Apparently there are 3 general levels of random computer bridge dealing.

[Douglas]

First is what passes for everything except Bridge Analyser today.
The best of these is probably BigDeal, more because it is
generally accepted now. But it is thin gruel compared to what is
possible, now and in the future.

Bridge Analyser has a good enough random number source to provide
unbiased random HANDS.

However when these same hands are evaluated as SUIT lengths,
they become ordinary pseudo-random in bias.

And finally, when evaluated as suit lengths in each separate
bridge hand, they display even more bias.

I recently evaluated quantum sourced "true" random numbers. They
were laughably biased, and volatile. z_values of 3.6 and 6.3.

However, the numbers from random.org (which are derived from
atmospheric noise) where quickly unbiased when dealing suit
lengths in each separate bridge hand.

So, I suggest that bridge dealing can be improved readily in
the immediate future with the adoption of cipher algorithms
(with their inherent security) with huge periods.

And when someone figures out a way to use an infinite source
of unbiased random numbers such as random.org in their bridge
dealing program, a "fair digital bridge deck" will become
possible. All deals, hands, and suit lengths possible, and
equally probable, in each and every computer deal.

Douglas

[p***@gmail.com]

Post by Douglas
And when someone figures out a way to use an infinite source
of unbiased random numbers such as random.org in their bridge
dealing program, a "fair digital bridge deck" will become
possible. All deals, hands, and suit lengths possible, and
equally probable, in each and every computer deal.
Douglas

Random.org already has a random card shuffle demo, which could be adopted for bridge shuffles. https://www.random.org/playing-cards/