View Full Version : Verification of Computer Results....

09-26-2003, 02:36 PM
I just finished writing a java program that will analyze three two card player hands and I would like to validate the accuracy of my code. Here is what I would like to know:

I analyzed Q /images/graemlins/spade.gif 7 /images/graemlins/diamond.gif vs. J /images/graemlins/club.gif 9 /images/graemlins/club.gif vs. 10 /images/graemlins/heart.gif 6 /images/graemlins/heart.gif.

My results:
There are 1,370,754 different possible boards.
Q /images/graemlins/spade.gif 7 /images/graemlins/diamond.gif will win: 455,340 times,
J /images/graemlins/club.gif 9 /images/graemlins/club.gif will win: 508,466 times,
And 10 /images/graemlins/heart.gif 6 /images/graemlins/heart.gif will win: 400,182 times,
And there will be 6,766 ties amongst all three players.

The reason I am posting this is because I figured that Q /images/graemlins/spade.gif 7 /images/graemlins/diamond.gif had a slight advantage but my numbers are telling me something different. Is there anyone that can validate or invalidate my results?


La Brujita
09-26-2003, 02:53 PM
twodimes.net is a hand analyzer many people use. It confirms J-9 is the favorite.

09-26-2003, 02:57 PM

<font class="small">Code:</font><hr /><pre>
pokenum -h qs 7d - jc 9c - th 6h
Holdem Hi: 1370754 enumerated boards
cards win %win lose %lose tie %tie EV
Qs 7d 455340 33.22 908648 66.29 6766 0.49 0.334
Jc 9c 508466 37.09 855522 62.41 6766 0.49 0.373
Th 6h 400182 29.19 963806 70.31 6766 0.49 0.294
</pre><hr />

09-26-2003, 05:23 PM
Cool. Would you mind showing us your code or explaining the algorithms behind it?

09-26-2003, 10:13 PM
Two-dimes is not Bruce's code, but a lot of that source code is available at sourceforge (http://sourceforge.net/projects/pokersource/) or you can get the source for holdem showdown at brecware (http://www.brecware.com/Software/software.html) .


09-27-2003, 01:33 AM
Wow, thanks Bozeman. I'll check that out.