Biased Roulette Wheels

When you are searching for a way to get a statistical edge over the house at the roulette table, there is one topic that appears more than any other: Biased wheels. The idea here is that roulette wheels are imperfect; they carry minor flaws from manufacturing processes, installation, human error, and wear and tear. When these minor errors cause some pockets to be more likely to get hits than others, then the wheel is no longer random, and becomes 'biased'.

The Casinos Perspective

Biased wheels on their own do not hurt the casino in any way. Even if some numbers show up more than average, players will still be betting 'randomly' as a whole, and there is no problem. What can hurt the casino, though, is the knowledge of a biased wheel. When players discover a biased wheel, they can bet on pockets that are more likely to win. The casino makes its profit assuming that there are 37 to 1 odds against any single pocket getting a win. Because they only pay out 35:1 there is a slight house edge, and in the long run they will make quite a bit of money.

Imagine though, that a group of players has discovered a biased wheel. They decide to act upon this information and only bet on the most likely numbers. Lets say on this specific wheel, by betting with the bias, the players are able to lower the odds to 33 to 1 against the numbers on which they are betting. The payout is still 35:1, but with the new set of odds, the players have a statistical edge over the casino.

Because this information can directly cut into the casino's bottom line, the house will do everything in its power to minimize wheel bias. While there will always technically be some imperfections in a roulette wheel, it is the casinos job to make these imperfections so small that they become negligible.

What Makes a Wheel Biased

Any imperfection in the roulette wheel that affects the outcome of the game in any way is technically biased. There are a few different ways these imperfections can be created, and the casinos do everything in their power to prevent and erase these errors. The most obvious, but least common cause of roulette wheel biases is manufacturer error. Because of tight quality control standards, roulette table manufacturers make wheels that are nearly perfect in symmetry, weight, and texture. Generally wheel biases occur after they are sold.

As roulette wheels are use, just like every thing in life, they begin to break down. Normal wear and tear varies from warped axels (crooked wheels) to chipped/battered pockets (formed by collisions with the ball. While normal wear and tear is anticipated by the casino, occasionally a wheel will begin to wear out much faster than expected, and for a window of time, players can take advantage of the resulting bias.

When being installed, a few basic precautions need to be taken to be sure that the wheels are fair and random. The most basic step in the installation is probably the most important, and receives a ton of attention: leveling the table. If a table is even a fraction of a degree from being perfectly level, in the long run, it can result in a bias, and expose the casino to the meticulous players who monitor the tables.

How Biases Are Detected

To find a bias, both the players and the casino are constantly monitoring the tables. To be certain that a bias exists, you need to have a huge record of spins. Ideally you will have seen at least 4,000 spins before you can find a wheel that is truly biased! When looking for biases, it is important to remember that whole sections of the wheel will be more likely to hit than others, but the pockets are not in numeric order. Click here to see the number sequence.

With the numbers lined up in the wheel's sequence you can create a histogram of past spins and see patterns as they occur on the wheel. You will probably see the effect of a bias spread over at least 5 numbers, if not more.

Example:

In the histogram above, you can see how many hits each pocket of a real American roulette wheel received in a test. The taller the bar, the more hits the number got. While it looks pretty even at first glance, put your mouse over the image to see the average. Clearly 24, 36, 13, 1, 00, and 10 all beat the average, and while it seems like a small difference on the graph, it is big enough to swing the odds in the favor of a bettor who knows where to bet.