06-30-2016, 10:16 PM

Noted elsewhere I was getting this all set up. Right now I've only got 1 game, 12 deals, but there's already a very suspicious statistic.

I counted the hand shapes for all the hands. I worked out the computation to know the frequency any shape should occur...see Pinochle Discussion sub-board.

So, with my 48 hand sample, I got a bunch of shapes. I took every shape where the probability of that shape was 2% or less. It turned out that there were 9 such. So, I defined a 'trial' as a hand being dealt with any of those shapes. The probability that such a hand would be dealt, is the sum of the 9 probabilities. In total, that sum was 10.3%. So, in 48 hands, the expected number of hands of any of those shapes, would be 5.

There were 11. To say this is an outlier is an understatement; it should happen less than 0.6% of the time.

I counted the hand shapes for all the hands. I worked out the computation to know the frequency any shape should occur...see Pinochle Discussion sub-board.

So, with my 48 hand sample, I got a bunch of shapes. I took every shape where the probability of that shape was 2% or less. It turned out that there were 9 such. So, I defined a 'trial' as a hand being dealt with any of those shapes. The probability that such a hand would be dealt, is the sum of the 9 probabilities. In total, that sum was 10.3%. So, in 48 hands, the expected number of hands of any of those shapes, would be 5.

There were 11. To say this is an outlier is an understatement; it should happen less than 0.6% of the time.