[Cryptography] Real-world crypto/PRNG problem: Bridge

[Bill Frantz]

On 8/23/16 at 1:37 AM, leichter at lrw.com (Jerry Leichter) wrote:

>Given the use case, shuffling a deck of cards and typing in the 
>sequence of cards should be sufficient...

I seem to remember reading about an issue that turned up when 
duplicate bridge tournaments moved from hand shuffling to 
machine shuffling many years ago. People noticed that there were 
many more distributional hands with the machine shuffling. 
Analysis indicated that the typical hand shuffling wasn't really 
mixing the cards enough, so the ordering left from the last hand 
presisted into the new hand. Machine shuffling was doing a 
better job of mixing.

Typically in rubber bridge, the cards get shuffled 3 or 4 times 
before being dealt again. This turned out to be far too few 
mixings. I think I remember recommendations for 7 or so mixings, 
but I'm not sure even this level is enough.

Moral: Random is hard. :-)

Cheers - Bill

-------------------------------------------------------------------------
Bill Frantz        | Re: Hardware Management Modes: | Periwinkle
(408)356-8506      | If there's a mode, there's a   | 16345 
Englewood Ave
www.pwpconsult.com | failure mode. - Jerry Leichter | Los Gatos, 
CA 95032