Page 1 of 1

PV lines sorting Bug

PostPosted: 20 Oct 2015, 13:32
by F. Bluemers
Hello HGM,
I accidenlty stumbled yesterday on the line sorting bug again,while debugging something in Dirty.
The bug happens on sorting matescores only.
I guess the cause is the (10000+moves_to_mate) convention,
which gives longer matelines a higher score,but i did not check the source.

Code: Select all
6r1/1p2kp2/p1b4p/n3p3/2P2p2/8/3r2P1/5BK1 w - - 4 33
 17       #4    131.1M   0:56.96   (c5) Rg5 Kh2 Rd1 Kh3 Rxf1 Kh2 Rh5#
 17       #4    100.6M   0:42.71   (c5) Rg5 Kh2 Rd1 Kh3 Rxf1 Kh2 Rh5#
 16       #4    30.4M     0:13.74   (c5) Rg5 Kh2 Rd1 Kh3 Rxf1 Kh2 Rh5#
 16       #4    13.4M     0:06.26   (c5) Rg5 Kh2 Rd1 Kh3 Rxf1 Kh2 Rh5#
 15       #4    12.1M     0:05.65   (c5) Rg5 Kh2 Rd1 Kh3 Rxf1 Kh2 Rh5#
 15       #4    9.15M     0:04.17   (c5) Rg5 Kh2 Rd1 Kh3 Rxf1 Kh2 Rh5#
 14       #5    3.20M     0:01.47   (c5) Bxg2 Bxg2 Rdxg2+ Kf1 Rc2 c6 bxc6 Ke1 Rg1#
 14       #4    8.65M     0:03.92   (c5) Rg5 Kh2 Rd1 Kh3 Rxf1 Kh2 Rh5#
 14       #4    6.87M     0:03.07   (c5) Rg5 Kh2 Rd1 Kh3 Rxf1 Kh2 Rh5#

Best
Fonzy

Re: PV lines sorting Bug

PostPosted: 20 Oct 2015, 13:55
by H.G.Muller
Yegh. Indeed I would have to use the reverse comparison for mate scores there.