06-30-2016, 03:24 PM

PlayOK does audit games. There's a PBN button in the History tab at the table. You have to copy/paste into something, but Notepad is fine. So I'm going to try to log a few games and get some data for analysis. I'm still seeing ridiculous things, IMO, like this:

KKJJJTTT.AAQT.T.AAKKKQQ

opposite

AAAKQQJT.KJ.AAKKJTT.AQJ

This is a *perfect* fit. Declarer started trump aces...and dropped Ax. The club suit (the suit order is S-H-D-C, because this is PBN...look at the rank ordering) set up dummy's QUEENS to win the last 2 tricks, after an obvious crossruff.

And, sure...MAYBE this can happen once. But later on in that same game, there was this:

K.AAKJ.KQQJJTTT.AAKQQJT

opposite

KKJT.AQQJ.AAAKQJJT.AAJT

Another 16 card fit? 2 in one game? I don't buy it. I think the correct...or at least very close to correct...expression for the probability at a partnership will have exactly 4 cards in a suit (combined) is

4 * 20c4 * 60c36 / 80c40

and that computes out to basically 0.7%. And if anything, this is a slight overestimate. There's some double counting, but it should be so incredibly unlikely that it wouldn't matter. So...twice? No.

BTW, I've also seen declarer get *hosed* by this, because the *defender* gets the wildly improbable suit. If one player has an 8 card suit, the probability anyone else has 8 should be *small*.

KKJJJTTT.AAQT.T.AAKKKQQ

opposite

AAAKQQJT.KJ.AAKKJTT.AQJ

This is a *perfect* fit. Declarer started trump aces...and dropped Ax. The club suit (the suit order is S-H-D-C, because this is PBN...look at the rank ordering) set up dummy's QUEENS to win the last 2 tricks, after an obvious crossruff.

And, sure...MAYBE this can happen once. But later on in that same game, there was this:

K.AAKJ.KQQJJTTT.AAKQQJT

opposite

KKJT.AQQJ.AAAKQJJT.AAJT

Another 16 card fit? 2 in one game? I don't buy it. I think the correct...or at least very close to correct...expression for the probability at a partnership will have exactly 4 cards in a suit (combined) is

4 * 20c4 * 60c36 / 80c40

and that computes out to basically 0.7%. And if anything, this is a slight overestimate. There's some double counting, but it should be so incredibly unlikely that it wouldn't matter. So...twice? No.

BTW, I've also seen declarer get *hosed* by this, because the *defender* gets the wildly improbable suit. If one player has an 8 card suit, the probability anyone else has 8 should be *small*.