As a new biased wheel player, i have made countless mistakes, such as wheel choices, betting when the statistical significance is not high enough, correlation of current data is not high to past data, betting ratio for different numbers with different advantage is wrong, betting without getting enough data.
Is there a general number of spins you need to collect before playing a wheel? The answer is it depends on the data of a wheel. For some wheels, 10,000 is sufficient while others you need 30,000+ or even 60,000 spins. In short, the more number that a wheel gives positive advantage and the higher is better, and you will also encounter less swings in your bankroll when playing. It is important to understand the concept of standard deviation when deciding whether you can start playing. I have also encountered mimicking bias where despite the standard deviation is as high as 4s.d.(data showing the wheel had a hit rate of 1/25 or 44%advantage) i had lost money playing it becasue it is physically impossible to have an advantage of 44% with this wheel and that the chisquare at that time was around 78 at 5000 spins, the number contributed to this were composed of many numbers which are just short of 2s.d. and using statistics, 2s.d. means that the number has a probability of 68% that it is biased (you need to subtract 1s.d. off it as this is approximate it using the normal distribution). At the same time, the wheel brand also is physically impossible to have those numbers that is that high advantage (around 20% or 1/30 hit rate) due to the design of the wheel and that defect seen is not significant. In the end, the chi square dropped to only 50 at around 15,000 spins. I was pretty sure it is the same wheel. If i had used the pearson correlation test for this wheel, i would had a red flag when correlating data for the 1st 5000, 2nd 5000 and 3rd 5000 spins as those would have been negative and i would have avoided loosing money.
Luckily, since then, i have read countless books about statistics on roulette and have improved and i have known how to beat wheels. However, understanding statistics is not enough and i am yet to investigate into how baromatic pressure, temperature, etc affect roulette wheels. I have a strong reason to believe that bias on roulette machines (automatic roulette ) can be casued by either the pump that pumps out the air or the motor of the rotot becasue i notice the bias are strengthed when the machine vibrates like crazy as you can literally feel the vibration and the ball just drop to the biased number/section.
I have also “discovered” or learnt from statistics material, computer simulatiom using R programing that 1 number with high advantage is not the key but having more numbers with advantage is way too important. In short, assume a number with advantage of 20% or hit rate with 1/30 in one wheel and a second wheel with 20%advantage on each number of perhaps 5 numbers, with the same risk, you can bet approximately 6 times on a wheel with 5 biased numbers. You would have to bring into a casino 800 units to play 1 number but using the same 800 units, you can now perhaps bet 6 units per spin while maintaing the same risk. Advantage is important but the payout and the probability in winning is also important. Having 5 numbers with hit rate 1/30, the chance of winning each spin is really 1/6. Imagine you are playing the lottery, and that you have a 100%edge, you wouldn’t sell all your houses/cars to bet becasue you will go ruin if you bet 100%of your bankroll becasue the chance of winning a lottery is still tiny.