Page 2 of 2

Re: Lc0 Vs Stockfish Match.

Posted: Sat Feb 23, 2019 4:50 am
by M ANSARI
Yes the match has been extremely competitive somehow. I certainly didn't expect that as in my tourneys it was very obvious that there were some serious weaknesses in Lc0 that were still not patched, and SF is an extremely well honed, mature and finely tuned world's strongest chess engine. This still could go either way, but that would depend mostly on the type of openings chosen in the last few games. SF already has 9 losses and that could have easily been 12 or as high as 15!! This match basically confirms what we have all noticed here on this forum, that NN is the way to push things forward in chess engine strength and Alpha Zero was really a game changer. TCEC is not really about checking which is the strongest engine as much as it is about entertainment. The positions chosen from book are purposefully chosen to get positions that will be entertaining to watch and provide exciting decisive games. I have to admit in most games this did happen and I have watched several exciting games live. For Lc0 this really shows that using NN is really a way to break through what seems to be a slow down in ELO strength. However Lc0 is still not an engine I would use for analysis on its own ... maybe in the version that plays TCEC 15 this will change. I think that is where the real battle will be.

Re: Lc0 Vs Stockfish Match.

Posted: Sat Feb 23, 2019 1:23 pm
by jp
M ANSARI wrote:
Sat Feb 23, 2019 4:50 am
The positions chosen from book are purposefully chosen to get positions that will be entertaining to watch and provide exciting decisive games. I have to admit in most games this did happen and I have watched several exciting games live.
I don't think so. The current game is 3 moves out of book & SF shows 0.00/100% D.

Re: Lc0 Vs Stockfish Match.

Posted: Sat Feb 23, 2019 11:23 pm
by sincerelybrent
Do they share the .pgn of the match? I'm not able to find it. Thanks,