GM Grigoryan on the "Myth" of Solving Puzzles

Regarding the “origins of the myth” as Grigoryan puts it. The “myth” is older than he is. Chess engines did not create it.

Nonetheless, I think his views are worth considering aside from their historical inaccuracy.

There have been many instances in my games where a move that looks good strategically fails because of some move that i missed.  This is what happens in a real game... a move that looks good strategically would still need to be calculated for and many times following strategically correct things fail so strategic principles seem to be harming people if they are not good in just seeing moves or forcing continuations...

For example in this game I played 9.. Nc4 thinking that my knight is pretty good on the c4 square since it puts pressure on the e3 pawn and it also puts pressure on the b2 pawn which makes it tough for his bishop to get out. If he plays b3 then i can retreat to d6 and my next move Nf6 and i have solid control over the e4 square...    I discarded Nxf3+ since i thought after Queen takes i am just bringing that queen into the game and it comes with tempo on f7 square  and his bishop is blocked pretty badly by my pawn chain so i felt no need to exchange my knight for his bishop..

It turns out, after Nc4 i am losing since i missed e4 move which threatens to win a pawn and if i take it then Nxe4 and white is coming to g5 with his knight and i am lost...

It is just an example of how even moves which look strategtically fine need to be checked with calculations anyway so unless a player has solid tactical eye and can "see" moves and variations and threats, following strategical concepts can be a big detriment.

I think there is something to be said about how you solve puzzles.

Finding each move one by one won't help you develop. Puzzles should be solved in their entirety, before making the first move. That at least helps with calculation and visualisation.

This is a really interesting article and idea to have posted. So thanks so much for sharing that. 

For positional puzzles, check out Jeremy Silman's lesson on chess.com: Roots of Positional Understanding. It is amazing! 

idk what u are saying, most games before the 2000 level ARE decided before the endgame lol

That raises the question, have you got those stats? 

So you have no stats but only "I have observed"... well i have also seen thousands of low rated OTB games where tactics decided games before the endgame phase... so now we're even

I've always thought that puzzles are overrated. I've done a zillion of them and I think that the time spent/extra wins ratio is just lousy. There are lots of puzzle warriors who have ratings that are a thousand points higher than their rating, and that in and of itself says something. Puzzles have solutions. 99% of chess positions don't. You can spend minutes calculating a puzzle knowing that there's a solution, but you don't have the luxury of that in real games. Sure, once in a while knowing a puzzle will set off your spidey-sense, but that's rare.

I think chess coaches tell students to do puzzles because it's an easy thing for them to do. The most I ever learned from a coach is when he set up a board and said things like:  where are the weak squares, how can you take advantage of them, which pieces are weakly defended, what would your plan be in this position... stuff like that. Puzzles not so much.

I have played millions of games OTB with over 40 years of competition, i am not taking opinions with noobs with only thousand games anyway

Puzzles are supposed to increase your look ahead ability. When you calculate various lines and try to look ahead the position down several moves then it strongly improves your chess abilities. That's the point of puzzles

Millions?

Why would they increase your look ahead ability though? What logic is there behind that? When you're analyzing a position you're looking ahead anyway, right? I am confident there is no possible rationale that you could have to argue that puzzles help train you look ahead better than normal analysis of a position. 

If puzzles have a place I feel like it's more a psychological thing, that it's fun to do and have a definite answer. They are like the burgers and fries of chess. And in fairness could have a respectable place in chess training if you are concentrating on eg. the final mating attack.    

The key to benefitting from puzzles is to not get locked into a single source. I use books with checkmates, tactics, and positional puzzles. I use chessdotcom, chesstempo, Lichess. I use several FB groups. I create puzzles (another form of solving is finding and/or composing).

The reason is because it's specialized training of one specific aspect of a game. That's why fps games have aim trainers even though when u are playing u are aiming anyways...

Normal chess games don't require you to look ahead that much. If i play d4 then u don't need to concentrate and calculate lines, you can just play d5. Many positions in normal chess games are extremely dry and there is not much to calculate.

Puzzles make sure to isolate positions where you absolutely MUST focus and look ahead. Will u get positions where you have to look ahead 10 moves in a real game? yes... will u get it as often? no.

