Well as soon as humans start playing 3900 rated humans. And as soon as those 3900 rated humans can memorize all lines of an opening then this will matter.
Engine refutation of popular openings?
Sort:
These openings were under clouds even before computer era. They were never as reliable as other openings and regularly went through crises.
As for how well an opening stands the most severe theoretical tests, you have to search a database of recent correspondence games (all played with huge computer assistance nowadays and the critical lines explored very deeply, much better than what you could do at home).
"Is there content online available that shows this same thing?"
++ See Figure 4
https://arxiv.org/pdf/2009.04374.pdf
Dutch Defence, Chigorin Defence, Alekhine Defence, King's Gambit
Another source is the TCEC superfinals. They impose 50 slightly unbalanced openings on the 2 best engines and let them play one game with white and one game with black.
Now there are 3 possibilities:
draw-draw: the opening is equal
win-loss: the opening is busted
win-draw: undecided
(win-win does not happen as the engines are of comparable strength)
King's Indian Defence is suspect.
"King's Indian Defence is riskier for black than King's Gambit for white" - Bronstein
Fischer did not dare to play it against Botvinnik in 1962 or Spassky in 1972.
Kasparov and Radjabov gave up on it.
GM Nefedov lost 2 games with it in the 30th ICCF World Championship final.
Black gets into trouble on the queen's wing and loses the endgame.
Those 3 openings may be bad practical choices these days, but they're not refuted.
For example the King's gambit is not lost, but as white you'll need to prepare for something like a dozen different black defenses. Even worse, stylistically, you may get anything from a boring position to complete chaos, and from easy moves to deep memorization is required to equalize. It simply doesn't make practical sense to make it your main opening.
Don't go to entertainment videos for your chess education.
Nerwal wrote:
These openings were under clouds even before computer era. They were never as reliable as other openings and regularly went through crises.
As for how well an opening stands the most severe theoretical tests, you have to search a database of recent correspondence games (all played with huge computer assistance nowadays and the critical lines explored very deeply, much better than what you could do at home).
Oh so true!
#10
"A position which is easily holdable for an engine and its very deep tactical calculation,
may well be utterly dangerous for practical play."
++ That is true, but the reverse:
"A position which is not holdable for an engine and its very deep tactical calculation,
may well yield good results for practical play." is true too.
It is often good in practice to play an inferior line that the opponent does not know how to refute or cannot refute over the board within the time limit.
The question of the original poster was about engine play or correspondence play,
i.e. from the theoretical, not the practical point of view.
#7
"And even following the 7...Nc6 Mar del Plata policy, the most consistent answer to 9.b4 is 9...Nh5! - Nefedov's 9...Ne8 is already (too?) risky."
++ The elder 9...Nh5 caused problems in correspondence as well as over the board,
that is why Nefedov and others turned to 9...Ne8.
Kasparov suffered a loss against Kramnik and gave up on the King's Indian Defence
https://www.chessgames.com/perl/chessgame?gid=1070932
As I mentioned in the original post, i understand that these openings are still alive and well and will probably continue to be used for a long while and with good reason. I am not debating that, nor is that what the youtuber i mentioned claims. I am not examining these openings as for their practicality, winrate and popularity. In this particular instance I am only interested in what the top engines think of them. Ironically the first reply to this thread ignores that completely, it's like people sometimes don't read what the post says. I am also not using the referenced channel for chess education, I do not play these openings nor am I interested in doing so.
Engines are not people and people are not engines. Take engine analysis in the opening with a grain of salt; they are notoriously dubious in some of their opening evaluations.
AggressiveMarketing wrote:
As I mentioned in the original post, i understand that these openings are still alive and well and will probably continue to be used for a long while and with good reason. I am not debating that, nor is that what the youtuber i mentioned claims. I am not examining these openings as for their practicality, winrate and popularity. In this particular instance I am only interested in what the top engines think of them. Ironically the first reply to this thread ignores that completely, it's like people sometimes don't read what the post says. I am also not using the referenced channel for chess education, I do not play these openings nor am I interested in doing so.
There are almost no openings that are forced losses according to the engine... I mean, off the top of my head, only the Latvian gambit and maybe the Grob come to mind (and I'm not even 100% sure about the Grob). The drawing margin of most endgames is quite large, and the opening is the most forgiving phase of the game. You have to play truly awful crap for it to be a forced loss in the first ~5 moves according to the engine.
At my level I could not possibly care less what some engine refutes. I play chess to exercise my mind and have fun.
nMsALpg wrote:
Those 3 openings may be bad practical choices these days, but they're not refuted.
For example the King's gambit is not lost, but as white you'll need to prepare for something like a dozen different black defenses. Even worse, stylistically, you may get anything from a boring position to complete chaos, and from easy moves to deep memorization is required to equalize. It simply doesn't make practical sense to make it your main opening.
Don't go to entertainment videos for your chess education.
Yes, this is all well put.
