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

BGonline.org Forums

Using two bots instead of a longer rollout: effect on confidence in result

Posted By: leobueno
Date: Friday, 21 December 2012, at 1:58 p.m.

I anticipate that most folk don't have super-duper fast PCs, so there are some positions whose rollout take a very long time to produce acceptable statistical confidence on our low-speed boxes. Of course there is a tradeoff between number of trials and confidence, but with slow PCs, most of us will never be able to reach a suitable balance. Thus we are left to the generosity of those of you who find a posted position sufficiently interesting to rollout on their speedy boxes.

I am wondering about using two bots simulataneously to analyze complex position that would take just too long to rollout, to say JSD > 3.5 on GNUbg or to 99.5% in XG.

For example, I am currently looking at Magriel's BG position 12-1 on page 150. Note this position was examined by Bagai in CLASSIC BG REVISITED, but I am rolling it with no Jacoby (I hate Jacoby), Beaver.

XG 3-ply RO says that after 886 games 8/7 8/5 has an 89.9 chance of being best. GNUbg 2-ply RO says that after 4450 games 8/5 8/7 is best with JSD of the second play at 1.803.

My understanding is that neither result gives us sufficient confidence that the top play is indeed the best play. However, what can we infer from the fact that *both* bots agree it's the best play? Is using two bots that agree on the best play a suitable shortcut?

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.