Poker-AI.org Poker AI and Botting Discussion Forum 2013-08-06T23:32:44+00:00 http://poker-ai.org/phpbb/feed.php?f=24&t=2524 2013-08-06T23:32:44+00:00 2013-08-06T23:32:44+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2524&p=4621#p4621 <![CDATA[Re: Optimize ICM calculation]]> Statistics: Posted by MerlinsBeard — Tue Aug 06, 2013 11:32 pm


]]>
2013-07-11T15:23:17+00:00 2013-07-11T15:23:17+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2524&p=4420#p4420 <![CDATA[Re: Optimize ICM calculation]]> Statistics: Posted by MrBrown — Thu Jul 11, 2013 3:23 pm


]]>
2013-07-10T12:45:39+00:00 2013-07-10T12:45:39+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2524&p=4408#p4408 <![CDATA[Optimize ICM calculation]]>
Here is the problem. In my tourneys payzone is top 9. Complete ICM computations for 9 places is heavy and it takes ~20 seconds to calculate NE ranges. This timing is unacceptable for real botting, so I pretend that payzone if top5 and it works much faster. But computed ranges are not true ranges. In most cases it does not matter but I suppose it matters on the bubble and my ROI would be higher if I could compute ICM values for 9 places faster.

Any Ideas to speed-up ICM?

I tried the approach from http://forumserver.twoplustwo.com/15/po ... s-1098489/ ,but for top9 payzone it takes the same time.

If it helps, about half of the stacks are equal to each other, because they are to simulate players on other tables.

Statistics: Posted by concretemixer — Wed Jul 10, 2013 12:45 pm


]]>