The point of taking ivermectin (horse goo) is to prevent you from getting Covid. That doesn't mean it does a very good job of it. I tell you that puzzles are bad at doing what they are supposed to do and you answer with telling me what they are supposed to do? All righty then.

Puzzles are not bad for most people, but people improve differently. 

Seriously, how many posts have we all seen that say something like "I'm rated X at puzzles but my blitz rating is only Y! What's the problem?!?!"

The problem is that they are weakly correlated. Spending days solving puzzles isn't going to make you a good player. Will it help some? Sure, but not much.

You speak from your own experience. But that isn't necessarily everyone's. I find puzzles quite useful as they confront you with realistic game situations. Of course in a puzzle you have the advantage of knowing that "there is something there", but solving them can make you more aware of critical positions and tactical patterns when they appear in a game. And yes, puzzle ratings are super inflated but that is not an argument that puzzles don't help...

Of course not. But in which sport, art, science or language can you expect to become good by practicing days?

People often spend more time on one puzzle trying to figure it out than they do in an entire game of chess, unlike real chess they always have a solution, and they often are from positions that never arise in the games you play.

That said, I do have one important caveat. I own the Alekhine chessable course, and in that course there is a tactics section. These are tactical positions that are likely to arise in the course of your games. This I found extremely helpful. I also have an old book that's all puzzles that have arisen from King's Gambit positions. This too I've found to be extremely helpful.

Random puzzles? Not so much.

"Chess is 99% tactics." (Or 95%, or whatever high%) -- I agree. However, it does not follow that the student should spend 99% of their time on tactics. And the top players did not get where they are by spending their time in this way.

"Games below xxxx Elo are decided by tactics." -- I disagree. The correct statement is that all games at all Elo are decided by tactics. The difference is that in a game between true beginners, the tactical errors appear on the board: hanging a piece, not taking a hanging piece, leaving a king in check, not noticing the enemy king is in check! Let's call this a "ply 0" error. Whereas in a game between top players, the tactical errors appear (mostly) in their heads. They leave a piece hanging at move 5 in some subvariation. (Although GMs probably don't leave the king in check at *any* depth of calculation, unless they are in extreme time pressure.)

By the way, even saying a game is "decided by tactics" automatically follows from our notion of "decided". Tactics is the transition from an advantageous position to a decisive position. Even a positional player who applies the boa constrictor squeeze needs tactics at the end to actually win the game. We shouldn't think the transitional phase at the end is the winning procedure. No, it was the positional play that caused the win, but even the greatest positional player in the world can't change the fact that chess is 99% tactics. But I digress. Back on topic.

The goal of tactical training is *not* to "remove tactical errors". It can't be done. The initial goal for beginners is simply to push the errors from ply 0 to ply 1. Teach them to sit on their hands, visualize the position that *would* occur after the move they want to make, and only make the move if it is safe. Once they learn that, if they sit down against a beginner who has not learned it, they will just mop up. But playing at ply 1 is still just terrible, they are making mistakes left and right, it's just that they infinitely better than before (1/0 = infinity). So there is the dead simple goal of tactical training -- to improve just enough so your opponent is making more tactical errors than you are.

We can say the exact same thing about openings, strategy, endgames, etc. The ideal amount to know is just enough so that you are making fewer mistakes than your opponents. But it's all on a sliding scale, and our results are largely determined by our biggest weakness. And weakness means not on an _absolute_ scale, but on a _relative_ scale vis-a-vis our opponents. For example, if I am a 1600 player who knows openings at a 2000 level, it is mathematically provable there is some aspect of my game that is well below 1600 level. And if you look at my games, I am probably doing well in the opening but doing (relatively) poorly somewhere else. It might even be tactics! This is why it's a mistake to study only tactics below any xxxx Elo. Knowing tactics at 2000 level is not enough to make me a 2000 player. I need to know *all* aspects of the game at 2000 level.

One thing I will say, though, is if you are going to make a mistake and over-emphasize some aspect of the game, probably the best thing to over-emphasize is tactics, and the worst thing to over-emphasize is openings.