What would you do with this?

Hello everyone,

I’ve read this part of the forum and i would like to ask you something about a supposed biased wheel i found in Belgium.

When i’ve collected 4000 spins on this wheel, it appears to be balanced but the “anti clockwise” shows a biased with a chi square = around 60. Maybe 2000 spins is definately not enough to declare something about that “bias”, but i’ve noticed that the rotor is wobbly with the distorsion of the light that reflects on the rotor when the dealer is spining. That’s the only thing i’ve noticed except some residue in the pocket and dust on the rotor.

I’d like to know if you could help me to exploit this wheel with the data i’ve collected.


We should take a look at the stream or total hita per number at 2000 trials.
The chi-square test isn´t the best tool to know what to do.

Let’s see what you have.
Based on what you’ve said so far, let’s see if the bias correlates with your observations.
First off, what brand is the wheel? Meaning what’s the name on the top of turret? (The turret is the tall post in the center of the wheel. The name of the manufacturer is usually printed here).

Let’s talk.


Hello everyone and thanks for replying!

Well, as my nick says, i’m a newbie bias player. That was the first wheel i was studying and this is why i don’t really know how to do with this.

First i’ve collected more than 8000 spins and i had a balanced results… This is the data clockwise and anti clockwise file attached.
I went there back and recorded 2760 spins anti clockwise, because clockwise direction showed a balanced result after 600 spins.

The anti clockwise showed a supposed bias for a long time with a Chi square above 60 and a sector of 7 numbers at 3.4 sd if i remember well, but suddenly yesterday after entered 500 spins more anti clockwise, i had a 44 chi square and a balanced result.
I’ve checked the wheel and there is still residue in the pocket, still dust on the rotor (so the wheel hasn’t been cleaned) and the ball scatter seems to be the same. :o that was probably just random or conditions have probably changed.

The wheel is a Caro. There is no wood. The fret are oblique on the rotor side and flat on the number side and are quite high. I tried to find a picture of the wheel on the net but i can’t find it.

What should i use to test numbers a bit more safely? Because even if this wheel is balanced, i have noticed another wheel Cammegh Mercury 360 with a lot of deep scratch on the rotor. If i can find a way to take a picture i’ll post it.

Have a good day

data clockwise and anti clockwise.txt (29.1 KB)

—data a. clockwise.txt (10.1 KB)

7766 both sides spins, random
2761 clokwise spins, nothing out of random

My impressions.

best regards

Thanks for your impressions. I’ll wait a little bit more next time before opening up a champain bottle when i face to 63 chi square. :smiley:

What should i use to test more safely? There is another wheel with scratch on the rotor in front a sector of 6 numbers and two other sectors of 2 numbers, anyone could notice it. So there is 10 numbers that i’d like to test. Is a sample of 1350 spins is enough to make an idea about it? or i’ll have to collect 10K anyway, even if i know yet that there is scratchs ?

I’ll try to take a picture of it soon.


It’s a long trip to master the best way to do what you asked.
Try the standard deviation test, read the math sideof an “unbiased” wheel and compare it to your wheels.

Ok thank you, i’ll take a look and make some research this week. This going to be a long trip but this time i’ll try to take the more details i can. By the way if you have any advices, i take everything :slight_smile:


It’s cool that you took the time to patiently collect so much data. Unfortunately this wheels too random. Don’t give up yet though, try tracking another wheel. :slight_smile: If you can share the data earlier, we can help prevent you from tracking wheels that are too random.


Yes it’s a lot of work, but i don’t think we can win without working hard at this game. I threw 5000 spins away before those ones because i didn’t separate the two directions.

Yes unfortunetely this is a random wheel, anyway for a first try, i would have been a lucky cow to find a biased one, especially if i’m tracking blindly.

I don’t have intention to give up yet, it is just starting :slight_smile: as i told before, i find another one with scratches on the rotor. There is 7 wheels and only this one got scratches.
I will send data and a picture of the rotor as soon as i got enough, thanks for prevent!

This is the data i’ll record next time, as i’ll be there, i should collect immediately the more info i can.

  1. I try to find something to identify the rotor/roulette easily if they move it later.

  2. I Check the alignment of the turret 's arms. I mean i check what numbers are in front of the arms. ( I don’t know if this can be changed, but if they can change it, i would know there was a maintenance done recently)

  3. So now i separate both directions, that’s the hard part because the dealers sometimes spins twice the same direction, so i need to see all spins :o

  4. For this second try, i’ll always precise the day and the time i start/finish.

If there is anything else that could be useful to note… like a ball scatter or the drop point data? I take every advices :smiley:


Here’s a tip that you probably already know:

Rather than standing around for each spin, track ALL of the wheels at the same time. Of course, make sure that you keep the data from each wheel separate.

