@Gordima It's OK. You gave the inspiration, which I appreciate, and I gave some SF lines to represent the absolute viewpoint. Its thesis, antithesis, and the reader and his games is the synthesis ;-)
@jahrzehnte
Sounds good! By the way, I would like to offer you a thought experiment regarding absolute evaluations.
1) What is an absolute evaluation if the Stockfish gives +0.7 and Leela gives +0.2? (If you analyze till the end, pretty much all evaluations below +1.5 are a draw).
2) When the new version of Stockfish comes out, will the new evaluation be absolute or temporary until the next version?
Likely, a reasonable approach may look like this: "If something works better - you do it". Then the question becomes, "Does the higher number mean that the line will work better for you?". My understanding is that it is not the case; there are factors such as remaining time and time control, price for the mistake, attacking/defensive nature of positions, standard/non-standard replies, etc
I'm not saying that Sicilian with f4-f5 is always better than Bg5 Najdorf for white of course. However, given the blitz time control - this line will likely give better results because: 1) You don't need to remember much and mostly make similar moves 2) Your opponent will not remember it well because they face it rarely 2) Making defensive moves in blitz games takes more time on average because the price for the mistake is higher.
If one of the best Blitz players plays some openings consistently, there has to be a point in that. And I'm trying to understand that and learn from it.
Sounds good! By the way, I would like to offer you a thought experiment regarding absolute evaluations.
1) What is an absolute evaluation if the Stockfish gives +0.7 and Leela gives +0.2? (If you analyze till the end, pretty much all evaluations below +1.5 are a draw).
2) When the new version of Stockfish comes out, will the new evaluation be absolute or temporary until the next version?
Likely, a reasonable approach may look like this: "If something works better - you do it". Then the question becomes, "Does the higher number mean that the line will work better for you?". My understanding is that it is not the case; there are factors such as remaining time and time control, price for the mistake, attacking/defensive nature of positions, standard/non-standard replies, etc
I'm not saying that Sicilian with f4-f5 is always better than Bg5 Najdorf for white of course. However, given the blitz time control - this line will likely give better results because: 1) You don't need to remember much and mostly make similar moves 2) Your opponent will not remember it well because they face it rarely 2) Making defensive moves in blitz games takes more time on average because the price for the mistake is higher.
If one of the best Blitz players plays some openings consistently, there has to be a point in that. And I'm trying to understand that and learn from it.
Great article! Not only do you analyze Hikaru's Blitz play style, but you even propose some lines to go against him.
One minor suggestion, the hyperlink "www.openingtree.com" does not function properly. It brings me to "lichess.org/@/Gordima/blog/hikarus-blitz-repertoire-2023/www.openingtree.com", which is obviously not what you intended.
One minor suggestion, the hyperlink "www.openingtree.com" does not function properly. It brings me to "lichess.org/@/Gordima/blog/hikarus-blitz-repertoire-2023/www.openingtree.com", which is obviously not what you intended.
@Gordima
> 1) What is an absolute evaluation if the Stockfish gives +0.7 and Leela gives +0.2?
An absolute evaluation is "won" "lost" and "even". Stockfish and Lc0 are different engines with different evaluation mechanisms. We can say that Stockfish has the most objective evaluation, as it is the strongest tool we have.
> If you analyze till the end, pretty much all evaluations below +1.5 are a draw
It is about probabilities. The probability to win a +1.0 position with white is higher than the probability to win a -1.0 position.
> 2) When the new version of Stockfish comes out, will the new evaluation be absolute or temporary until the next version?
It will be the most objective evaluation we have at that point.
> Likely, a reasonable approach may look like this: "If something works better - you do it". Then the question becomes, "Does the higher number mean that the line will work better for you?". My understanding is that it is not the case; there are factors such as remaining time and time control, price for the mistake, attacking/defensive nature of positions, standard/non-standard replies, etc
Agreed, there are a lot of factors one has to consider when playing a line. The objective evaluation is one of them, even in a bullet game. On average, a higher evaluation also means the position is simpler to play. Yes, there are exceptions. I don't always play what Stockfish suggests to me. But I would not play f4-f5 :-) I would look if there are ideas in the Bf1-b5 line and probably then conclude that 2.Nc3 and 3.f4 is not good enough for my taste.
> I'm not saying that Sicilian with f4-f5 is always better than Bg5 Najdorf for white of course. However, given the blitz time control - this line will likely give better results because: 1) You don't need to remember much and mostly make similar moves 2) Your opponent will not remember it well because they face it rarely 2) Making defensive moves in blitz games takes more time on average because the price for the mistake is higher.
1) There are also solid lines against Najdorf. E.g. I scored well in the Adams line (h3) with Nd4-f3 on e7-e5. White is not worse in that line, has ideas and a solid position.
2) But if they remember, white is worse.
3) The moves made here by black seem to be normal development moves.
> If one of the best Blitz players plays some openings consistently, there has to be a point in that.
Call to authority alert.
> 1) What is an absolute evaluation if the Stockfish gives +0.7 and Leela gives +0.2?
