Best Way to Study GM Games for Improvement - Your Counsel, Please?

Hi All,

I ran across this blog comment by Urk from a long ago thread:

"...  the most important training technique, which is this:

USE A PHYSICAL CHESS SET WITH REAL PIECES (preferably wooden) TO PLAY OVER HUNDREDS OF COMPLETE MASTER GAMES FROM BOOKS AND MAGAZINES.

This will burn the coordinate squares into your brain so that you can speak our chess language.

It also gives you openings, middlegames, endgames, tactics, and checkmates.

Play over lots of master level games at a rapid pace without dwelling on lots of sub-variations, and make sure you get a lot of old games 1850-1950, since they're much more clear and understandable than modern struggles between 2800 titans.

This is my old school approach."

This spurs the following question(s).  I'm an old guy, and I remember back in the day that I was told to get an index card and cover up the moves after the opening has been played, and to "guess the move" by the master.  This took a heckuva long time to go through a game, but the counsel was that it was a great way to go over the game and improve your chess skills.

Q:  But if you do it this way, there's no way you can go over lots of GM games.  So what's better?  Long way (cover up and guess the move) or just go over lots of games and things come in through "osmosis"?

Q:  A lot of advice given to absorbing chess pattern recognition.  This argues for quantity, I would think.  If so, then wouldn't it be better to click through lots of master games in a 2D format, and thus forget the slow, laborious "guess the GM move"?

Q:  Or should a student do a blend of the two?  If so, what's a recommended blend, roughly speaking?

Thanks in advance.

                                                                        __________________________

We have talked about this before.  If you have purchased ChessBase, the  ChessBase games data base which contains 5 million + games which very likely will contain the game you are studying.  The game will be built into your opening tree. Your training sessions in the opening and  middle game visualization pattern memory banks will bring that particular game into your repetition exercises, if  it  contains a valuable position that you need to  build into  your visualization pattern memory bank.  

hello

Thanks FoF.   Quantity is better, you're saying.  Do you ever do "predict the next move"?  Just curious.

                                                                               _______________________

Are you talking about on Chess24 or the feature on ChessBase?

Actually neither, lol.   I really meant covering up the part of a page in a chess book with an index card.  I'm a dinosaur!

Studying with chessbase is much easier, but that doesn't necessarily mean it is better.

If you use a book and guess the move, especially going down side variations, you have to physically move the pieces and then remember the moves to go back to the start of the variation.  If you forget, oh well, you replay the game again and get the ideas reinforced.  If you remember, even better, as you're actively training your memory.

When I study a game with a book, I often end up memorizing it by the end, just because of physically moving the pieces; I don't mean to, it just happens.  I also take longer and get more out of it.  With Chessbase, I tend to go faster, forget faster and rarely if ever memorize games, intentionally or otherwise.  The temptation to just use the arrow keys and zip to the next move is very strong.

Personally, for what it is worth, the bulk of my improvement came from studying books and guessing moves, not using the computer for anything but blunder checking afterwards.

Remember SeniorPatzer:

With ChessBase the software will build your opening tree in 2-3 years.  Without software it will take you 8-10 years to build your opening tree by hand.

"Personally, for what it is worth, the bulk of my improvement came from studying books and guessing moves, not using the computer for anything but blunder checking afterwards."

Thanks SmithQ.  It's a counterpoint to FoF's thesis.  Hence, I asked if there was a blend, optimal blend, between "predict the next move" and rapidly going over GM games in 2D in 3-5 to acquire pattern recognition.  

I dunno, 50/50 time spent between "Predict the Move" and Rapidly Going over GM games.

guess the move is best, looking at games at a fast pace reminds me of watching rain drops fall down a window, passive and pointless

Yeah, that's an odd language thing. "Master" can mean "the best" so when someone says "study the masters" they might mean "study the super GMs"

There is definitely a time to go through some games quickly.  If you're learning a new opening, for instance, you can flick through a half-dozens very quickly.  Pause maybe 5sec on each move, see what your intuition says, and then advance.  You can learn a lot this way about general motifs and plans, especially if it is all the same opening.  Using a modern database is essential here, and I wouldn't want to do it any other way.

However, I believe there's a difference between gaining knowledge and gaining skill.  Going through games quickly, seeing plans and move orders, that's knowledge.  Training yourself to find the best move in any position?  That's a skill, and it takes something different.  Going through a thousand games quickly won't help you here, but spend one hour going through one game, that will help a lot.

I once took about an hour per game for a dozen Botvinnik games (over a few weeks).

I didn't feel like it helped me at all. Just saying.

Maybe if I were a GM I could get something out of it, but it seems to me (so take it with a grain of salt, I'm no expert trainer or anything) that going over these games quickly is a lot better. You want the general patterns and ideas, not the specifics (unless something happens to catch your eye).

I feel the same way.  I respect everyone's opinion here, but I think my chess skills would have been greatly enhanced if I had the access to non-master games and tactics that's available today.

I've played correspondence games at a 2100 level since the pre-computer 1970's but am awful at short time limits and I think a lot of it is due to the fact that most of my early studies were Bobby Fischer's 60 Memorable Games, and books by or about Alekhine, Capablance, Keres, Tal, etc. and they tend to be cautious games and you don't learn how to take advantage of non-master mistakes.

Only as I've returned to chess this year have I realized how much better I need to see tactics, how to recognize patterns, etc.  Also, there are now books that address errors in thinking processes that use non-master games like Dan Heisman's The World Most Instructive Amateur Game Book and Jeremy Silman's The Amateur's Mind.

Half of the comments here are about building an extensive personal openings database - aimed at a 1358 rated player!

I coached a usually-county-champ-top-five-in-state high school team and when we would have a NM come to talk to our mostly 900-1400 USCF OTB standard rated players, he would say about the openings that while the basic moves and ideas behind several openings is necessary, he does not memorize long lines and that all he needs out of the opening is to get to a playable middlegame.  When I play, I'm usually out of the book by the 7th move.  The last two times I played the Caro-Kann, I was out by moves 2 and 3.

The middlegame is where you need to be able to figure out what to do with the initiative or how to gain the initiative, where you are strong enough to attack, etc.  If you're much less than, at least, an 1800 player, you need to study tactics and strategy over a larger range than you'll see in grandmaster's games alone.  Those of us with stronger Endgame skills are usually not afraid to trade-off pieces when we're a pawn ahead.

THAT, in my opinion, what will get your rating up faster than the masters' games.

I think he meant >2600 not <2600

I think that's why Urk was saying to study games 1850-1950. First of all the best players back then had very clear and logical styles. None of the modern dynamic stuff trying to confuse matters. Some stuck to their principals to a fault, like Nimzo vs Tarrasch.

And of course some of the people Morphy played against were basically beginners.

I definitely don't recommend studying modern GM games, since at your level, there will be plenty of tactics for you by just studying games of old masters.

At the same time, remember that the "dynamic stuff" and heavy tactics that are in today's games, largely due to engines, is very very important.