And Uhoh, the reason I dislike black is that black has no proper counterplay after you castled and yet king in the middle is vulnerable to piece sacs. I evaluated it to a slight edge for white but that's trying to be objective. Black is far harder to play than white. You don't like my O'Kelly, apparently, and yet I would have had to have made a mistake to get an equally bad position to what you have achieved here apparently by playing the main line, as you claim.
Sicilian defense: Scheveningen variation with a6
Sort:
Jul 28, 2020
0
#63
Uhohspaghettio1 wrote:
Alright, I checked it out and apparently Stockfish is giving something similar to those numbers even at a large depth. Thanks for alerting to that as it's really interesting, a rare find.
I have to admit I had to do a double take and for a second started to wonder whether black was somehow in massive trouble after all. However Black is completely fine, as is verified by several authors and engines. Black has clear counterplay down the kingside and white will likely give up the h-pawn according to Negi. I've never personally experienced such a buggy evaluation myself, so again nice find.
Negi's book was was being written in 2013-2014. He does not state it in his books, but I suspect he (or someone from Quality Chess) went through the lines with an engine just to make sure there were no surprises, but they likely did not spend any time updating his notes from "better" to "winning", as there would be no point in doing so.
This is not a buggy engine evaluation.
I do not have Houdini, but Lc0, SF11, and Chess.com's Komodo MCTS all show similar results. The only thing they disagree on is exactly how much White is winning.
Uhohspaghettio1 wrote:
Glitched evaluations like this do happen from time to time, for example Stockfish gives this position from the middle game -18.1 when it's actually a forced draw:
https://chess.stackexchange.com/questions/24867/bad-stockfish-evaluation
It was known for years that closed positions give AB-engines difficulties. That is why Hikaru (and others) have used things like the Hippo when playing engines. In the position posted in that Stack Exchange question, you run into the horizon effect. Specifically, White draws by sacrificing more material (something that will quickly get pruned out by the engine because it thinks it is nonsense), followed by forcibly closing up the position and moving around for 50 moves while Black (in theory) has a massive material advantage. The problem is the engine cannot see 100-ply down the road and realize that it has no way to break through, so it is relying on material considerations. This can happen with very closed positions.
The problem is the position in the game is not a closed position at all. In open and fluid positions, the likelihood of finding an error in evaluation when multiple engines agree with GM-praxis is so small that you have a better chance of winning the MegaMillions a couple times in a row.
Uhohspaghettio1 wrote:
Supposedly the King's Indian Defence is full of totally faulty evaluations at the highest theoretical levels.
The KID tends lead to closed positions. See previous discussion about closed positions and engine evaluations.
Uhohspaghettio1 wrote:
I put it through Houdini on my own machine for as long as possible and it gives +0.48 at depth 23, which is more than reasonable for a double-edge opening that gives black chances of winning, especially a (nowadays) offbeat one where you may catch your opponent off guard.
"As long as possible" and "depth 23" do not coincide. On my machine, SF11 reaches depth 23 in about 15 seconds (and gives a ~+2.5 eval). I'm not sure how you have Houdini configured, or what you are running it on, but your result seems faulty.
Uhohspaghettio1 wrote:
You're claiming Negi's book is old.... bruh that's not it, I am pretty certain the evaluation is just wrong. If something is good enough for Houdini at depth 23, Negi and other engines it's good enough for me. I somehow doubt Stockfish really has it worked out to a forced win, it's a faulty evaluation.
As I stated earlier: in the game you have been following, White blundered into a draw. But, I'm sure you know better and I wait for you to prove Negi, Shaw, de la Villa, Illingworth, SF8/9/10/11, Lc0, Komodo, and of course everyone else on this thread wrong.
Jul 29, 2020
0
#64
Uhohspaghettio1 wrote:
It's a pity it's the liar and troll PawnTsunami that is coming up with this stuff because I would like to answer some of those arguments but I just can't when he says so many openly trolling and false things.
The irony in this statement is amusing.
You can save yourself a lot of time and energy if you just learn to look in the mirror and say the words "I was wrong". Continuously trying to make the claim that "black is fine" when Black is already losing is simply ridiculous.
White's position is now fully coordinated and white's dominance of the board's squares is beginning to be keenly felt. White is beginning to generate multiple threats and it's difficult to expect black will successfully counter them. I think it was a fundamental error for black to 0-0-0 instead of b5. Now it's too late and I think black may be lost.
That is just via a cursory, visual analysis of the position for two or three minutes, of the type that I could not achieve were I to spend the rest of my life attempting.
