@@manuelgarrido5602 Leela did do the same kind of restraint on a rook in one game against Stockfish, and succeeded in capturing the rook. Stockfish didn't recognize the danger until it was too late.
Its not practical for a large search space engine to recognize the danger : it requires too much processing of each position. Leela, by using a much smaller search space can see the dangers obvious to humans.
@@Bobby-fj8mk The problem is the origin and evolution of Stockfish. Much of its strength comes from the size of its search space. Given the limited resources available, it can only execute a very simple evaluation of all positions. Your correct, it should have a 2nd evaluation run of the dominant lines, but that would demand a much more complex program, making it more difficult to determine the optimal 1st run. Its a matter of accepting weaknesses to allow a faster overall development.
@@michaeledwards2251 - yes - if you play against Stockfish as a human you will get beaten even though it does terrible things like giving itself doubled pawns - sometimes many of them or locked in pieces like that Rook. That is something only a computer - able to analyse billions of board positions - can get away with. Therefore - some computer solutions to chess problems can be ridiculous for a human player to employ.
He said at the beginning that the engines are not winning with this opening, wich is proof of it's substandard characteristics. If you still want to play garbage, no one is stopping you 👍
Great great game by Leela, as you said. I like your comparison between Stockfish and Leela. It does seem that Leela's chess is "elegant" in compared to the brute forced moves of Stockfish.
Hey Jozarov, would u please show us how Stockfish would continue after rook getting trapped? As Stockfish evaluate that position advantageous for black....do this man😢
Leela has always been really good at saccing in order to make the enemy pieces useless, great game
Chess: the only game where 'sickness' signifies health..
Hello! This is not the first game where Leela put the stockfish rook in jail. This aslo happened in the last final of tcec in one game.
It's not Stockfish, but Berserk.
@@MarcoMate87
To Stockfish also, I think...
@@MarcoMate87
May be I was wrong and not in TCEC
ua-cam.com/video/YUC96f29mjE/v-deo.html
@@manuelgarrido5602
Leela did do the same kind of restraint on a rook in one game against Stockfish, and succeeded in capturing the rook. Stockfish didn't recognize the danger until it was too late.
@@michaeledwards2251
Tx. Then, yes it looks like this is a patern understood by Leela and not Stockfish
This is Leela WDL btw
Berserk must have played with the contempt parameter at maximum to allow this white win.
Amazing game by leela , thank you for analysing
From 9:23 position, I played Stockfish against itself and it drew.
Leela spots Beserk a pawn...and wins
Black should never have allowed its Rook to be trapped like that.
Its not practical for a large search space engine to recognize the danger : it requires too much processing of each position. Leela, by using a much smaller search space can see the dangers obvious to humans.
@@michaeledwards2251 - it's faulty programming.
@@Bobby-fj8mk
The problem is the origin and evolution of Stockfish. Much of its strength comes from the size of its search space. Given the limited resources available, it can only execute a very simple evaluation of all positions.
Your correct, it should have a 2nd evaluation run of the dominant lines, but that would demand a much more complex program, making it more difficult to determine the optimal 1st run. Its a matter of accepting weaknesses to allow a faster overall development.
@@michaeledwards2251 - yes - if you play against Stockfish as a human you
will get beaten even though it does terrible things like giving itself doubled pawns -
sometimes many of them or locked in pieces like that Rook.
That is something only a computer - able to analyse billions of board positions - can get away with.
Therefore - some computer solutions to chess problems
can be ridiculous for a human player to employ.
@@Bobby-fj8mk
I remember Magnus Carlson remarking "Its like being crushed by an idiot".
❤❤❤❤❤ brutal, mesmerizing, genius. Cheers Josip
Incredible!
In many Leela vs Stockfish that is posted on UA-cam, I feel that Leela is always given weird/disadvantageous opening line..
They both Are given this positions
Helo. Beautiful chess game, best regards.
Brilliant!
I want to see a game like this..
Tell me how and why these chess players keep spewing “bad opening” it’s obviously bad if u play it bad
He said at the beginning that the engines are not winning with this opening, wich is proof of it's substandard characteristics. If you still want to play garbage, no one is stopping you 👍
I wonder what in Leela's programming causes it to trap pieces like that...
Alpha Zero would do the exact same thing...
Lc0 came from AlphaZero technical paper/report that was released by Deepmind.
It’s because it has learned heuristics from the game of Go, which is all about space and suffocation of the opponent
@@thomasmckane9334 It's the same neural network architecture, not literally the same neural network. Lc0 Neural network didn't see any Go games.
@@alpha007org
Because Leela is self learning, the positional evaluation is strongly based on danger and potential.
Great great game by Leela, as you said. I like your comparison between Stockfish and Leela. It does seem that Leela's chess is "elegant" in compared to the brute forced moves of Stockfish.
31... Kg6 ?! is controversial 31... Rd8 !! is probably better
Ledner Ports
Hey Jozarov, would u please show us how Stockfish would continue after rook getting trapped? As Stockfish evaluate that position advantageous for black....do this man😢
White obvius becouse the rook trapped and you just win the pawns and then the rook
#7