Re: [Bug-gnubg] Error rate of predefined settings?

[boomslang]

I cannot give you Snowie error rates for the several settings, but I can give you an estimated FIBS rating for various noise settings. These results come from a bot I ran in 2006; I let it play with different settings for noise. The following table are maximum likelihood estimates and 95% CI's based on 67000 matches:
                                                                                        
                              
Setting       plies  noise*1000  FIBS    95% CI   
                                 ------------------
supremo         2        0       2106 (2075, 2138)
expert          0        0       2035 (2011, 2061)
 ??             0        7       1958 (1909, 2008)
advanced        0       15       1825 (1776, 1874)
 ??             0       22       1643 (1594, 1691)
 ??             0       31       1468 (1413, 1522)
intermediate    0       40       1220 (1169, 1270)
casual player   0       50        940  (888, 992)
beginner        0       60        714  (664, 764)

Looking at the table, I would say that the noise settings for intermediate, casual player and beginner are set too high.

A few remarks: (1) The bot played only 1 ptrs. (2) Results are based on a 2005 version of GNU Backgammon. Nets have recently improved. (3) Due to droppers the estimates are prolly 5-10 biased downwards.

Gr, boomslang

On Friday, 28 February 2014, 17:55, Timothy Y. Chow <address@hidden> wrote:

In the GNU documentation it says the following about predefined settings:

Beginner: This setting uses no lookahead and adds up to 0.060 noise to the
evaluation. With this setting GNU Backgammon will evaluate like a
beginner.

Casual play: This setting uses no lookahead and adds up to 0.050 noise to
the evaluation. With this setting GNU Backgammon will evaluate a bit
better than the beginner setting but not much.

Intermediate: This setting uses no lookahead and adds up to 0.030 noise to
each evaluation. It still plays an intermediate game.

Advanced: This setting uses no lookahead and adds up to 0.015 noise to
each evaluation. This setting plays a good game.

My question is, approximately what Snowie error rate do these settings
translate into?  Has anyone tested this out?

Tim

_______________________________________________
Bug-gnubg mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/bug-gnubg