Hi Ryan! You are correct - g4 gives me a roughly -1.4 evaluation, while f3 reads roughly -1.0 Just 1. f3 is widely accepted as the “worst” move in terms of difficulty to play, so used it for the title. Cheers! And sorry if a bit clickbait of a title :(
Hi Armaan! Just tested and this worked with LC0 for me - any engine file should work. I am not sure if playing at a specific node count is possible yet since the program is still in development. But you can adjust based on time/depth, and even choose to have the engine play completely random moves XD Cheers
@@joekempsey Hi, thank you. Developer if you see this feel free to add a node feature too Can you also check if you can change the weights for Leela. Since that's how you can make Leela stronger or weaker
Hi! The Wesley Bot was hanging on fairly well (~+1.00 position eval) until move 28, where Wesley played his king up to h6. Here, the eval jumped to +3.50, as it simply gave White a free move and a closer king to target. Better was immediate infiltration with the rook to d2 - delaying Rdd2 by even just a move was a costly mistake for the Wesley Bot xD
@JoeKempsey I find how low depth Stockfish 15 and high depth Stockfish disagree to be very fascinating! During the game, the evaluation bar was saying that Wesley was slightly better even after the blunder 28. Kh6. It did not even realize how bad Wesley's position was. The lopsided evaluation reminded me of a game magnus bot vs sf (on Gotham chess), where sf took advantage of a seemingly normal move by sacrificing its queen. Incredibly, the evaluation bar didn't understand the sacrifice until much later. Is there a way to approximate the difference in elo between the two engines? Thanks for responding!
@@tmpwow4282 I think it would be possible if you had the time to run a large sample size of games between the chess.com analysis engine and stockfish @ high depth. My guess is that the chess.com analysis engine is around 3200 ELO ish, but it’s hard to judge. Also, the analysis engine is currently limited to only stockfish 11 so it’s a bit worse than normal right now actually :O Stockfish @ high depth I would say is at least 3700+. Chess.com’s engines run on cloud software and at limited depth though, so unfortunately they end up performing a lot worse than Stockfish I run on computer hardware. With this in mind, the chess.com game review feature/ analysis can’t be 100% trusted 😅 Cheers!
Well you see that's an older version of Stockfish judging it's newer self. So that's not actually correct. Maybe a 100 depth latest Stockfish would rate it 99.5? maybe...
@@happygood18 Yeah unfortunately game review still ends up having a lot of hiccups analyzing top engine level chess. It’s an excellent indicator of how the match went overall and in case you miss any clear tactics, but just isn’t built for Stockfish 15.1 NNUE @ 50 Depth 😅😅
Stockfish really premoved the mate, the disrespect 😂
NO MERCYY
Isn’t g4 worse at high depth
Hi Ryan! You are correct - g4 gives me a roughly -1.4 evaluation, while f3 reads roughly -1.0
Just 1. f3 is widely accepted as the “worst” move in terms of difficulty to play, so used it for the title.
Cheers! And sorry if a bit clickbait of a title :(
@@joekempsey thanks
@@joekempsey its a bait title XD
@@happygood18 Yeah I’m probably gonna change it - thinking to put the WORST in quotes or something haha
@@joekempsey nah it's fine. will get you some extra attention :D
Can you only set up Stockfish, or other engines?
Like I would love to set up my maia bots to play at ~ ___ nodes
Hi Armaan!
Just tested and this worked with LC0 for me - any engine file should work.
I am not sure if playing at a specific node count is possible yet since the program is still in development. But you can adjust based on time/depth, and even choose to have the engine play completely random moves XD
Cheers
@@joekempsey Hi, thank you. Developer if you see this feel free to add a node feature too
Can you also check if you can change the weights for Leela. Since that's how you can make Leela stronger or weaker
dragon 3
@@armaanmehanger4154 Sure. I am redesigning the entire configuration for it to allow more flexibility.
@@shambhav9534 Do you have a discord I can contact you by?
Damn, that was a brutal finish! After e4 I swore White was busted, but it managed a shocking comeback.
Easy moneyy
Should've done the full bongcloud. Even then stockfish would've won tbh
Bongclouds will come don’t worry :D next video is gonna be really fun
When did Stockfish 15.1 begin evaluating its position as equal or better?
Hi! The Wesley Bot was hanging on fairly well (~+1.00 position eval) until move 28, where Wesley played his king up to h6. Here, the eval jumped to +3.50, as it simply gave White a free move and a closer king to target.
Better was immediate infiltration with the rook to d2 - delaying Rdd2 by even just a move was a costly mistake for the Wesley Bot xD
@JoeKempsey I find how low depth Stockfish 15 and high depth Stockfish disagree to be very fascinating! During the game, the evaluation bar was saying that Wesley was slightly better even after the blunder 28. Kh6. It did not even realize how bad Wesley's position was. The lopsided evaluation reminded me of a game magnus bot vs sf (on Gotham chess), where sf took advantage of a seemingly normal move by sacrificing its queen. Incredibly, the evaluation bar didn't understand the sacrifice until much later. Is there a way to approximate the difference in elo between the two engines?
Thanks for responding!
@@tmpwow4282 I think it would be possible if you had the time to run a large sample size of games between the chess.com analysis engine and stockfish @ high depth.
My guess is that the chess.com analysis engine is around 3200 ELO ish, but it’s hard to judge. Also, the analysis engine is currently limited to only stockfish 11 so it’s a bit worse than normal right now actually :O
Stockfish @ high depth I would say is at least 3700+.
Chess.com’s engines run on cloud software and at limited depth though, so unfortunately they end up performing a lot worse than Stockfish I run on computer hardware. With this in mind, the chess.com game review feature/ analysis can’t be 100% trusted 😅
Cheers!
Wow 😯 97.7
Well you see that's an older version of Stockfish judging it's newer self. So that's not actually correct. Maybe a 100 depth latest Stockfish would rate it 99.5? maybe...
Cheat codes were activated XD
@@happygood18 Yeah unfortunately game review still ends up having a lot of hiccups analyzing top engine level chess.
It’s an excellent indicator of how the match went overall and in case you miss any clear tactics, but just isn’t built for Stockfish 15.1 NNUE @ 50 Depth 😅😅
@@joekempsey lmao cheat codes
You Fox
🦊🦊