TalkChess.com

I have long held a theory in my own personal over-the-board human chess which I haven't seen in a any chess book: that there is an advantage of having two knights side by side, especially if near the opponent's king. Knights often make each other redundant, but when they are side by side, they don'...

The same at g4 h4!! at move 34 and knights in h4 h5 at move 20 [Event "Lc019rc2Sf8Initial25plus10LeR1_i7_4790S"] [Site "Ros_4790S"] [Date "2018.11.07"] [Round "28"] [White "Stockfish8"] [Black "Lc019rc2_11261"] [Result "*"] [BlackElo "2000"] [Time "09:42:30"] [WhiteElo "2000"] [TimeControl "1500+10"...

I have detected a curious maneuver of knights in g5 and h5 by lc0 in two games of a test that I am doing. In both cases the maneuver leds to a winning attack. The test compares lc0 with AlphaZero against Stockfish 8 without opening book. Lc0 version 0.19 rc2 with net 11261 on Asus Geforce GTX 1060 3...

Leela Chess Zero has found very original ideas in the opening and is positionally superior to alphabeta engines, which makes it ideal for preparing openings. However, its tactical weakness makes it necessary to carefully check these ideas with a program like Stockfish.

Damir wrote: ↑

Thu Nov 08, 2018 11:47 am

rc3 is already released.

https://github.com/LeelaChessZero/lc0/releases

Thanks, I will finish my tests before upgrade to rc3.

... Thanks, seems to suggest Lc0 can setup traps effectively against strong engines in fast time controls. Well it is not so clear. Replaying the games at 25 min plus 10 sec the results are: Sf9 +3 -2 =5 lc0 +2 -3 =5 victory for Stockfish9 with 55% In the game with the poissoned knight (24.Qxg6), S...

... If you want to prove how better lc019rc2 is than lc0181 why do not you play lc019rc2 and lco181 against each other? If you want to demonstrate the power of lc0 against Stockfish you ought to use opening book with at least two ply depth because without opening book Stockfish are obviously in a d...

When lc0 starts beating latest Stockfish comprehensively, like AlphZero did, let me know. Till then I'm done with computer chess using alpha-beta engines. Well, this thread is about improvements in fast games in lc0 0.19rc2 version and not about Stockfish 8, but I am doing another test between lc0 ...

... There's also another strange Lc0 Sicilian endgame where an advanced c-passer + N defeats the R. Looks like some eval tweaking and fine tuning for SF9 endgame eval function is needed. Quality of the above games doesn't look like a 1min+1sec for humans, more like Finals in a Candidates match mayb...

A curious position was reached in the following game, where the White King is trapped in h1 from move 26 until the end. [Event "Lc019rc2_11261Sf9Initial1plus1_2CPU1024"] [Site "Ros980X"] [Date "2018.11.05"] [Round "24"] [White "stockfish_9_x2"] [Black "lc019rc2_11261"] [Result "*"] [BlackElo "2000"]...