GM Grigoryan on the "Myth" of Solving Puzzles

I don't think that would be insightful, because the result would apply only to a particular player. More so, one player playing this many times will tweak the result. If I play this position as white against someone of my rating, after 10 or so games I'd get to the point where I'd win almost every game, because by then I will have figured out the best way to convert the advantage. As an experiment, only a tournament would make sense, and one with a selection of curated opening positions that are known to favour one side but not as obvious as the one proposed. But then again, we already have that information from the lichess database. 

This whole discussion is getting increasingly pointless, repetitive, and in circles, as plenty of arguments (and even data) are ignored in favour of a dogmatic interpretation of casual comments by some GMs, and the goalposts are constantly shifted. When one argues for positional understanding, then an argument for the importance of calculation and tactics is made, which no one denies. 

I thought only a player for Black and different ones (one try each) for White, but I can see your point and agree. To add to your argument, it has already been said that Leela, set at 1 node (this prevents calculations completely) plays at master level. To be honest, I think that it has some basic tactical knowledge, so maybe it can see simple forks, skewers and pins "at a glance"; but try to make it find the correct moves in a puzzle suited for people rated 1600 OTB: it fails miserably most of the times!

Since it is truly unbelievable that one would be a master just avoiding forks and pins, the theory that strategy cannot be useful until you reach that level is clearly debunked.

GM Sadler said that he didn't work much on his evaluation ability during his career, but studying openings (not just memorizing variations), endgames and typical middlegame positions does improve one's strategic ability. So if a player with normal talent waits to be a master to do so, he might never get there.

@kartikeya_tiwari, I agree with you that strategy can get complex, and can involve a lot of tactical decisions.

But positional understanding can start at a much lower level, too. And it can be done without needing much tactical vision.

Here's a position that was reached, in a friendly game, with a player whom I occasionally analyze with.

It was white's move, and he was looking for what to play. He ended up playing this:

Why did he play that? Because he wanted to develop his queen bishop. And the only way he could see to do that, in this position, was to put his bishop on b2. Seems logical.

But a quick glance at the position should tell a player that the bishop has a rather bleak future on b2, due to white's center pawns being locked on d4+e5.

I explained to him a basic positional idea: that, when one has fixed central pawns, we can view the central pawn chains as "arrows", guiding each player on which side of the board they should be playing on.

White's pawns point toward black's kingside - so that's the side that white should be moving his pieces (and pawns) toward. Especially since black no longer has Nf6 as a defensive resource - which makes the kingside even more vulnerable to attack.

Black's central pawns, meanwhile, point in the opposite direction - so that's the side that black should be moving his pieces toward.

These are positional ideas - basic strategical guides- that require little to no calculation at all ... yet they would still guide the player toward making logical moves that are in harmony with the position.

And this is the kind of stuff that players should be striving to grasp - learning to let the position guide their moves - along with whatever they do to improve their tactics, in the meantime.

Well this claim of yours that tactics and looking ahead don't matter at all can actually be tested. You are implying that strategically understanding a position is more important than "looking ahead" since u said that Leela played at master level while not being able to calculate...

Well if that is true then i urge you to find anyone who is lower than me in tactics in rapid(say a 1500 in rapid) but knows a particular opening very very well, understands the concepts behind it. Infact, he can study that opening for a week understanding basic ideas, then we can have a 45 min game with him playing that opening and i would still beat him even though i know absolutely 0 theory and strategy about any position from that opening. He will understand it way better than i would but he would still lose.

I am ready to test it out if u want. 

I didn't say anything like that and didn't imply anything, while you should demonstrate you don't use any positional concept during play; we only have your word.

Leela at 1 node does not calculate anything. This is a fact and I don't see how this fits with your theory. I and GM Sadler say that LC0 at 1 node plays like a master, since for you matters so much what GMs say (and, please, don't use arguments like "yes, but Carlsen and Nakamura are stronger than Sadler"). To me ipse dixit is important, but it is not a proof.

Let me make an observation. The thing that is most important is being able to see where the game is going to go in the next few moves. The other thing that is important is being able to tell when one outcome is better than another without analysing further.

Tactics are simply where the differences in the outcomes are large. In practical chess the differences in the outcomes are not always large: this is called positional play. But the ability to analyse - seeing good lines - is common to both.

