Chapter Seven: Biased Wheel Strategies.
It is impossible for casinos to keep their roulette wheels in perfect condition. Many wheels develop minor imperfections that produce non-random results that players can detect and exploit.
The facts that biased wheels exist, can be detected, and produce predictable and profitable results is very well documented in a number of books already on the market. I personally recommend Russell T. Barnhart's book, Beating the Wheel. Barnhart points out that casino roulette wheels are, after all, mechanical constructions that experience wear and tear with use. By accumulating tiny nicks, dents, loose frets, etc., some roulette wheels generate non-random results. Barnhart relates a number of stories of how "wheel clockers" have tracked results to discover biased wheels and then won thousands and thousands of dollars on biased wheels.
Since there are 38 slots on the roulette wheel, over a period of time one would expect to find each number hitting an average of once every 38 spins.
However, with a biased wheel, some numbers will come up far above such a rate of once-every-38-spins. If one keeps track of the numbers appearing on a given wheel and discovers which numbers the bias of the wheel is favoring, once could win a good deal of money (as Barnhart's anecdotes demonstrate).
Now, the key to this strategy is patience. If you have played roulette for any length of time, then you know that a given number might hit 5 times out of 20, which looks like the results must be biased. Not necessarily. The Variance Demon giveth and the Variance Demon taketh away. You might get 10 hits on a number in 100 spins, then that number might disappear for 300 spins. Both will happen in the normal course of events on an unbiased wheel.
Accordingly, the only reliable way to know if a wheel is biased or not is to track it for a large number of spins. Based on a number of statistical studies Barnhart has generated a simpler formula for determining at an 80% level of confident if a roulette wheel is producing random results: n/38 +.4 n1/2. That is to say, if you have 100,000 spins, then you calculate 100,000/38 + (.4 x 100,0001/2) = 2758. If some numbers appear more than 2758 times, then the wheel is probably biased or the computer program flawed. The formula for being 95% confident is the same only it is .5 instead of .4.
No, you do not have to wait for 100,000 spins! That was just to illustrate the point. Barnhart suggests that you have at least 500 spins worth of data before deciding whether a wheel is biased enough to bet on. If any numbers have hit 23 or more times in a set of 500, or 33 times or more in a set of 800 spins, then you should assume the wheel is physically biased and will produce a disproportionate frequency of those numbers. If you want more details, then you should purchase Barnhart's book.
The problem with Barnhart's suggested strategies concerning biased wheels is that they involve an investment of many hours and may not produce a biased wheel. But that is if chasing biased wheels is your only strategy.
If you are playing Roulette 2000, you have lots of good strategies to keep you busy while you write down the wheel's results (this is also called "clocking the wheel").
My suggestion is quite simple. If you are intrigued by the idea of a biased wheels, then to become more expert about the subject read Barnhart's book. While you are playing on a table using Roulette 2000, make sure your notation system allows you to later count up the hits on individual numbers. You can do this one of two ways. You can go through the lists of numbers reading them aloud while a partner marks on a sheet of paper each time each of the 38 numbers hit. Or, use a system of marking on the list of pairs that allows you to tell which number hit each time. You can put the hashmark in the upper left-hand part of the space to record the first number of the pair, and the lower righthand corder for the second number in a pair. Or use two different color pens to distinguish between odd and even numbers in each pair. It does not matter as long as you are consistent.
That night after you are done playing for the day, count up your totals for each number if you have at least 500 spins worth of information from a given table. If you find numbers that hit more frequently than 23 hits per 500 spins, then bet on those numbers the next day.
Generally speaking, you should use strict flat betting for this strategy (1 unit per spin). If the number is biased, it will hit often enough to outweigh the house advantage.
I am not a huge fan of this strategy because it takes so much time and effort for uncertain results--most tables you will find do not produce non-random results worth betting on. But if I did find a wheel that had a clear bias, I would certainly gamble up to 200 units trying to see if the bias kept up.
Strategy #10 is to track individual spins for at least 500 spins, then flat-bet on biased numbers, if any occur.