@ Caleb regarding your biased numbers. Genuine questions

[Turner]

Caleb

DISCLAIMER: at no point am I doubting the validity of the spins. I am accepting they are genuine and taken from a biased wheel.


I ran your biased numbers through RX and found them to fail a chi square test.(I used 10K to match the random.org spins below)

#17 is the number that hits well above average.....and #24

Numbers average around 268 hits with a low of 224 but 17 and 24 sitting around 360 hits

I ran, also, 10k random.org and it passed a chi square test.

12 underhit at 235 spins where av. was 268 spins, with #24 hitting 298 times.(I can see 360 is un naturally higher than 298)

So....you go to the casino armed with the fact the wheel is biased on #17 and #24

Now, these numbers hit 360. A lot of that is natural and within random, but some is bias and not within random.

So, this time, this wheel could naturally under perform on #17 or #24 like 12 did above, and with the bias hits added, look normal.

in other words, you dont know how well naturally that number will perform, and your calculated 15% or 20% edge could be nothing like that when you re visit the wheel.

1. Its not a known. some luck is needed.

2. Also, that chi square test didnt turn FAIL for 400 spins. Thats random too. You could be in the hole deep after 8 solid hours play.

3. Also, you are relying on past spins for the natural times your biased number hit and it can perform so differently over the next 10 k

(this is a genuine attempt to understand. Please treat it as such)

[nottophammer]

caleb looking for the wobbly wheel


Such a wise monkey

[The General]

1. Its not a known. some luck is needed.

I don't know what you're trying to say here.  Are you trying to say that you're not sure if the wheel is biased?  I would hope that you'd be absolutely sure that it was. 

2. Also, that chi square test didnt turn FAIL for 400 spins. Thats random too. You could be in the hole deep after 8 solid hours play.

And why are you running a chi square on just 400 spins? 

3. Also, you are relying on past spins for the natural times your biased number hit and it can perform so differently over the next 10 k

When measuring the fitness of the gambling device you can,  of course, use past spins.  Actually I have probably another 20k or 30k spins from that wheel.  I played off an on several different times.  The wheel has an indexable number tape, meaning they were able to rotate the numbers.  However, whenever this happened it was quite easy to find out how much it was rotated by scratch marks.etc.  There's no doubt that the wheel was biased.   Number 17 was at 7.01 standard deviations and number 24 was at 6.18 standard deviations.  That's not chance.

[The General]

Turner,

Do you know how many spins it took me to determine that the wheel was biased?

[Turner]

Ii will reply tonight....thanks for the replies.
Granddaughters birthday and cant reply properly on this phone

What I will say is this.

You show also 17 is high at 7SD

I am trying to say something else

Lets imagine I have something in my eye and it makes me blink too much.
Some expert could count my blinks and determine I have something in my eye because I blinked 50 times per minute instead of 20. He didnt examin my eye....just counted blinks.

Some of those blinks were normal automatic blinks....some because I have something in my eye

The expert cant say which was which...just that I am blinking unnaturally.

Some of 17s hits are normal and that varies.
When I come to the table....I dont know if 17 will underperform regardless of the bias.
You have determined bias from a big sample
You play in a tiny sample.

Understand?

[Steve]

Need more variables. Do you have tourette's?

[Bebediktus3]

Some of those blinks were normal automatic blinks....some because I have something in my eye

The expert cant say which was which...just that I am blinking unnaturally.

Some of 17s hits are normal and that varies.

But for expert does not need to sort which hit was natural which was because the number is biased 
The same like which is the difference if you braked leg jumping from 1 meter or from 2 meters high
Important is only result and that result is that concrete number is biased and because of that it has the chance to catch ball more times than mathematical in future.

You have determined bias from a big sample

yes that to put to gypsum your braked leg you first must enter to medical university finish studies so spent for that 10 years...

[Turner]

Thanks..I get you..but in a nutshell....I am not saying this or that is so.
I am sayin...I am wondering if this is so

If 17 is biased....its high SD isnt all down to bias
If you filled 17 in with bluetack....yes...it never hit all down to the blue tack....

[Turner]

Need more variables. Do you have tourette's?

Well...steve...I really looking forward to you commenting because you can see the point I am trying to make rather than taking the piss

[Steve]

Ok serious response.

Some expert could count my blinks and determine I have something in my eye because I blinked 50 times per minute instead of 20. He didnt examin my eye....just counted blinks.

Thats just a statistical anomaly. If we rely only on statistics, then we need a lot of data. The more data, the more assured we are the anomaly will continue.

You can greatly reduce the amount of spins needed if you verify bias visually. There are numerous ways. Still it isn't 100% assurance of bias, but you don't need 100%, like the casino doesn't need you to lose on every spin.

[Steve]

Also you might test 100 spins and see which won most. If theres real bias, theres a better chance it might show as hot numbers. But theres too much variance to be sure. Do you need to be sure? No. But theres too much of a chance the variance is too great for reliable analysis. Taking such a shortcut might turn a potential +1% edge into a -2.6% house edge. Or you might be wrong and avoid the bias, and end up with -2.8%,which is worse than the house edge.

So you must take into account more data. Not just spins, not just visual, but also what the data says.

For example, you can notice a half-wheel bias in as little as 500 spins. You should still verify it though. And again a statistical anomaly is just that, and not a guaranteed bias.

There's a lot more to it. But basically sector bias like half wheel requires fewer spins.

In any case, it's a big waste of time and impractical to consider only winning numbers alone.

[Turner]

Ok serious response.

Thats just a statistical anomaly. If we rely only on statistics, then we need a lot of data. The more data, the more assured we are the anomaly will continue.

You can greatly reduce the amount of spins needed if you verify bias visually. There are numerous ways. Still it isn't 100% assurance of bias, but you don't need 100%, like the casino doesn't need you to lose on every spin.

Ok...
But am I way out by wondering if a biased #17 isnt all bias. Some hits are natural and those natural hits can underperform like sleeping for 600 spins in the short term at the table. It will hit because of sime defect but not above average
Just trying to understand

[Steve]

Yes again what you described could easily be just a statistical anomaly. Simply more data means more assured. We can never be 100% sure but don't need to be.

The data may suggest bias for 2 wheels but it's just statistical anomaly and we lose $5000. Then on a third wheel, we have a genuine bias and careful milk it for $30,000.

Thats just looking at winning numbers. Again you look at much more.

I personally wouldnt bother with bias in most cases. There are better and quicker ways.

[Roulettebeater]

caleb looking for the wobbly wheel


Such a wise monkey

Is this wheel biased ?
I am not believing my eyes, the ball is always coming to rest  on the left side

[Roulettebeater]

Hey Steve

I have a question, my system is providing me two hot points on the wheel ( the two points are -18 pockets far each other, so a diameter of 180 degree).

As I don’t have a computer that measures the speed accurately I am using these two points with their neighbours in my bet.

Everything looks fine, except when the wheel travels so fast, then I noticed that the former predictions are no longer valid, instead of the -18 pockets distance (or 180 diameter), the distance under actual conditions drops to 90 degree ( -9 pockets ).

Unfortunately, it’s hard to distinguish the wheel conditions on time, therefore i am looking for a way to solve this problem.

Do you have any advice in this regard?

Looking forward to your reply