[Pachi] Regression ?
lemonsqueeze at free.fr
Mon Apr 4 19:33:13 CEST 2016
Oh wow, looks like 32e02d4 is good, but ac859d8 is bad (!)
Either i missed something or something funny is going on ...
S GAMES WINRATE S.D. PAIRING
? 13 0.615 0.135 19-7.5-1-pachi_org-pachi_32e02d4
/ 19 0.211 0.094 19-7.5-1-pachi_org-pachi_ac859d8
On 04/03/2016 01:05 PM, Petr Baudis wrote:
> On Sun, Apr 03, 2016 at 12:48:26PM +0200, lemonsqueeze wrote:
>> Looks like we might have a regression:
>> I just played some games (19x19, 180k sims, large patterns, no dcnn) between
>> current master branch 83985a5 and 8e6758e (2016-01-11), i'm getting around
>> 27% winrate ...
>> Played 11 games so far + 15 between my tree and master branch with same
>> winrate, so about 26 games total. (my tree has some fixes not in master, but
>> none that i'd expect to make it 1 stone stronger !)
>> I'll see if i can find out more ...
> Thanks for the heads-up, that'd be great.
More information about the Pachi