[ QUOTE ]
[ QUOTE ]



If you have 1-million players all playing -EV blackjack at a gigantic casino (about 1-2% disadvantage) then after 100 you will probably have 45% of your players showing a profit and 55% in the red.

This does not mean that blackjack is beatable for 45% of players though.


If you let them keep playing until they reach 1,000 hands then you will probably only have 40% or less showing a profit.

after 10,000 hands then probably 10% will still be ahead.

after 1,000,000 hands I would guess that less than 1% of your 1-million players will be break-even or higher.



So - back to the original point. If you have 1-million players who all played 100 or 1000 hands each you still don't have a sufficient sample size to determine how many of those players are true winners.
In this game, 0% win in the long-run because everybody is at a disadvantage.
but to somebody who doesn't know how to interpret the data it will LOOK like 35-40% of players are actually winners (when it's obviously just a function of variance).

[/ QUOTE ]


According to my PT database, this isn't true. It is NOT a function of variance.


I have 547,250 hands - many of them through datamining.

If I use a date filter all the way back to when I first started, here's what I get :

Hands % winners

8521 39.17
15,335 39.94
35,902 40.89
57,065 41.27
123,600 41.77
273,665 41.76
478,594 41.64
532,233 41.82
535,064 41.77
547,250 41.65


The % winners does NOT decrease as the total # hands goes up, as many posters have suggested.