Poker-AI.org Poker AI and Botting Discussion Forum 2013-11-09T13:31:55+00:00 http://poker-ai.org/phpbb/feed.php?f=26&t=2614 2013-11-09T13:31:55+00:00 2013-11-09T13:31:55+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2614&p=5216#p5216 <![CDATA[Re: New Open Test Bed]]> bob101 wrote:
Nice find

Statistics: Posted by spears — Sat Nov 09, 2013 1:31 pm


]]>
2013-11-09T13:20:23+00:00 2013-11-09T13:20:23+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2614&p=5215#p5215 <![CDATA[Re: New Open Test Bed]]> In that case Server will be a simple php page(or some pages) at any hosting.
And no any problem like:
[strike]server must have statik IP[/strike]
[s]sevrer and its owner had to wait a long for anoter side configure his software[/s]
and more more more...

Statistics: Posted by nefton — Sat Nov 09, 2013 1:20 pm


]]>
2013-10-21T09:53:50+00:00 2013-10-21T09:53:50+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2614&p=5127#p5127 <![CDATA[Re: New Open Test Bed]]> http://www.cubeia.org/index.php/poker

bob

Statistics: Posted by bob101 — Mon Oct 21, 2013 9:53 am


]]>
2013-10-19T08:53:58+00:00 2013-10-19T08:53:58+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2614&p=5114#p5114 <![CDATA[Re: New Open Test Bed]]>
http://www.computerpokercompetition.org ... otocol.pdf

So I don't see issue to have java client...

By the way I think that there is a java client for ACPC in the FO2 sources... but for ACPC v1 and not v2...

Regards,

MrNice

Statistics: Posted by MrNice — Sat Oct 19, 2013 8:53 am


]]>
2013-10-18T22:14:35+00:00 2013-10-18T22:14:35+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2614&p=5112#p5112 <![CDATA[Re: New Open Test Bed]]> Statistics: Posted by cantina — Fri Oct 18, 2013 10:14 pm


]]>
2013-10-18T21:55:29+00:00 2013-10-18T21:55:29+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2614&p=5110#p5110 <![CDATA[Re: New Open Test Bed]]> Nasher wrote:

If you setup the ACPC server we can all compete.

I plan to do that, but is ACPC faster than OTB for simulations? Do you know if the current version of ACPC is compatible with the Java version protocol?

Statistics: Posted by corintio — Fri Oct 18, 2013 9:55 pm


]]>
2013-10-18T21:52:38+00:00 2013-10-18T21:52:38+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2614&p=5109#p5109 <![CDATA[Re: New Open Test Bed]]> shalako wrote:

Quote:
Or I'm loosing my time with it and there are better ways to test our bots (at least the Java based ones) that I'm not aware of?


I like to test live in play money games as you get to see a wide range of player types. It is slow but it works. I watch it play and take notes of issues.

I also like to test it against itself vs random boards and situations to ensure its actions are polarized. I use this for leak testing quite extensively. An example is running thru 100k hands preflop and see what its PFR, 3 Bet, OOP call etc percentages are. These percentages change based on the player type but its important that the baseline strategy is near GTO as possible. I do this for flop games as well to see how it handles a turn barrel, river check raise, how often it c-bets etc.

Yes, there is nothing like testing it in the "production" environment :) Unfortunately I'm not there yet, still working on the AI.

But there are cases where you need to run thousands (millions?) of hands (for training ANNs, for gathering statistics on PT, etc...) so I think that for these situations OTB is still a valid option, right?

Well, if someone is interested, I just uploaded my version to GitHub: https://github.com/corintio/opentestbed It has initial support for Tournaments

Any comments, patches, issue reports are welcome!

Statistics: Posted by corintio — Fri Oct 18, 2013 9:52 pm


]]>
2013-10-18T21:23:12+00:00 2013-10-18T21:23:12+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2614&p=5108#p5108 <![CDATA[Re: New Open Test Bed]]> Statistics: Posted by cantina — Fri Oct 18, 2013 9:23 pm


]]>
2013-10-18T20:31:56+00:00 2013-10-18T20:31:56+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2614&p=5107#p5107 <![CDATA[Re: New Open Test Bed]]> Quote:

Or I'm loosing my time with it and there are better ways to test our bots (at least the Java based ones) that I'm not aware of?


I like to test live in play money games as you get to see a wide range of player types. It is slow but it works. I watch it play and take notes of issues.

I also like to test it against itself vs random boards and situations to ensure its actions are polarized. I use this for leak testing quite extensively. An example is running thru 100k hands preflop and see what its PFR, 3 Bet, OOP call etc percentages are. These percentages change based on the player type but its important that the baseline strategy is near GTO as possible. I do this for flop games as well to see how it handles a turn barrel, river check raise, how often it c-bets etc.

Statistics: Posted by shalako — Fri Oct 18, 2013 8:31 pm


]]>
2013-10-18T21:53:36+00:00 2013-10-18T13:29:23+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2614&p=5105#p5105 <![CDATA[New Open Test Bed]]>
I'm making some improvements to the old Open Test Bed project (Tournaments for example), and would like to know if people here still uses it. Does it make sense to take over the project (I've being talking with the original author about that), so we can continue to push it forward?

Or I'm loosing my time with it and there are better ways to test our bots (at least the Java based ones) that I'm not aware of?

UPDATED: Here's the full source: https://github.com/corintio/opentestbed

-Corintio

Statistics: Posted by corintio — Fri Oct 18, 2013 1:29 pm


]]>