[gnugo-devel] how influence influences owl

[Arend Bayer]

Whew, --experimental-influences currently causes 70 PASSes and 104 FAILS. Btw,
after Dan had mentioned that regressions take only 12 hours on his
Celeron-433, while they took ~24 h (that was including --owl-threats) 
on my K6-II 400, I realized that I might actually be so kind to my gcc
to tell him about the processor. He returned the favor of adding the flag
-m486 in engine/Makefile by producing a huge performance gain, appartently
more than 1.5. Maybe we should add a hint in the INSTALL file that one
should care about the $CFLAGS environment variable? At the moment
this only mentioned in the ./configure --help output.

Among these FAILS are some owl failures which I do not understand. Can
someone more owl-knowledgeable explain me how the influence function affects
the owl code, and e.g. change the moves tried? That would be very helpful.

Arend

./regress.sh . owl.tst
109 unexpected FAIL: Correct '1 (B5|C4)', got '1 B4'
118 unexpected PASS!
170 unexpected FAIL: Correct '0', got '1 B15'
194 unexpected FAIL: Correct '1 G18', got '0'
228 unexpected FAIL: Correct '0', got '1 D5'
./regress.sh . owl_rot.tst
118 unexpected PASS!