets, or a build-up of dust. The important point isn't what causes a "biased" wheel, though. The important point is that biased wheels exist, and can be taken advantage of.

Why would a casino let this happen? Roulette wheels are expensive, and so they are not often replaced, unlike cards and dice, which casinos replace daily. This means that if there is a bias, it sometimes remains for months. I know for a fact that managers where I worked were aware of the problem, but as long as the table made money overall, they were too lazy to worry about one guy making money on it.

Charting a Roulette Wheel

John (not his real name) came in initially with two friends, as he explained to me. They took turns "charting" the wheel. This is nothing more than writing down the number that comes up on every spin. They did this for weeks on both roulette wheels in the casino. It is an incredibly boring and crucial part of the process, often amounting to nothing, since there may not be a bias.

You see, if the number eighteen comes up ten times in a hundred spins, it is meaningless. The sample is too small. You need to know if the bias is real, and therefore will continue, or is just a statistical fluke. Without getting into probability theory and a discussion of standard deviations, suffice it to say that if you want to be fairly certain the bias is "real," you need a sample of about 5,000 spins or more. They eventually had 15,000 spins recorded in a little notebook.

As it turned out, the number "0" was coming in 1-in-28 spins. This surpassed the 1-in-33 threshold they needed to make it safe and worth their while. Only John had the patience, though, to continue sitting there night after night, placing one bet on one number, over and over. Within a couple weeks, his friends quit. They didn't have the patience required, and probably also didn't like the fact that even with the odds in their favor, they had nights when they lost as much as $700.

So night after night John sat there discussing politics and philosophy with me and the other dealers, placing one bet on 0 for each spin of the wheel. He was making between $50 and $100 per hour depending on the number of spins per hour, and assuming the bias was consistent in the long run. Eventually, after more than a year, and $80,000 in profits for John, the casino got a new wheel.