Hi everyone, I would like to know what happened to stockfish here. I wanted to know if my bishop sacrifice was good, but the engine gives a very weird evaluation, with big peaks and drops for no reason. Has anyone got an explanation please ? How can I know if my sacrifice was good and if I had a better continuation ? Thanks to the people who will help me to understand.
http://fr.lichess.org/CDRWHfNF/black#0Problem is the depth Stockfish is calculating. I let the engine run for some time and it gives around -1.8 until around depth 24 and goes down to 0 from 25+ (with Kg1 or g3 as playable moves).
So stockfish found a refutation which secures a draw for white, a 25 move long refutation. So from a human standpoint I would count it as good (and fun). Actually a well played game from both of you.
Only question: Why did you guys agree to a draw? There is so much to play for in this position.
No idea why the lichess-stockfish evaluation goes up to +1.7 again after Ng4 and Qg6 though. It doesn't change on my computer. Weird stuff.
(BTW, All the moves played after Bxh2 were basically the top moves)
The game was drawn because I told him that after fxe3 I didn't see anything else to play besides the perpetual check (Qxg3+ Kh1 Qh3+ Kg1 Qg3+ Kh1...). I don't see any way to continue my attack and I'm down a lot of material, so the perpetual was the best thing I could get as far as I can see.
And yes I was quite happy about this game, especially since I spent like 3 hours analyzing my bishop sacrifice lol, but I have to admit that I was sure to checkmate him. He defended extremely well and I wouldn't have played that if I saw that he could get a draw.
So basically stockfish (or at least the way lichess uses it) is not strong enough to give a good evaluation of this position ? It's strange because this game doesn't look THAT complex for an engine, since most of the moves are forced to avoid a quick checkmate
If you want to call it like that, yes, lichess Stockfish isn't strong enough.
Keep in mind though that this stockfish here has to analyze 1000s of games so for each game it probably only has a few seconds or so, which then again divide into a fraction of a second only per move. When calculating only a fraction of a second it's understandable that the output may not be perfect. ;)
I think there is sufficient defense for white, if it can locate the knight in b3 on d4 and the other knight in d2, deffending well f3. In addition to advancing the pawn to g3. If there are changes of knights and bishops, the white lady should not lose sight of the point f3 to prevent check mate in g2. In my opinion is why Stokfish consider it wrong. Regards
According to my local Stockfish, your sacrifice 14 ... Bxh2 was not that good (leads to a draw). 14. ... Bf3 was better with an evaluation of around +2 (but was hard to find)
The Bf3 line goes:
14. ... Bf3 15. gxf3 Qc8 16. Qe2 Qh3 17. f4 Nh5 18. f3 Bxf4 19. Qf2 Bxe3 20. Qxe3 Re5 21. N1d2 exf3 22. Rxf3 Rxe3 23. Rxh3 Rxh3 24. Re1 Rh4
Lichess' Cluster Stockfish just didn't have enough time to calculate this correctly
Stockfish gives up on b5 at Depth 17 (where it switches to Bd7, before ultimately settling on Bf3!). That should give you some idea of how much time it gets to evaluate the positions on Lichess.
Stockfish thinks Bxh2+ loses until it reaches Depth 29, where the evaluation drops to 0.00
Hi,
I'm the guy who played against Lockpick. Interesting game ;) I was really surprised by your sacrifice. I knew that with your level, it was not a blunder, so I was a little bit afraid.
16..Bf3 was also surprising for me. And I thought, life would be so wonderful if I could have two bishop for nothing ! So I was careful. I wasn't sure of real danger, but 17.bxf3 would be really dangerous for my King. Hopefully, the other alternative was 17.N1d2 (I develop my knight and attack f3, but I was afraid to lose the ability to move my queen).
17..Ng4 : Now, I see the danger. After if your Queen can come to h2, I'm checkmate. So, my best simple option was 18.Nxf3. The engine propose g3, but I don't really understand him.
18..exf3 : Again, I have to prevent to go on h2. So the move was easy to find. 19.g3
19..Qg6 : I know what you want to achieve, hopefully for me, I have time to prevent you 20.Nd4 will put my knight in the right position to avoid the checkmate. I think it's the unique solution.
For me what is surprising, isn't really the evaluation of your bishop sacrifice, but the situation after 21..Qh3. There is a lot of different option. I've chosen Qc3 (protect bishop and free the bishop). But the engine advices me to do Bd4 (what ???). I've tried to see what happens after Bd4, but I don't understand why it's always a draw for the engine...
