[gnugo-devel] twogtp breakin results

[bump]

Here are the results of 115 gnugo/gnugo games. The conditions of the match:

twogtp --black "gnugo --without-break-in --quiet --mode gtp --resign-allowed" 
--white "gnugo --with-break-in --quiet --mode gtp --resign-allowed" --games 100 
--sgffile match_break_.sgf

There's more than 100 games because the match was stopped and restarted
to run the regressions once in the middle.

I note that allowing resignations slightly complicates the problem
of deducing statistical conclusions from these results, and it is
also necessary to have a similar run with the colors reversed
before on can draw any conclusions about whether the breakin code
seems to give an advantage that is statistically measurable from
twogtp results.

Dan

B+R
B+R
B+R
B+R
B+R
B+R
B+R
B+R
B+R
B+R
B+R
B+R
B+41.5
B+38.5
B+34.5
B+28.5
B+27.5
B+27.5
B+25.5
B+25.5
B+25.5
B+22.5
B+22.5
B+21.5
B+19.5
B+18.5
B+17.5
B+16.5
B+15.5
B+13.5
B+12.5
B+12.5
B+12.5
B+12.5
B+11.5
B+10.5
B+9.5
B+9.5
B+9.5
B+8.5
B+8.5
B+8.5
B+7.5
B+6.5
B+6.5
B+5.5
B+3.5
B+3.5
B+3.5
B+3.5
B+2.5
B+2.5
B+2.5
B+2.5
B+1.5
B+0.5
B+0.5
B+0.5
B+0.5
W+0.5
W+0.5
W+1.5
W+1.5
W+1.5
W+2.5
W+3.5
W+3.5
W+4.5
W+5.5
W+5.5
W+5.5
W+5.5
W+7.5
W+7.5
W+7.5
W+8.5
W+8.5
W+9.5
W+10.5
W+10.5
W+11.5
W+12.5
W+13.5
W+13.5
W+13.5
W+13.5
W+13.5
W+14.5
W+15.5
W+15.5
W+15.5
W+19.5
W+19.5
W+22.5
W+22.5
W+26.5
W+27.5
W+29.5
W+29.5
W+30.5
W+30.5
W+31.5
W+33.5
W+35.5
W+38.5
W+38.5
W+40.5
W+R
W+R
W+R
W+R
W+R
W+R
W+R
W+R