<<White's position is now fully coordinated and white's dominance of the board's squares is beginning to be keenly felt. White is beginning to generate multiple threats and it's difficult to expect black will successfully counter them.>>
To put it differently, I would have to have blundered with my O'Kelly to have acheived the same effect.
Jul 29, 2020
0
#69
Uhohspaghettio1 wrote:
And what about Golubev - Barskij 2018? You're saying the first deviation from theory by an International Master was a losing mistake, keeping in mind that it was used in a otb game almost certainly prepared with the aid of a computer? It hardly seems like the sort of opening black is going to waltz into without knowing what he's doing and the result was a draw. So I will continue with him for now:
20... Nd7
We can always check this with computer analysis later, highly doubt black is losing. You could also describe why you think white is winning, especially if you're confident black can't escape.
Keep following that game, please ... so you can see exactly where White blundered into the draw (Hint: you do not have to wait too long now).
You're studying the thing and I'm just glancing at the online diagrams so everything being equal, your impressions take precedence over mine.
But specifically, I think white is starting to control b6 and has options of Bxa6, Qf1, e4-e5 (all and/or) and if the response to e4-e5 is d5 or de, then maybe Nb-c5 and the overall effect is that white is threatening your king, which would have been safer left in the centre, with more pieces than you will be able to defend against given such limited maneouvering room. I don't like black's position although it's a difficult position when one doesn't have a board to move pieces around on. I just instinctively think black's position is fragile and so white is winning. Could be wrong.
White also has Rb3, swinging it onto the k-side because the knights will have moved, or even Rb5 sacking the exchange and removing one of your defensive pieces. White's other rook is strong on the e file and your spare rook is tied down defending g7 and losing that pawn may well be enough to lose the game given black's lack of counterplay. So that's more moves lost before you can connect rooks. It doesn't look good but maybe you have a surprise resource. If I could see it, it wouldn't be a surprise.
cool
Thought you'd play b5 ... otherwise you don't have space to breathe.
White has no obvious win, or Spaghettio wouldn't be playing this line. But white has all sorts of combinations of ways to feed pieces in and break up black's queenside, which is why the K should stay in the centre. Is it technically defensible?
Jul 31, 2020
0
#78
Optimissed wrote:
White has no obvious win, or Spaghettio wouldn't be playing this line. But white has all sorts of combinations of ways to feed pieces in and break up black's queenside, which is why the K should stay in the centre. Is it technically defensible?
He is playing this line because a game between 2 IMs ended in a draw. What he has failed to realize is that White blundered into the draw on move 23 by attacking with the wrong pawn.
Uhohspaghettio1 wrote:
22... Bb7
Optimissed wrote:
White has no obvious win, or Spaghettio wouldn't be playing this line. But white has all sorts of combinations of ways to feed pieces in and break up black's queenside, which is why the K should stay in the centre. Is it technically defensible?
*sigh* That guy is going by that odd evaluation he found by Stockfish (and that family of engines which are notorious for giving very similar evaluations while others say black is completely fine) and he thinks that white made some "blunder" coming up that gives up the draw and that FizzyBand will see it and play the "correct" move. Imagine how foolish he will feel when his computer analysis doesn't work when played out as happens all the time. What I posted with Houdini has the time in seconds next to the evaluation, so you can get a picture of how long I left it on - like 20 minutes, I'm using a really old laptop. It's extremely unlike Houdini wouldn't suddenly see something on a higher depth but try it by all means if you think otherwise, Houdini that used to win against Stockfish, and still possibly does against the free versions.
Now there's lots of tactics to come, there may well be huge mistakes on either side there. I'm not going to spend tons of time trying to obsessively calculate everything to prove a point, just so we're clear.
I find your comments here a bit hard to follow.
I thought you knew that I don't use chess computers, Stockfish or otherwise, except occasionally the Chess.com analysis tool after a game. In situations like this I always do it in my head and never spend more than a few minutes on it, which is why you and people like Pfren get to laugh at me when I make a mistake, which keeps everyone happy and makes us all feel good.
Jul 31, 2020
0
#80
Optimissed wrote:
White has no obvious win, or Spaghettio wouldn't be playing this line. But white has all sorts of combinations of ways to feed pieces in and break up black's queenside, which is why the K should stay in the centre. Is it technically defensible?
Spoiler Warning: For those that do not want to see what I was referring to, skip this post ....
Okay, with that out of the way:
But Stockfish is known not to be so strong in a complex middle game. It's known to be strongest in endings.