11-17-05 [poker] - 1

11-17-05 [poker]

As usual, Party Poker has thoroughly fucked up on its deal making algorithm. They use a totally incorrect formula which doesn't take the prize structure into account correctly at all.

The correct way to do it is to first assume that each person's chance of winning 1st place is equal to their chip count divided by the total chip count. Then you look at the remainder assuming each of those did happen. eg. assume player 3 won first place, now assume that the chance of winning 2nd place if that happens is equal to each person's stack divided by the total remaining chips (with player 3 gone), and recurse.

So, I just coded this up and put in the GoldBullion poker suite. (but I haven't uploaded the new code yet).

For example, the correct payouts in the example they provide are :

Teddybear : 1556.918743
JellyFish : 1141.400077
Ace       : 1041.893275
Commandor : 572.407355
Fantastic4: 517.167609
MrRomeo   : 458.542746
TangoKing : 347.258367
Gladiator : 345.989316
Franklyn  : 340.909027
Bluesky   : 227.513486

The big difference there is a lot less for the chip leader and a lot more for the smaller but not smallest stacks.

No comments:

old rants