On 5/7/2008 8:01 AM, Matthew Johnson wrote:
> (sorry for the long post!)
>
> Being a poor student club we can't afford a duplimate machine but is
nice to
> have some of the benefits, like handouts and double dummy analysis etc.
It can
> also be nice to generate specific hands for certain events.
>
> To solve this, I invented a scheme for producing duplicated boards
without a
> duplimate machine. It also doesn't allow the people involved in the
process
> to know what the hands are so it can be done by the players themselves.
>
> A more complete description was presented at the 2008 International
Workshop on
> Security Protocols and the paper is available for download[0]. The
software
> used to run it is also available[1] along with a how-to do***ent for
people
> doing the duplication[2].
>
> I'm posting to rec.games.bridge to see what the wider bridge community
thinks
> about the system and what types of game it might be suitable for use in.
>
> A quick description of the system is as follows. The computer generates
random
> hands, or loads them from sims files, or whatever, and does DD analysis.
There
> are many programs which will do this some of which are free. It's
obviously
> easy to then print hand records. The duplicating program then generates
two
> pseudo-random permutations which individually reveal nothing about the
desired
> hand, but when applied in order, starting from a sorted deck of cards,
but when
> applied in order produce the desired hand.
>
> These permutations are given to two separate players (possibly on
different
> teams) who each do half of the process. The first player stacks up the
piles
> after their deal and pass them to the other player. The second player
puts the
> resulting hands in the boards.
>
> Permutations look like:
>
> 3121 2131 4431 4431 4242 ...
>
> and give the order to deal the cards into four piles.
>
> We've used it so far for our club nights, pretty successfully. It
doesn't take
> much longer than dealing the boards by hand (particularly if you shuffle
> properly, which many people don't) and most of the time gives the
correct
> result, although occasionally people screw up, usually resulting in a
single
> card transposed between hands. It also takes a similar length of time to
> running a set of boards through a duplimate machine.
>
> Given that the alternative is to have no computer-dealt hands or
analysis,
> having all but a few correct is still a big win for us. It's
particularly nice
> to be able to have some more interesting hands generated for our
Christmas and
> Easter parties.
>
> We have wondered about suggesting this for league matches and such,
where hand
> records would be nice, but have been unsure of what oppo would think.
Our
> players are happy to accept the security analysis we've done, but it
might look
> a bit suspect when suddenly presented with this to opponents. In reality
it
> should be treated the same as duplicated boards produced by a duplimate
> machine.
>
> So, does this sound interesting to anyone? Would you complain if your
opponents
> tried to use it for a match? The software is free to download (it's open
> source) and written in Java, so it should work on any system (although
I've
> only really used and designed it for Linux).
>
> Matt
>
> 0. http://www.matthew.ath.cx/publications/2008-JohnsonOwen-duplimate.pdf
> 1. http://www.matthew.ath.cx/projects/pescetti/
> 2. http://www.matthew.ath.cx/projects/pescetti/dealing-howto.pdf
>
It is certainly a reasonable method of player duplication, and solves
your problem of wanting computer-analyzed hands.
At tournaments (which for some reason won't get duplimate machines in
the US), they usually pass out hand records and then make sure the
boardset you dealt is the one you skip. This works fine if you have
enough tables for that, and student clubs rarely do.
At clubs, the usual approach is for the director to deal the hands out
himself, from dealing instructions, ahead of time. The dealing
instructions look basically the same as your solution, except that you
only need to do it once, because when you are dealing out 24 boards, you
really can't remember even a single card from all those 1s and 4s, even
if you end up having to play. A practiced director can easily do 24
boards in under an hour (even under a half hour), if the boards are
pre-sorted.
If neither of these methods is feasible for you, your solution seems
very practical.
--
Cheers,
Alan (San Jose, California, USA)
|