The ability to analyse must be mainly about devoting an appropriate amount of time to different options - humans don't have the luxury that computers have of looking at millions of variations, so this is way more important for us (it's fair to say even computers wouldn't be that great if they were not selective. Brute force search with no preferential treatment of different lines (plus a reasonable evaluation function for endpoints) would be very weak compared to better designed engines. To get an indication of how weak, it might see 5 moves ahead at standard time controls (based on example Stockfish nodes per second starting with the opening position, which actually has rather fewer possibilities than an average position because so many pieces are blocked).

You did imply that when u said that an engine which isnt calculating AT ALL is playing at master strength and i don't buy it. What if i setup a 2 move combination? the engine isn't going to see it? 

You can check by yourself, if you do not believe me or Sadler. Just download LC0 and set it to examine just one node of the tree of variations. It will examine only the root node, i.e. the current position (check its logs, if you don't believe it) and evaluate it according to its NN (it does it instantly). Yes, it fails to see combinations that even players rated 1200 see, but try to beat "her" at standard time control. Good luck and let us know how it goes, ok?

As i said, you are speaking from a place of already having the tactical vision. Go back to the time when u were 1200... would you really have tried to make deep plans? plans mean nothing when u are hanging your queen every move lol.

...

My point is, the goal of weak players should be to make sure that a move is not a blunder... "activating one's bishop" is not important, what's important is that u don't hang a pawn or miss a move which gives black a huge advantage.

This is what actual chess coaches do.

A proper coach does more than just work on tactics alone. I know you're convinced otherwise, but if you were to hire a titled coach, I believe you'd be surprised (hopefully pleasantly) by the amount of positional and strategic ideas that your coach would help you with.

Yes, tactics are a big part of learning. But they aren't everything. Tactics don't exist in a vacuum. You still need to know how to logically move forward in a position ... otherwise, the only thought process going through a player's mind would be, "Where is the tactic? Where is the tactic? Where is the tactic?" on every single move.

Consider this position. What's a good, logical move for white here?

If you thought of Rook to c1 - congrats! You just made a strategic move. The needs of the position clearly suggested a logical move for white: improving his rook, by placing it on the open file.

Yet you'd be surprised by the number of lower-rated players who would struggle to find Rc1 in this position.

Why?

Because they're looking for tactics.

They're so used to tactical puzzles, that if you show them such a position, and ask them to find a move, they immediately assume that there must be a winning combination somewhere. So they hem and haw, and begin calculating sequences.

Case in point: a while ago, I showed this to a young player (around 900 in strength), and asked him to tell me what his first thoughts were. He got stuck on a single move, and was unable to see anything past it:

Pawn to g5

Why pawn to g5? Because, in his words: "I can tell there's something there, kind of. I just can't find it."

Why is that? Because he was, at the time, so used to thinking of chess as mostly tactical. So when faced with a position like this, he still assumed that there was a winning tactic somewhere ... if only he could just "find it".

Instead of thinking of the position in a logical way, he viewed it as a "puzzle" to solve, and immediately looked for the most forcing move available.

When I pointed out Rc1 to him, he looked disappointed. Disappointed because it seemed so "easy". And also disappointed because Rc1 was nothing special - there was no firework combination that ended with won material, or a forced mate - as tactical puzzles have taught so many of us to expect.

Rc1 is a quiet, fundamental move that improves white's position ... and our young player didn't like it. He still asked me to go over pawn to g5 with him ... just in case there was something there. (There wasn't.)

Now yes, I agree with you that we don't want to miss tactics, or to blunder. I don't think anyone on this thread thinks that tactics are worthless. My main point is that tactics and strategy go hand in hand ... but too many players think that tactics are everything and all one needs ... when, in reality, they're missing half the picture ...

Rc1 makes the most sense. It seems like a move that can't be wrong here, just taking the open file. The position is calm, so probably there is no rush. At first I was thinking about Ne4 as well, to improve the knight but it is not stable there, so just d5 and I've lost a tempo. So that is stupid.

Other than Rc1, perhaps some pawn move is ok as well, gaining space b4 or d4. That is probably how I would think here.


Edit: Hm, engine is fine with b4, but for some reason doesn't appreciate d4 (it doesn't hate it, but it views it as a slight inaccuracy). d4 is somewhat tempting because it makes his bishop worse, at least for now. My best guess why engine slightly doesn't like it is because it weakens e3. Do you think there is something else wrong with it?

