Biased Roulette Wheels

Are biased roulette wheels so few and far between nowadays that nobody is out there searching for them. I remember reading an interesting (I think out of print) book (I don't know how true it was) about folks who tracked roulette wheels in hopes of finding ones that are out of 'true,' if that's the right phrasology. The book indicated some had had success at the endeavor.

Anyway, in one of the hotel reviews on this site, the author made a note that the Wizard asked him to note that the roulette board pictured behind the female model indicated a certain number had appeared a disproportionate # of times. If I am not mistaken, I believe the Wizard calculated the odds of the number coming up that frequently was over 100,000. So, it made me think, do casinos do such a good job inspecting their wheels for imperfections, or are there perhaps 3% or so of wheels worldwide that have an exploitable bias?

Obviously one cannot take paper and pen into a casino to record past roulette results off the board, so do folks plant hidden cameras to track wheels' results (if wheel-tracking is even practiced nowadays.)?

I doubt wheel tracking is practiced anymore and as an activity I think it would just about edge out that famed alternative activity: watching paint dry!

Wheels are no longer subject to manufacturing defects or physical abuse by being moved about in traveling gambling halls or moved abut by casino cleaning crews.

You can spend your time tracking a wheel all you want. It won't make you any money.

By the way, I don't know what photo you are referring to but I wonder: as to that oft repeating number, do you plan to bet on it or against it? If the wheel is out of true then why are there prior winning numbers that are no where near that particular repeating number?

I almost think you CAN track numbers by paper and pen, specifically because it makes not one bit of difference.

With the tech today, finding a biased wheel is nearly impossible. Maybe in some rundown, half assed joint that gets their equipment second or third hand and doesn't have the resources to maintain them, maybe. But I don't know of where this place is, or even if it exists. Most wheels today are many thousands of dollars, receive monthly if not weekly maintenance, and are checked every time the games are opened for balance. Of course as with everything, this relies on personnel to perform their jobs properly and up to standards, which makes a biased wheel plausible. I wouldn't bet the rent money, though.

Quote: StevenBlack

Anyway, in one of the hotel reviews on this site, the author made a note that the Wizard asked him to note that the roulette board pictured behind the female model indicated a certain number had appeared a disproportionate # of times. If I am not mistaken, I believe the Wizard calculated the odds of the number coming up that frequently was over 100,000.

You're talking about this photo:

It's from the Bally's review.

Quote: Bally's Review

Bally's has seven 00 roulette tables. Limits ranged from $10 to $15.
The Wizard wanted me to mention that in the history board on the picture above the same number hit 6 out of 15 times. He says the probability of that is about 1 in 15,800. If you don't consider the bottom four numbers, it hit 6 out of 11 times, for a probability of about 1 in 172,000.

While that is quite unusual, it does not signify a biased wheel. What's most amazing to me is that, apparently, the table closed shortly after that event.

I use that photo, and two other interesting history displays, on my Poker For Roulette website, Misc page.

The second is on the Four Queens review page.
The third I took myself at Sands, Bethlehem, PA.


----


To answer your original question, the average gambler, attempting to track wheel bias, hasn't got a chance.

Those history displays do a lot more than entice players to bet on black when 7 reds in a row show up.

TCS John Huxley, the premier manufacturer of Roulette wheels, also makes the sensor and history display. At the same time, they make software to track and report possible wheel bias, among other things.
Admin note: removed image www.djteddybear.com/images/wheel_bias.png
To the untrained eye, it would appear that there is a bias in the chart shown. If so, the casino will know about it, and correct it, long before a gambler can find and exploit it.

For more info: http://www.tcsjohnhuxley.com/en/gaming-systems-and-security/roulette-wheel-analysis.html

Quote: DJTeddyBear

That display on the right is unbelievable - 1 triple & 5 doubles in 17 spins. It's enough to put even MrJJJ off.

----

Quote: DJTeddyBear

Admin note: removed image www.djteddybear.com/images/wheel_bias.png



............To the untrained eye, it would appear that there is a bias in the chart shown. If so, the casino will know about it, and correct it, long before a gambler can find and exploit it. ....

Hopefully, any corrections are not carried out remotely.

Quote: DJTeddyBear

Admin note: removed image www.djteddybear.com/images/wheel_bias.png
To the untrained eye, it would appear that there is a bias in the chart shown. If so, the casino will know about it, and correct it, long before a gambler can find and exploit it.

For the record, I believe the data used to generate that chart is either a particularly small sample, or perhaps specifically selected numbers to produce that interesting chart. The peaks and valleys are too close to be a long-term chart.

I would think that a real chart would have a much smoother, circular form. A wheel with a bias will have a chart that looks either circular, but off-center, or oval.

At least in my untrained opinion.

I think there is evidence that some roulette wheels are deliberately constructed as biased. An easy way to do this would be make the slots for some numbers a little wider than others. A description of a wheel that I think may have been constructed in that way is to be found in an article "Roulette Bias Exposed" by Jeff Murphy Link.

Murphy's wheel had a chi-square value of 94.59, which is high. Chi-square for an unbiased wheel ranges from 20 to 40. Murphy says the industry standard for a biased wheel is 55 or over. Murphy noted a strong bias in favor of the numbers 0, 25, and 29. Zero is a number that strongly favors the house because it does not pay off on side bets. As Monk would say, here's what happened.

A slippery casino commissioned a wheel with a bias toward one of the house numbers, 0 or 00. A slippery carpentry shop built it but included a "back door" so that someone affiliated with the shop could take advantage of it once it was installed. Plausible? Think about it! That is a really biased wheel.

The chi-square test is used in such cases because it is easy to calculate, however it has been criticized as being an approximation valid only for large samples and substantial numbers of degrees of freedom. In this computer age it may become possible to evaluate wheels using the multinomial distribution, which gives exact results, however it is a bitch to calculate. Stattrek.com has a multinomial calculator, however it will not accept more than ten variables and even with that number takes about a minute to return a result. It might be possible to program a routine for calculating the multinomial probability of a roulette wheel in Matlab or whatever. One would input a vector of frequencies, a vector of probabilities and Matlab would calculate the exact probability of that vector of frequencies.

Quote: statman

I think there is evidence that some roulette wheels are deliberately constructed as biased. An easy way to do this would be make the slots for some numbers a little wider than others. A description of a wheel that I think may have been constructed in that way is to be found in an article "Roulette Bias Exposed" by Jeff Murphy Link.

Murphy's wheel had a chi-square value of 94.59, which is high. Chi-square for an unbiased wheel ranges from 20 to 40. Murphy says the industry standard for a biased wheel is 55 or over.

I looked over the critical values for a two-sided 95 percent chi-squared test and they are 22.106 and 55.688, as there are 37 degrees of freedom. An unbiased wheel would be in the 95 percent confidence interval [22.106, 55.688]. So, yes, above 55 would be suspect.