[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Bug-gnubg] Aborted Rollout Analysis
From: |
Ned Cross |
Subject: |
[Bug-gnubg] Aborted Rollout Analysis |
Date: |
Tue, 24 Sep 2002 11:53:50 -0700 |
Here's the output I get from "copy" (in the annotation window) after a
rollout that I stopped after 64 trials:
1. Rollout 18/16 18/13 Eq.: +0.818
0.692 0.269 0.007 - 0.308 0.066 0.001 CL +0.666 CF +0.818
[0.006 0.007 0.004 - 0.006 0.003 0.000 CL 0.019 CF +0.049]
Full cubeful rollout with var.redn., with rotate rolls
432 games, seed 1153103104 using Mersenne Twister
When I stopped the rollout, GNUBG wrote on the rollout status bar "(Finished
64 trials)", but then saved the analysis as if the full 432 trials had been
completed.
The output seems to assume all rollouts will be performed to completion, but
sometimes they are stopped early due to time constraints, because the info
is "good enough" by that point, or because I'd prefer to start over with
different settings.
Better would be if the output contain the actual number of trials completed.
_Ned