When I downloaded Stickfish 17 I immediately setup an engine tournament against SF 16.1 and the result instantly told me that they got something wrong with SF17 since it lost the tournament against its previous version..
More like the other engines are improving faster than it, we gotta give the other developers credit. Stockfish 17 is stronger than 16, but not by enough and now other engines are catching up. Despite that stockfish is still in 2nd place
Wow what a game. End games are so fascinating arent they?! Its so strange to see all this stuff right in front of you but not be able to solve “simple” problems. It feels like a magic trick, but theres no trick. Everything you need is right on the board and yet its so difficult and even impossible to see it! How these “machines” manage to make everything move together as one harmonious body is absolutely transcendental, just like Liszt’s etudes.
Leela is learning from the best counterpart!! That tells you, humans might not need to build stronger AI in the future as these titans of chess engines are gradually scouting on each other's tactical ideas everyday without the artificer's knowkedge(human)
All strong engines choose 16. d6 (@5:52) almost instantly. I'm not denying that it's a nice move, but it's a totally obvious one for any modern engine (Stockfish 8 finds it in 2 seconds).
likely. By all measures SF17 is weaker than SF16. Also title is slight clickbait. LC0 is momentarily half a point ahead of SF17 half way thru a tournament. This is testament to how dominant SF has been over LC0 for the last 10 tournaments. The slightest hint of a possible hiccup in SF total dominance sends an earthquake around the whole computer chess media.
@@plasmaastronautNo, Stockfish 17 is +40 elo stronger than Stockfish 16, and the tournament you said was “halfway through” had only 2 games played between Stockfish and Leela (Stockfish won 1, Leela won 1), the rest is swiss pairing luck
The reason you think that stockfish is weak is because Jozarov is covering games where lc0 won. In reality Stockfish still dominates Leela (and any other engine) in tournaments
@@许玄清 +40. I'll believe it when its reported on ccrl. Sounds like the elo gain per version 10 years ago. These days an engine is lucky to gain +4 elo per version.
This is a very interesting game, and I love your analysis and comments, thanks. However, and this is not intended as a criticism, I would say that after the opening (8. f3...), Stockfish 17 evaluates that Leela is already +1.00 up, and so is it a surprise that Leela eventually wins, against a weaker engine? I'd be intrerested to see the "return" game, same opening but with Leela as black, to see if it can also win that game.
Great game by the second-best engine in active tourneys! (AlphaZero is more for exhibition matches, but is considered around Stockfish's level or even better, depending on how many resources are allocated to it!) My theory on some of Stockfish 17.0's strange failures is that it's so aggressive in pushing for the win that it opens itself up for defeat as well. By attacking, it makes itself vulnerable. Perhaps 17.1 will address this?
Stockfish is in second place very early in the match, having played the tougher engines (based on the current standings) and has beaten Leela in their only match so far. So I thinks it’s misleading on you part to say Leela is taking over. Perhaps you wrote the title to create a clickbait video. The Fish will be in first when it’s over.
A fascinating endgame. And a wonderful dynamic win for Leela. Thanks Josip.
Thanks Jozarov - that is such an unusual game.
The White King is so exposed and Black has 3 wonderful connected pawns.
You'd think White was lost.
Leela stays the strategic&positional genius! Great game! I always learn lots from your great analysing commentary! Thanks Josip!
SF17 growing pains are real. Lc0 though has always been better in longer time formats so this doesn't fully surprise me
When I downloaded Stickfish 17 I immediately setup an engine tournament against SF 16.1 and the result instantly told me that they got something wrong with SF17 since it lost the tournament against its previous version..
Stockfish 17 looks like Stockfish 8
Lol. Don't disagree...Tbf... they made SF play very hard openings against strong engines to start the Swiss.
More like the other engines are improving faster than it, we gotta give the other developers credit. Stockfish 17 is stronger than 16, but not by enough and now other engines are catching up. Despite that stockfish is still in 2nd place
What does this video have to do with Stockfish?
