[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

The Michael Depreli Bot Comparisons

Posted By: Tom Keith
Date: Tuesday, 15 February 2011, at 3:53 p.m.

In Response To: The Michael Depreli Bot Comparisons (Nack Ballard)

Thanks for the comments, Nack.

About the strength-vs-speed table, I admit I'm a sucker for pretty graphics. Xavier did a nice job of presenting this information alongside the strength numbers. And it is helpful to know the relative speed of the bots so you can compare them at similar settings.

But perhaps the bigger question is: Can a single set of numbers properly convey the speed information? The relative speed of Snowie is going to depend a lot on the number of cores in the processor you use for testing because Snowie is not optimized to use multiple cores. Is this unfair to Snowie? In some ways yes, in some ways no. It would be nice to have a more comprehensive study of relative bot speed, as Xavier suggests.

Also, offhand it seems to me that the bot chosen to do the arbitrating rollouts has an "unfair" advantage (Snowie in 2005, but more importantly GnuBG in 2010). I daren't venture a guess whether the effect is trivial or significant. (Perhaps the only fair way to do the study is to roll out all positions with all bots, though it would be highly impractical.)

One way to look into this some more would be to take a random sample of the positions that Snowie got wrong and roll them out in Snowie to see if any of the results are overturned (or if the size of the error changes significantly).

Messages In This Thread

 

Post Response

Your Name:
Your E-Mail Address:
Subject:
Message:

If necessary, enter your password below:

Password:

 

 

[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums is maintained by Stick with WebBBS 5.12.