GM Grigoryan on the "Myth" of Solving Puzzles

I agree with the GM. Probably the hardest thing to do is not to learn tactics or strategy per se, but to harmonize them. As he says, they should go hand in hand.

I guess the easiest way to put it is, why do people lose at chess at different rating levels? take the games of anyone who isn't a master, they lose because they blunder something or miss a certain move... super GMs lose since they underestimate certain positions. They need strategy, we don't.

Strategy in any activity is useless until the person has strong fundamental skill. For example in FPS games any form of team coordination and strategy is useless unless u have the aim. Same for chess, if u are missing moves in variations then it doesn't matter how well you understand your position, u will lose.

That's why i think that strategy, openings or anything like that has absolutely no value in low-intermediate chess.

Puzzle rush is on the clock - fun and entertaining for many players. 
As a tool - puzzle’s, especially for new players should not be practiced as such. A good puzzle to be solved most often takes a good deal of time. 
The perspective is very valid. Seen in the past as the best way of learning - to take your time and get it right. 
However - a puzzle rating was given. Myself and many others pointed out this made no sense - a puzzle rating when unlimited time and tries was at one’s disposal had no meaning.

It was pointed out for a puzzle rating to have any meaning it must be on the clock - the same as a rated chess game. Debate went back and forth. Many claimed putting puzzles on the clock defeated it’s purpose - the only way to practice puzzles was at one’s leisure. At this time came along whomever was In CC’s marketing division who recognized the potential of an app already on the market elsewhere. It was not very popular and seldom used - puzzles with the clock running. 
You know the rest of the story. Puzzle rush was born - but it’s success depended on hyping it’s tremendous value as a learning tool. Everybody must have it - to practice every day !

Finally someone's telling the truth.

I prefer real puzzles:

I believe that taking One chess puzzle, and solving it with an engine from every possible aspect, is more Fun - therefore you'll remember it better, and also more beneficial. And like in many things in life - learning several rules / concepts, is better than memorizing thousands of details. But I need to wait for a GM to say it, so, here I am waiting...

Yes, we lose because of tactics in many cases, but I wanted to say that tactics is not the whole story. Purely tactical mistake is just a simple out of the blue blunder. Those happens as well, especially on novice level, but any other type of tactical mistake happens for some deeper reason.

When you put some pressure on the opponent and he blunders, is it really accurate to say: "Yeah we shuffled pieces and then the game was decided by a tactical mistake and everything before was meaningless"?

In some cases this sentence might be true, but in many I don't think that it is. For instance in the game I posted earlier I tried to explain why that e6 move was more or less the deciding mistake of the game, even though the engine says black could hold afterwards. In human terms the game was that much more difficult from that point, that black lost in just a few moves.

And I have many examples in my games. In one of them, the opponent was about to lose a pawn. He could defend it in a few ways, but if he does, it is a blunder because he will lose a piece in a few moves. And that is exactly what happened. Already a slightly bad position lead to even worse blunder.

That is why I say other aspects of chess go hand in hand with tactics. Other aspects of chess will often lead to slightly better positions that will make opponent crack sooner or later.

Of course, there are opposite examples as well. You could be better for the whole game and blew it, rendering all the strategy and positional themes meaningless, but they matter in many games as well.

Thats like saying the sky is blue......nobody cares.

GIGO....if you stuff your face with garbage you are garbage.

He want to blame the soda companies? The bacon companies. The donuts companies. Capitalism itself. Global warming.

Rich people. Thats whose to blame...first world problems.

Lazy so- called chess coaches perpetuate the myth. So simple to hand the student a sheet of puzzles like homework, tell them get to solving with a review and grade to follow. Lazy is being generous. Such coaching practices are a great disservice and contribute little to the chess community. Rather a coaches medal and expertise is judged by how well they understand and communicate a students particular needs.

I agree with the article, actually.

Tactical puzzles can be extremely helpful, but they'll mainly only help you with one thing: improving your calculation, when there's a forced sequence that will help you win material (or avoid losing material).

But there's more to chess than just calculation. You should also work on the skill of identifying the needs of the position, too.

That's a skill that tactical puzzles don't always help with.

Consider this position, for example:

Show this to a player who is used to solving puzzles, and they'll immediately start looking for forcing moves. They'll consider captures. Potential exchanges. Sacrifices.

