Caspianhero's Blog

You seem to be saying that the 1,000 Real World simulations is not in exact accord with expectations... should it be?

Also if you convert 00058 to a percent for 100 games and then multiply by ten for 1,000 games, its not so rare to have a streak of ten.

This would make for a good thread in a forum.

I can make one if you do not know how after NCAA Basketball tonight.



Are you counting the ties in the streak, meaning a tie breaks the streak? That will change the probabilities a bit.



How did you get 0.00058?

This does not look right.



Banker win prob

In an 8 deck game

0.458597^10 counting ties 0.000411445

0.50682483^10 not counting ties 0.001118353



In a 6 deck game still real close values.



I gots to run. so add some comments from my questions and maybe start a new thread.



Counting ties in 8,000 hands (about 1000 shoes)

(Per shoe probabilities are even a bit smaller. I do not have those results with me at work.)



For Banker

expected # of runs 1.78

at least one 10 run 0.832105832341954

at least 2, 10 run 0.531808302991253

at least 3, 10 run 0.263956469967264

For Player

expected # of runs 1.39

at least one 10 run 0.750518832600418

at least 2, 10 run 0.403513638543370

at least 3, 10 run 0.162821896198648



Not counting ties coming later

The above numbers are not correct. The problem of being in a hurry



From The Wiz 1,000 8 deck shoe

I get

83,640 hands

(streaks are Not counting ties

This is per shoe)



For Banker

expected # of runs ~43.1

actual 41

For Player

expected # of runs ~33.1

actual 31



Counting ties in 1000 shoes coming later