Yeah, I noticed with Traxler that SF can make terrible evaluation misjudgments.
Sicilian Dragon was busted, 3500 + rated SF died in 26 moves
Yeah, I noticed with Traxler that SF can make terrible evaluation misjudgments.
What position of Traxler, how many million nodes of SF you use for analysis? If you can post FEN, I will check with my SF. In our tests, 500 million nodes or 1.5 minutes of positional evaluation for SF is generally enough for causal evaluation. It might be 3300 strength. SF at depth 1 is 1600 rated. SF at 10 secs is approx 2800-3000 rated.
In this game SF died not because of his positional evaluation during attack, it was due to tactical blunder at move 19 in which he did not see that he will be losing a piece at move 37. ( Tactics blunder at 16 ply ).
In this game at move 19. g4! (not his principal variation, the main reason must be King safety penalty evaluation in SF , he did not expect that his opponent will be exposing his king at black bishop diagonal ). Let us say SF searched about 1 billion nodes in that position at time control, he might have searched about 10% of his resources, or 100 million nodes. It is fair enough to say that SF did not see that tactics at that high depth due to complexity of the game.
Today I busted 1.e4 and 1.d4. The only good move is 1.g4
Haha, You like "Busted" as well. I copied it from @staple13.
Well, I agreed, a bit misleading, like fake news titles! (but it to attract audience)
If I just post it as a normal game, people wont read it.
But the game is worth watching!
But people dont read on my previous post!
Now, you can see how Title attract readers! @staples13 , I copied your title, it works!
from 3.Nc6 to 10...b6, the game is indistinguishable from a game between a couple of 1500 rated players. ICBW, but I'm pretty sure I was one of them. I don't know what current theory is, but 40 years ago (when, with a good tail-wind, I just might have beaten a 1500 player) it was commonly assumed that ...a6 doesn't often work in the dragon. Only when white has made some sort of positional misjudgement, such as playing Bd3, for example. No doubt the computers can explain for us simple folk. With small words we simple folk can understand.
But after 12.f4-5, the position became pretty wild. White's tactical control was impressive! Give credit where due, But did black really play like a 3500 titan?
Really? 1500 play
But Kingcrusher think AB engines(Stockfish) is dying soon by Neural Network engine.
FYI, The chance of SF losing
against top 20 pool or 3300 = 1%
against top 3 pool or 3450 = 3%
Boy! Another day, another opening "refuted!"
Well, such kind of fun topic is started by staples13. It is fun, isn't it.
drmrboss: I'll try to find/recall that Traxler position tomorrow.
Ah, yeah, I remember now: it's the whole Traxler with 5. Nxf7 evaluated by SF9+ at +1.8 (d=31, Cloud) while the true evaluation should be zero (it was shown my smart masters ).
It was TCEC bonus game between 43 cores SF vs 4xP100 Leela. TC 15+5 ,I think.