Simply walk by each reader board every 20 -30 minutes and write/record the board.
As you do each wheel, simply glance down and note the direction that the wheel is spinning.
Next assume that it’s going to alternate and add the direction for all of the previous spins on that specific board. If for some reason the direction doesn’t line up, keep the data, but note that mistake took place during that board. Reset the direction with each pass.
Most of the time, even if a mistake took place, some of the spins will still be accurate. Record the board from the bottom to the top.
When you have collected 7500 spins, the occasional partial board of mixed numbers isn’t going to sufficiently randomize your data. In between passes, stay away from the wheels. This will also help cut down on heat from the casino.


Hello everyone,

Thanks for the tip Snowman, i started to track 3 wheels for the begining. I recorded 300 spins today but only 55 on the wheel i want to track first; it opened a bit later than the others…

Today i shot the wheel i was talking about, the picture is a bit blur but we can see in front of 7 and neighbours that there is scratches on the rotor.

I found another wheel that have scratches on the rotor too, but a bit less visible than this wheel above.

Don’t hesitate to let me your opinion about it, thanks.

Nobody? :frowning:

A Cammegh Classic wheel. The ball travels the less by the rotor. Those scratches probably are useless(or not)

I cannot see that scratches would give you an bias advantage (except if you think that the scratches suggest rough handling of the wheel, which might cause bias I guess) The scratches by themselves though… cannot see how they could influence bias.

Hello and thanks for replying,

"A Cammegh Classic wheel. The ball travels the less by the rotor. Those scratches probably are useless(or not)"

Hello Toby, that was short but clear… thanks for saving my time. When i remember the behave of the ball, i admit that you are right.
May i ask what kind of bias should i pay attention on those Cammegh wheels?

I cannot see that scratches would give you an bias advantage (except if you think that the scratches suggest rough handling of the wheel, which might cause bias I guess) The scratches by themselves though..... cannot see how they could influence bias.

Hello Toxic, well i guess that is because i’m a newbie on this subject so i guess the way i think can look a bit silly sometimes. Actually i’ve read Dr spock 's topic and i saw that scratches could cause bias, then when i was at the casino, i’ve noticed those scratches and thought that it may cause a bias.

Anyway thanks for your impressions.

The scratches are all on rhe cone. That would mean the ball would have to run over the cone alot for any scratches to have an effect. I have not seen wheels where this happens often. If you had deep scratches on the track, maybe that would cause a dd somewhere other on the track but a track has to be pretty warped or twisted to cause a dd in my opinion. The scratches on the cone could be a sign that the rotor is mistreated when removed. Maybe the spindle is bent? You might have a wobble on the rotor. If so, it may very well have an bias effect. Personally I do not worry too much about bias. Too time consuming. I once tracked a wheel together with two other guys. 16.5.24 sector definately showed bias after 7000 spins. When we started to play those numbers went dead. Only after noticing a dent on the turret that wasn’t alligned with 22 like the previous dayse (the dent was now at 1) we realized that the number ring was turned.

Coincidence? Whether yes or no, we lost interest right there.

I once tracked a wheel together with two other guys. 16.5.24 sector definately showed bias after 7000 spins. When we started to play those numbers went dead.
Here is very important how you tracked, if you only writed falling numbers that not always will show you what is in reality. Numbers must be tracked in relationship with some other thing such as ball moving direction, wheell speed, faling point, scatering distance... when you track 7000 spins then you do tracking maybe in diferent conditions which can compensate bias or oposite increase . So after tracking you have this way ball faling density acordingly some say averidge conditions. But when you play condittions are not averidge they are some specifick, and you not know which numbers can be favorites in theese specifical condittions. If you track in relationship with certain condittions, you can sort all hits acordingly every feature and when play you can recognize which condittions is in concrete moment and play on hits which was in the same conditions before. This way results can be very very diferent. Of course if the number ring was turned then all is bad, but tracking in right way you not need so many spins and with that do chance for turning - less.

I have a complete method to detect bias wheels with out defect spotting.
I sell a copy if you are interested, you can send me a PM.

It is a valid method.


Hey Lucky. How many spins you have to track.