After d4 Nf6 c4 e5 dxe5 Ng4 Nf3 The stockfish cloud is recommending Nxf2. After this stockfish likes a3 for some reason. Maybe I'm missing something and this is a brilliant queen sacrifice but I believe it is a bug.
Wow, that is really weird and a great find. Even Nxf2 is the proposed move - as long as you don't add more lines. With 2 or more lines everything is back to normal. I have no clue how this can happen, definitely a bug and surely not the only one. I remember old SF versions having the same problems in Crazyhouse openings. There were just wrong cloud results stored and in the most played openings black was supposed to get mated after a very solid 3rd move, which was totally incorrect.
It does indeed look like a wrong cached "cloud" value, once you let your local engine run long enough to get past depth 51, the "best line" switches to 4... Nc6 with +0.5. But when I played Nxf2, it shows 5. a3 Nxd1 with "0.0" and depth 52/99 again. I left the evaluation run again so that there is a chance that my result will replace the cached value.
BtW, the value of "0.0" (which clearly does not match the 5. a3 Nxd1 "best line" by far) is a hint that this is actually a glitch ("lost" value) rather than a result of an imprecise calculation.
BtW, the value of "0.0" (which clearly does not match the 5. a3 Nxd1 "best line" by far) is a hint that this is actually a glitch ("lost" value) rather than a result of an imprecise calculation.
Tak bohužel... I když jsem nechal obě pozice (po 4... Nxf2 i po 5. Kxf2) spočítat do větší hloubky (54 a 53), "cloudová" hodnota se neaktualizovala a v novém okně se zase ukazuje jako nejlepší ten nesmysl. :-(
@mkubecek said in #4:
> Tak bohužel... I když jsem nechal obě pozice (po 4... Nxf2 i po 5. Kxf2) spočítat do větší hloubky (54 a 53), "cloudová" hodnota se neaktualizovala a v novém okně se zase ukazuje jako nejlepší ten nesmysl. :-(
Oops, sorry... A translation for the majority unfamiliar with this weird language:
Unfortunately, even if I got both positions (after 4... Nxf2 and after 5. Kxf2) analysed to a bigger depth (54 and 53, respectively), the "cloud" value wasn't updated an a new window still shows the nonsense as "best". :-(
> Tak bohužel... I když jsem nechal obě pozice (po 4... Nxf2 i po 5. Kxf2) spočítat do větší hloubky (54 a 53), "cloudová" hodnota se neaktualizovala a v novém okně se zase ukazuje jako nejlepší ten nesmysl. :-(
Oops, sorry... A translation for the majority unfamiliar with this weird language:
Unfortunately, even if I got both positions (after 4... Nxf2 and after 5. Kxf2) analysed to a bigger depth (54 and 53, respectively), the "cloud" value wasn't updated an a new window still shows the nonsense as "best". :-(