Roulette numbers

I heard there was a merit for 1000 roulette spins (there is). I didn't start recording at the beginning because at first I was just putting it on red and black or even and odd, to maximize my chance of breaking even (knowing that the ball landing on a zero would be a net loss and so I was seeking to minimize net losses).

But then I realized that this is a game. And the way roulette was coded may not be purely mathematically implemented. There could be a flaw in the coding which could enable me to win more or less. So I started recording what numbers I spun. That also means I only got 878 numbers because I have no desire to keep spinning that wheel. Anyway, here's my data from Excel:

The numbers arranged from least spun to most spun (in no particular order): 17, 22, 19, 9, 13, 28, 29, 12, 32, 2, 4, 8, 21, 36, 34, 27, 26, 5, 18, 31, 14, 0, 3, 25, 35, 30, 23, 16, 6, 1, 15, 10, 20, 11, 33, 7, 24

The average number of spins for each number was 23.7 ± a standard deviation of 5.6.

The least spun numbers were 17 and 22 with 16 spins each or just about one and a third standard deviations outside the norm. The median number of spin was 23 spins. The most spun numbers were 7 and 24 with 38 spins each or about 2.5 standard deviations outside the norm.

The ball landed in the first dozen 297 times, in the second dozen 283 times, and in the third dozen 283 times.
The ball landed in the... uhm, first column or bottom row according to this setup 278 times, in the second column or middle row 276 times and in the third column or top row 300 times.

I spun a 0 24 times.
I spun a 1 26 times.
I spun a 2 19 times.
I spun a 3 25 times.
I spun a 4 20 times.
I spun a 5 23 times.
I spun a 6 26 times.
I spun a 7 38 times.
I spun a 8 21 times.
I spun a 9 17 times.
I spun a 10 30 times.
I spun a 11 33 times.
I spun a 12 19 times.
I spun a 13 18 times.
I spun a 14 24 times.
I spun a 15 28 times.
I spun a 16 26 times.
I spun a 17 16 times.
I spun a 18 23 times.
I spun a 19 17 times.
I spun a 20 30 times.
I spun a 21 21 times.
I spun a 22 16 times.
I spun a 23 26 times.
I spun a 24 38 times.
I spun a 25 25 times.
I spun a 26 22 times.
I spun a 27 21 times.
I spun a 28 18 times.
I spun a 29 18 times.
I spun a 30 26 times.
I spun a 31 23 times.
I spun a 32 19 times.
I spun a 33 35 times.
I spun a 34 21 times.
I spun a 35 25 times.
I spun a 36 21 times.

[image: i.imgur.com]


It kind of seems to me, and this is just my gut feeling and may be off, that there are some barriers between the bet and non-bet numbers because it seemed like the ball flowed easily between numbers that were all green, or that were all gray, then started bouncing off of the borders between green/gray segments. But who knows, I could be completely off. Point is, I think the chances of a ball landing on a number might change depending on how you bet, meaning it would be impossible to see whether something was a good bet without actually betting that exact way for at least several hundred spins if not a full thousand spins.

In other words, roulette seems to be a big waste of time and money -- you might win, you might lose, but with no way to actually have a good known chance of winning it comes down to pure luck and is ultimately a waste of your money.

Edit: Added in zero. Brought the standard deviation down from 5.7 to 5.6. Nothing else changed. Also, I added the following graphic, showing how I originally arranged my table, plus the 0 I added on later:
[image: i.imgur.com]

I mean roulette really does have a 3% house advantage no matter how you bet (assuming all of the numbers are being picked equiprobably).

And no, the animation isn't the actual thing that'll determine it. It's already determined, then the animation plays.

> assuming all of the numbers are being picked equiprobably

What, though, if we get a large enough sample size that we can perhaps see that the numbers are not being picked equally? What seed is determining the random numbers?

An arranged list in no particular order, excel tables made unreadable by forming sentences, bar graphs that don't start at 0, a conclusion that you just need to look at overall statistics to know.
Good post.

I'm wondering why you didn't include 0 in your statistics? Could that have some effect on your calculations?

I am going to conduct a similar test in Excel and record all my numbers and see if my results are similar to yours, Escaso. If they are, a roulette player could potentially bet on the most commonly picked numbers which may increase their odds of winning slightly.

Well, hunh. You're right, I left 0 out because initially I wanted to see which spread was more valuable and I kind of forgot about 0. Hold on...
0 was spun 24 times... Ok, fixed and uploaded a new chart. :)

^^