Tactics are not only where "difference of outcome is large"... tactic doesn't only mean sacrificing your queen for checkmate. There are many, many small tactical motifs present in literally EVERY SINGLE chess game....  Super GM games have hundreds of tactical motifs.. the difference is they are good players so we never see those motifs in games since they stop all of them...

True for practical purposes, but you will find that every tactical puzzle you can do on chess.com (and probably most other places) are characterised by a large difference in evaluation between the correct line and any alternative. If the difference is not decisive (i.e. almost certainly affecting the result with accurate play) they are not included. I have noticed that chess.com puzzles end sometimes when one would expect them to continue because it happens that rather than there being one best move there are two or more. Of course, a position where there are two moves leading to one result and all the rest leading to an inferior result can still be a tactic - the reason they stop is the convenience of only dealing with a single winning line. So the puzzles are an even narrower category than tactics - they are tactics with a single best move. This permits reasoning inappropriate to chess - "if move A and move B seem equally good here, I must be making an error of reasoning".

What you describe is more like my description of analysis in general. The "motifs" guide the choice of lines even in a position where there is nothing decisive.

Tactics are the soul of chess. Strategy is strictly for people who have already mastered those small tactical motifs. Making plans before that is futile

If there is a game between a guy who is 3000 rated in understanding strategical concept of a position but is 100 rated in tactics (can't see a single move ahead) vs a guy who sees 10 moves ahead but doesn't understand the position at all, then the latter would crush the former... this is how chess is.

I was going to say there are no such players, but if you have a top engine and set it to 1 ply analysis, you get something pretty close to what you describe (except its strategic understanding is probably not that good - engines are over 3000 more because of their powerful analysis than their genius strategic understanding).

A proper coach does more than just work on tactics alone. I know you're convinced otherwise, but if you were to hire a titled coach, I believe you'd be surprised (hopefully pleasantly) by the amount of positional and strategic ideas that your coach would help you with.

Yes, tactics are a big part of learning. But they aren't everything. Tactics don't exist in a vacuum. You still need to know how to logically move forward in a position ... otherwise, the only thought process going through a player's mind would be, "Where is the tactic? Where is the tactic? Where is the tactic?" on every single move.

Consider this position. What's a good, logical move for white here?

If you thought of Rook to c1 - congrats! You just made a strategic move. The needs of the position clearly suggested a logical move for white: improving his rook, by placing it on the open file.

Yet you'd be surprised by the number of lower-rated players who would struggle to find Rc1 in this position.

Why?

Because they're looking for tactics.

They're so used to tactical puzzles, that if you show them such a position, and ask them to find a move, they immediately assume that there must be a winning combination somewhere. So they hem and haw, and begin calculating sequences.

Case in point: a while ago, I showed this to a young player (around 900 in strength), and asked him to tell me what his first thoughts were. He got stuck on a single move, and was unable to see anything past it:

Pawn to g5

Why pawn to g5? Because, in his words: "I can tell there's something there, kind of. I just can't find it."

Why is that? Because he was, at the time, so used to thinking of chess as mostly tactical. So when faced with a position like this, he still assumed that there was a winning tactic somewhere ... if only he could just "find it".

When I pointed out Rc1 to him, he looked disappointed. Disappointed because it seemed so "easy". And also disappointed because Rc1 was nothing special - there was no firework combination that ended with won material, or a forced mate - as tactical puzzles have taught so many of us to expect.

Rc1 is a quiet, fundamental move that improves white's position ... and our young player didn't like it. He still asked me to go over pawn to g5 with him ... just in case there was something there. (There wasn't.)

Now yes, I agree with you that we don't want to miss tactics, or to blunder. I don't think anyone on this thread thinks that tactics are worthless. My main point is that tactics and strategy go hand in hand ... but too many players think that tactics are everything and all one needs ... when, in reality, they're missing a big part of the picture ...

After more analysis g5 is actually bad but not because it's bad strategically but because Ne4 gaining a tempo and forking the pawns doesn't work since black has a way to track down and win the h pawn in return, thus gaining a pawn...  so even in this dry position there were tactics present, nice. 

However, as pointed out earlier, none of this matters since lower rated players like us win or lose in the middlegame... ain't no one getting so such dry positions lol