One thing that makes conversation difficult is that people sometimes use the word "refuted" to mean different things -- and may in particular mean different things when they call a Black opening refuted than from when they say the same of a White opening. I wouldn't call any of those three openings refuted if what we mean is that they lose by force. But sice we typically assume that White ought to be able to emerge from the opening with at least a slight advantage, some people might feel comfortable calling the King's Gambit refuted in the mere sense that Black can achieve equality or maybe even a very small advantage. That's different from the stronger claim that (as Fischer said) it loses by force, but it could be justified if you use the word that way for White openings, as some poeple do (but as nobody does for Black openings).
All of this is irrelevant to your main point which is that the more pertinent question about them is whether they're practical choices. There are indeed good practical reasons to avoid them, especially the KG and the Dragon -- but even so, it's a very personal decision, and there are people who have managed to have good practical results with all three.
While it's possible Rozman said this, I'd kind of like a citation before taking it on trust. It seems possible he might have said something more defensible, which was garbled into this indefensible claim in the transmission. Or maybe not.
tlay80 wrote:
It seems possible he might have said something more defensible, which was garbled into this indefensible claim in the transmission. Or maybe not.
Yes, this was my first thought too.
And even if he said "refuted at the engine level" he probably didn't actually mean in engine vs engine play... he probably meant that these can't be used at the professional level because the professional's peers would use the engine to become extremely well prepared, and score much better against that player than if that player chose something else.
I remember Nakamura saying something like that about the French defense... that it's by no means a bad opening... but OTB if he had a choice between forcing his opponent to play 1...e5, 1...c5 or the French, he'd choose the French 100% of the time.
The Portuguese line in the Scandinavian (1 e4 d5 2 exd5 Nf6 3 d4 Bg4) looks to be on very thin ice these days. In Smerdon's very good book on the 2...Nf6 Scandinavian he says that if there is a refutation to 3 d4 Bg4 then it must be 4 f3 Bf5 5 g4!. Stockfish appears to not like black's side at all in the Portuguese.
tlay80 wrote:
One thing that makes conversation difficult is that people sometimes use the word "refuted" to mean different things -- and may in particular mean different things when they call a Black opening refuted than from when they say the same of a White opening. I wouldn't call any of those three openings refuted if what we mean is that they lose by force.
Refuted doesnt mean lose by force. Only people who care about winning or losing care about that. In fact, I've lost many a 5 min game in superior evals. Even long games...but the deal is im not looking at one game. If you are looking to win one 2 min game, hardly anything is 'refuted'...but for me, if im learning thats a plus than winning a no effort required.
brink2017 wrote:
The Portuguese line in the Scandinavian (1 e4 d5 2 exd5 Nf6 3 d4 Bg4) looks to be on very thin ice these days. In Smerdon's very good book on the 2...Nf6 Scandinavian he says that if there is a refutation to 3 d4 Bg4 then it must be 4 f3 Bf5 5 g4!. Stockfish appears to not like black's side at all in the Portuguese.
I forgot which GM said it, but during the candidates commentary he pointed out that many openings with something that looks bad (like +1.2) tend to go to a draw at very high depth / as you explore them more, simply for the reason that eventually everything either goes to mate or draw.
They were talking about Caruana's habit of purposefully giving himself "bad" positions that are eventually a draw with very good play.
nighteyes1234 wrote:
tlay80 wrote:
One thing that makes conversation difficult is that people sometimes use the word "refuted" to mean different things -- and may in particular mean different things when they call a Black opening refuted than from when they say the same of a White opening. I wouldn't call any of those three openings refuted if what we mean is that they lose by force.
Refuted doesnt mean lose by force. Only people who care about winning or losing care about that. In fact, I've lost many a 5 min game in superior evals. Even long games...but the deal is im not looking at one game. If you are looking to win one 2 min game, hardly anything is 'refuted'...but for me, if im learning thats a plus than winning a no effort required.
I think you’re proving my point, which is that people mean a range of different things by “refuted.” Standards differ from person to person and even from situation to situation (one often thinks refutations in different terms for black and white). That’s why it often becomes helpful to slow down and clarify what you mean by “refuted.”
#22
"people mean a range of different things by “refuted.” "
++ refuted or busted means: it loses by force
Any other meaning makes no sense: when it draws it cannot be refuted
Fischer already wrote that about the King's Gambit 'it loses by force' and about the Dragon Variation 'weak players beat Grandmasters with it' and implied it about the King's Indian Defence 'He had come well-armed for my King's Indian'.
The engines only confirmed his verdicts.
Popular chess youtuber GothamChess has mentioned on multiple instances that some popular openings such as the king's indian defense, the king's gambit and the sicilian dragon have been (virtually) refuted at engine level. Now I understand that these openings may still provide significant value and utility to those that play them, and therefore have not been truly refuted, especially since some of the lines may require perfect play to retain the advantage as is often the case with computers. However I would very much be interested in examples of said computer analyses/theory, that illustrates the engine's approach to facing these openings. Would an engine analysis made at home, say with stockfish and 10k knodes be sufficient to illustrate these openings' slight weaknesses and the point made by Gotham? Is there content online available that shows this same thing? In other words where does he get that from, and where can I see it for myself? Apologies if this has been answered before, and thanks to anyone that answers any of these questions.