@@dinofx35 If you would have subbed him than you know
Great Game Leela that was quite a battle nice offensive game Thanks Jozarov
Wow what a game. End games are so fascinating arent they?! Its so strange to see all this stuff right in front of you but not be able to solve “simple” problems. It feels like a magic trick, but theres no trick. Everything you need is right on the board and yet its so difficult and even impossible to see it! How these “machines” manage to make everything move together as one harmonious body is absolutely transcendental, just like Liszt’s etudes.
LEE-LA LEE-LA LEE-LA
The entire endgame I felt white was losing.
White up a piece the Bishop and it’s passed pawn making progress tight game for awhile though
Leela is learning from the best counterpart!! That tells you, humans might not need to build stronger AI in the future as these titans of chess engines are gradually scouting on each other's tactical ideas everyday without the artificer's knowkedge(human)
True, Lc0 got some craazy positional knockouts against Stockfish in last few years, the games are just mesmerizing
stockfish will come back 😮
All strong engines choose 16. d6 (@5:52) almost instantly. I'm not denying that it's a nice move, but it's a totally obvious one for any modern engine (Stockfish 8 finds it in 2 seconds).
Its weard because my sf 17 dont play like the one in the tcec even in 100+ depth he Plays different moves and go for litteraly different plans
What's your system specs, and what is your stockfish settings?
Maybe this new S17 has some bugs that the 16.1 didn't have? It playing strenght seems inconsistent at times.
likely. By all measures SF17 is weaker than SF16. Also title is slight clickbait. LC0 is momentarily half a point ahead of SF17 half way thru a tournament. This is testament to how dominant SF has been over LC0 for the last 10 tournaments. The slightest hint of a possible hiccup in SF total dominance sends an earthquake around the whole computer chess media.
@@plasmaastronautNo, Stockfish 17 is +40 elo stronger than Stockfish 16, and the tournament you said was “halfway through” had only 2 games played between Stockfish and Leela (Stockfish won 1, Leela won 1), the rest is swiss pairing luck
The reason you think that stockfish is weak is because Jozarov is covering games where lc0 won. In reality Stockfish still dominates Leela (and any other engine) in tournaments
@@许玄清 +40. I'll believe it when its reported on ccrl. Sounds like the elo gain per version 10 years ago. These days an engine is lucky to gain +4 elo per version.
I would of liked to see 16.1 in the same comp for a good comparison!!!
This is a very interesting game, and I love your analysis and comments, thanks.
However, and this is not intended as a criticism, I would say that after the opening (8. f3...), Stockfish 17 evaluates that Leela is already +1.00 up, and so is it a surprise that Leela eventually wins, against a weaker engine?
I'd be intrerested to see the "return" game, same opening but with Leela as black, to see if it can also win that game.
The reversed game was a draw
AFAIK engines really don't like/underestimate the caro kann
Great game by the second-best engine in active tourneys! (AlphaZero is more for exhibition matches, but is considered around Stockfish's level or even better, depending on how many resources are allocated to it!) My theory on some of Stockfish 17.0's strange failures is that it's so aggressive in pushing for the win that it opens itself up for defeat as well. By attacking, it makes itself vulnerable. Perhaps 17.1 will address this?
Alpha Zero games are much more entertaining than stochfish tho.
@@mithunchakravarthi8423 I like them both!
I liked the challenge question at 15:25 !
The reason this Fish underperforming is that the hand crafted evalution is off S17 is now fully self eval.
Stockfish is in second place very early in the match, having played the tougher engines (based on the current standings) and has beaten Leela in their only match so far. So I thinks it’s misleading on you part to say Leela is taking over. Perhaps you wrote the title to create a clickbait video.
The Fish will be in first when it’s over.
U need to set free of the fish so no stak loss anymore
isnt that just immediate checkmate? QxC7#