>I think I've mentioned the error in game_std.h but the bug in
>eval_low8.c makes me think nobody has used this lib to do any high/low
>split investigation.

Thanks for the bug report; will fix. Will eventually get a public CVS
server running so you can do that yourself...

>Brian -- I believe at one point you were working on a o/8 evaluator.
>I know have an o8cmpn.c that does it by brute force. If you are not
>going to have a reasonable evaulator done in the near future I'd like
>to post this source.

I'm pretty sure I've got working low and low8 evaluators in my sadly
neglected library.

My slighly-optimized plan for Omaha evaluation was this: Let H be the
(four) cards in your and C be the common cards. Use the evaluator to
evaluate M = eval(C union H), which would give you the best possible hand
you could make if you broke the Omaha rules. Then do a brute-force
enumeration of the 2-card subsets of H until you either get M (in which
case you know you're done) or you exhaust the possibilities.

I suspect this will likely yield slightly less than 2x factor of speedup
over the brute force approach.



--
Brian Goetz
Quiotix Corporation
brian@... Tel: 650-843-1300 Fax: 650-324-8032

http://www.quiotix.com