An absolute evaluation is "won" "lost" and "even". Stockfish and Lc0 are different engines with different evaluation mechanisms. We can say that Stockfish has the most objective evaluation, as it is the strongest tool we have.
> If you analyze till the end, pretty much all evaluations below +1.5 are a draw
It is about probabilities. The probability to win a +1.0 position with white is higher than the probability to win a -1.0 position.
> 2) When the new version of Stockfish comes out, will the new evaluation be absolute or temporary until the next version?
It will be the most objective evaluation we have at that point.
> Likely, a reasonable approach may look like this: "If something works better - you do it". Then the question becomes, "Does the higher number mean that the line will work better for you?". My understanding is that it is not the case; there are factors such as remaining time and time control, price for the mistake, attacking/defensive nature of positions, standard/non-standard replies, etc
Agreed, there are a lot of factors one has to consider when playing a line. The objective evaluation is one of them, even in a bullet game. On average, a higher evaluation also means the position is simpler to play. Yes, there are exceptions. I don't always play what Stockfish suggests to me. But I would not play f4-f5 :-) I would look if there are ideas in the Bf1-b5 line and probably then conclude that 2.Nc3 and 3.f4 is not good enough for my taste.
> I'm not saying that Sicilian with f4-f5 is always better than Bg5 Najdorf for white of course. However, given the blitz time control - this line will likely give better results because: 1) You don't need to remember much and mostly make similar moves 2) Your opponent will not remember it well because they face it rarely 2) Making defensive moves in blitz games takes more time on average because the price for the mistake is higher.
1) There are also solid lines against Najdorf. E.g. I scored well in the Adams line (h3) with Nd4-f3 on e7-e5. White is not worse in that line, has ideas and a solid position.
2) But if they remember, white is worse.
3) The moves made here by black seem to be normal development moves.
> If one of the best Blitz players plays some openings consistently, there has to be a point in that.
Call to authority alert.
I'd like to see Kramnik's repertoire.
interesting
Why study the opening repertoire of a chess player who is suspected of cheating?
what about the bongcloud?
@jahrzehnte
I enjoy how you so vigorously point out the so-called “logical fallacy,” even though that was merely a reason for me to start the research. That was quite rude. But I guess you can edit the post like the first one.
I agree regarding the absolute evaluation. Some positions are won in 100% of cases against all versions of engines. Let's say a +2 advantage will be won in all correspondence games.
However, the absolute evaluation if the engine gives +0.7 is a draw. The number doesn't represent the probability of you winning the game. After analyzing and comparing many positional/tactical variables, it tells you the conclusion. Let's say black has some weak pawns and attack; engines may give you +0.5 because they see how to refute the attack. Nevertheless, if one needs to find several complicated moves quickly, the results may favor the side with an attack. (you may look at the line and say - oh, I see how to refute it as well, and all the defensive moves look natural. Well, welcome to the logical fallacy, then).
I will play all of these "crappy" lines as you named them and share the results. In the comments, people may share the results as well. This particular line may work only for Hikaru and will be a disaster for me and everybody who tries that after reading my article. However, I don't like religiously following engine numbers. For me, analysis of the experience of a very strong player and my intuition, aka "calls for authority" and "anecdotes," are also important. That's just the way these things work.
I believe that I expressed my position quite clearly and will not reply further, but you can continue writing ofc. The engagement is very important for the article to grow and is very appreciated!
I enjoy how you so vigorously point out the so-called “logical fallacy,” even though that was merely a reason for me to start the research. That was quite rude. But I guess you can edit the post like the first one.
I agree regarding the absolute evaluation. Some positions are won in 100% of cases against all versions of engines. Let's say a +2 advantage will be won in all correspondence games.
However, the absolute evaluation if the engine gives +0.7 is a draw. The number doesn't represent the probability of you winning the game. After analyzing and comparing many positional/tactical variables, it tells you the conclusion. Let's say black has some weak pawns and attack; engines may give you +0.5 because they see how to refute the attack. Nevertheless, if one needs to find several complicated moves quickly, the results may favor the side with an attack. (you may look at the line and say - oh, I see how to refute it as well, and all the defensive moves look natural. Well, welcome to the logical fallacy, then).
I will play all of these "crappy" lines as you named them and share the results. In the comments, people may share the results as well. This particular line may work only for Hikaru and will be a disaster for me and everybody who tries that after reading my article. However, I don't like religiously following engine numbers. For me, analysis of the experience of a very strong player and my intuition, aka "calls for authority" and "anecdotes," are also important. That's just the way these things work.
I believe that I expressed my position quite clearly and will not reply further, but you can continue writing ofc. The engagement is very important for the article to grow and is very appreciated!
@Minucii said in #37:
> Why study the opening repertoire of a chess player who is suspected of cheating?
Hello! As far as I know, he is not convicted. I believe in the presumption of innocence
> Why study the opening repertoire of a chess player who is suspected of cheating?
Hello! As far as I know, he is not convicted. I believe in the presumption of innocence