They might eventually find a way to win a pawn:

The queen can retreat on the next move, and black has essentially won a free pawn, with no apparent tactical drawback.

"Tactical puzzle solved!" a player might think.

But was winning that pawn really a good sequence?

Positionally, we could actually consider it a "blunder" for black. Because now white's d3 knight, which was previously white's worst-placed piece, now has a clear path to improve itself, and to give white an extremely comfortable game:

An engine will tell you that the game is even, here ... but engines don't look at positions the way humans do. For most human players, having the black pieces now would feel quite difficult, as white now has a firm grasp on the d5 square, and clear and easy paths for further improving his pieces.

Black, meanwhile, will likely struggle to find a harmonious plan.

This is the kind of stuff that tactical puzzles won't teach you ... but it's also the kind of stuff that players should strive to learn.

The unfortunate part is, there aren't really "positional puzzles" out there (not many, at least), to teach positional understanding, in the same way that tactical puzzles teach calculation ...

(P.S. What could black have played, instead? Nd7-Nb6-Nc4! A solid positional improvement, which would play "against" white's strong bishop, and threaten to disrupt white's queenside.

How would black find such a plan? By being identifying the position's best-placed, and worst-placed, pieces, and by being familiar with ideas that are thematic to certain structures. In e5 Najdorf setups, for example, Nd7-Nb6-Nc4 is a very common thematic idea ... and by having this positional knowledge, a player is far more likely to consider such a maneuver than a player who is, instead, only used to looking for tactical solutions ...)

For sure.

5+

??

These problems are not connected to Chess game... I think.

If your chess hasn’t improve with puzzles; it means you are not doing enough, because I’ve been cutting the board like butter and thats with 4 days of training. Ohhhhh I’m so exited and I can’t hide it, am about lose control and I think I love it 🎼. Goodluck guys

the value of puzzles increases alongside their difficulty rating.  accurately visualizing & evaluating every branch a 10+ sequence of moves, while foreseeing & accounting for every possible tricky defense, is excellent training for planning ahead in a real game... boosting your creative problem-solving skills as well.  the lack of a time control with puzzles & survival rush allows for some serious mental weight-lifting.

This might sound simplistic but anyone below masters has really only one "need", to work on their visualization and calculation

There was a study that showed players rated 2000 calculate the most. Masters calculate less because they're more efficient and beginners calculate less because they're bad at it.

And anyway, yes, any "you only need to do ____" advice is overly simplistic.

I'd recommend spending a few minutes to listen to the Perpetual Chess Podcast's recent interview of GM Matthew Sadler.  He points out that current chess engines do an excellent job of playing the game (he estimates they can play at about FIDE 2500 strength) even with their tactical search depth set to only look at the coming move.  In other words, they'll play at that level using only static analysis of the current position.

A human who played that way would tend to look at certain positions that were very tactically favorable and say "I don't know why, but this move just feels very promising to me."  The intent behind drilling puzzles is to be able to start to memorize and learn to identify those patterns that lead to tactics.  Of course, remembering the whole series of how the tactic works is very helpful to executing it, but it's that first moment of looking at the board and thinking "I believe there's a tactic here" that can focus and direct the calculation to find it.

(Also worth noting that GM Sadler's book "The Silicon Road to Chess Improvement," which discusses some of these ideas and how to think about applying them, is coming up soon. I'm eager to read it despite that the value for players at my level is probably low.)

Fair point, however you are not considering their skill of calculation, you are only considering their calculation "time" which is a bad metric.

Super GMs have already spent a decade in this phase of intense calculation. They started calculating when they were 4 years old... for them at that point in their lives calculation has become second nature. This is not the case for intermediate players.

Super GMs beat intermediate players in both the efficiency and the skill of calculation

I have heard that part of the interview and I have run some tests on LC0 at 1 node too. I have made it play against some non top engines and they could beat "her" only when they managed to get into very tactical positions, so it is likely that for a human rated less than 2500 it is not easy at all to win, because they don't have the tactical ability of an engine. But for the topic and for completeness, in that interview Sadler said that he didn't focus very much on his evaluation ability during his career: he trained openings, endgames, typical middlegame positions and tactics...

True, although he also seemed to me to suggest that he might approach things differently knowing what he knows now.

yes