The Casino Answer Man

You walk up to a roulette table and see on the board that the ball has landed on number 17 three times in the last 12 spins. What do you do?

Do you jump on board, betting on 17 or on combinations that include 17, figuring it's a hot number?

Or do you avoid 17 altogether, thinking that it's due to go cold?

I've had people send me betting systems based on both scenarios. One fellow sent me a brochure and asked for advice on how to market a system based on avoiding numbers that have hit recently. Another man insisted that the way to win at games of chance, especially roulette, was to ride streaks and trends. If a number has been hot, he said, keep betting it until you're sure the streak is over.

There's no mathematical reason to expect a hot number to stay hot, or for it to suddenly turn icy. The wheel doesn't know what numbers have hit. Every spin is a new trial, and past results have no effect.

Still, when I hear such things from enough people, I like to do little tests. A few years ago, craps players were telling me the way to make money was to wait for two passes to establish a trend, then to bet the pass line. I responded by charting 1,000 incidents of the shooter making two passes, then recording the next result. The test found no tendency for the dice to stay hot--the results stayed pretty close to the mathematical expectation.

For roulette, I've done a shorter test, a wholly unscientific, fun little trial to see if either approach to a hot number could bring a profit. It's too short an experiment to have real statistical validity. Nonetheless, I wanted to see if any trend toward hot numbers recurring or going cold would emerge.

Over the course of several months, I kept an eye on the roulette results boards at the casinos on my rounds. Whenever I saw a board that showed the same number had hit three or more times in the last dozen spins, I took note. I also took note of the next 38 spins, to see if the number in question would hit more or less than the 1-in-38 average we expect of any number on an American double-zero wheel.

I kept track until I had 100 trials, leaving me with a record of 3,800 spins of the wheels.

In 3,800 spins, any given number should turn up an average of 100 times. In any one such trial, a number will turn up a little more often or less often just by random chance. But to be of any use to a systems player, the hot number has to show up either much more often or much less often than expected.

These results were not encouraging for systems players. My 100 hot numbers came up a grand total of 102 times in the 3,800 ensuing spins. There was one trial in which the hot number came up four more times in the next 38 spins; two in which it occurred three more times, and 10 in which it occurred twice more. On the negative side, there were 15 trials in which the formerly hot number didn't hit again at all.

Where does that information leave the bettor? In roughly the same position he or she would be in with no knowledge of any previous results.

Given that the starting numbers occurred slightly more frequently than the average expectation, we can safely say the results didn't support a system based on avoiding these numbers.

What about jumping on the streak and betting the hot numbers? Let's say we wagered $1 a spin on single-number bets on these numbers. We'd risk a total of $3,800. Single-number bets pay 35-1 odds, so on each of our 102 winners, we'd get $35 in winnings plus the return of our $1 wager, giving us a total of $3672.

Even though our numbers came up a tiny bit more often than we'd expect on average, we'd still lose $128 in this trial.

Essentially, the results of this trial are close enough to the mathematical expectation that we can say the odds held up. Instead of once per 38 spins, our numbers turned up once per 37.3. In a short test like this, that difference isn't worth worrying about.

So what good do the boards that track recent roulette results do us? Not much, in terms of helping us devise winning systems.

But they do enhance enjoyment of the game. It can be fun to try to ride a hot number and see if maybe we can luck into a wheel with a real bias. Perhaps there's a slight track worn on the wheel, or unevenly sized frets between numbers are causing some numbers to occur more often than expected by random chance.

That's not going to happen very often at a casino that is at all diligent about maintenance of its wheels. Most of the time, we're going to be playing on wheels at which any given number has a 1-in-38 chance of turning up on any given spin. We can always hope to hit a streak, but don't expect past results to help us see into the future.

For more information about roulette, we recommend: