TurboGenius When does it lose ?

I'll get back to your post with the answer. Needs a bit of coding.

 

thnx

 

at spin 37 we have 24 numbers hit at 1/37
these are removed.

 

On spin 74 we have only 9 numbers left who are still behind their 1/37.

5 numbers showed up 2 times, these were back at 1/37 (2/74) and were removed.

Numbers that are still behind would need to hit a total of 3 times to get back at 1/37 (3/111).

 

at spin 111 we have only 7 numbers left, wich means two have reached their 1/37 (3/111) and were removed.

now at during the next cycle, the remaing numbers would have a total of 4 hits before they can be removed...

 

More think , - less write, will be much better.

 

one did and is bein" removed (1/37 4/148).
6 numbers remaining. etc etc well you get the point.

 

I coded the test. The results are very dispersed. I repeated the test for 50 times, and here are the max number of spins that it took for all the numbers to meet 1/37 ratio in 50 independent trails:

40580
10967
1401
8748
11968
29203
20455
1174
4621
22533
7441
2047
824
10306
13363
149883
450
2842
5380
10561
70310
7249
7932
1842
3565
27643
1769
9593
1138
89475
48389
7807
1126
11226
93767
379
678812
2792
15851
2421
18955
12298
882
2575
630
12720
3382
2102
81498
269481

Good luck.

 

I didn't expect that. Playing sleepers gets scary very often lol

 

If you read my post I asked to investigate the details of regression to the mean for all locations.

 

What post?

 

Thanks pal, for the time and effort. much appreciated.

 

Anytime.

 

https://www.gamblingforums.com/threads/gigis-hr-tests.21346/page-4#post-127483

 

My current understandings make me answer "Chaotically" to both questions since I cannot realize any obvious pattern in the above movements. I have spent hours looking at the changes over time, but it didn't spark any ideas. I tried to guess the answers but had no luck. I wonder how likely it is for someone to stumble upon the answer randomly or by brute-forcing the problem. If I had a penny for every way I tried to answer the above questions, well, I would have $10 .

 

The way to solve a complex problem is to simplify it.

The way to simplify a complex problem is to first count the number of elements that contribute to this complexity.

Then reduce complexity by reducing the elements.
Best to start with the simplest form.

 

I find variance to play the biggest role in this complexity. This so-called variance is that crazy randomness that chooses how each number or location regresses towards the mean every time an imbalance is present in the game.

 

Five random snapshots.

Focus on the 2hits in the first 37 spins(blue).

Is there any orderliness by the 3rd 37spins?

Solve this you solve rtm.

 

Regression to mean is not what many think it is. It's actually the average (i.e. the ratio) that regresses to the mean, so hit / non-hit is what regresses to expectations.

So toss a fair coin 10 times and get 7 heads vs 3 tails, the ratio is 7/3 or 2.33.

Now continue tossing the coin 100 times and get 67 heads vs 37 tails, the ratio is 67/33 or 2.03

Now continue tossing the coin 500 times and get 300 heads vs 200 tails, the ratio is 300/200 or 1.5

The actual count of events do the opposite and diverge (i.e. drfit away) from the mean over many spins. So after a million spins you may have a thousand more reds than blacks.

The divergence in regresion to the mean becomes a fraction of a decimal point over millions of spins.

 

So here is what regression to the mean does as it's the ratio of heads vs tails (or any EC bet):


And here is what the actual accumulated counts do (notice how they drift further and further away).


There is no paying back the house when short term variance favors you. The wheel, card, dice or coin (for two up) doesn't make up for past "errors" to even up the score, but the overall average will drown out the "indiscretions" in little sir Anyones viewpoint at infinity.

Sessions are played, won and lost on the occurance of events where variance is king.
The count of events will diverge over time and get further and further apart (whilst the ratios of those events get closer). The divergence aligns with physical laws of entropy/disorder increasing.

The law of large numbers is absolutely irrelevant to what happens in a session. Who cares after 1 mission spins what the net ratio is. The ratio and counts within a session is what matters.