Hacker News new | past | comments | ask | show | jobs | submit login
Training in chess: A scientific approach (2005) [pdf] (semanticscholar.org)
72 points by lainon on Jan 18, 2017 | hide | past | favorite | 11 comments



Cool Article.

I played a ton of chess in my high school / college years, progressing to just over 2000 USCF before taking a break from the tournament circuit. In my opinion, if youre not an expert / A - level player and want to progress rapidly, here is where you should focus:

1. Tactics, tactics, tactics - So many games under the expert level are decided by a tactical slip. A strong knowledge of opening or endgame theory means nada if you walk into a combination midgame and get mated or lose a ton of material. Work on exercises to train yourself to look for tactcial opportunities in game. The exercises on apps like LiChess and Chess.com are good places to start. Chess training pocket book by Lev Alburt is a handy reference too (although this isn't entirely focused on tactics)

2. Pick 1-2 openings as white and black and focus on them. Don't try to master a dozen openings. Focus not just on the moves, but the key ideas once the opening is over. Also, and the OP pointed this out, focus on the key ideas for both white and black, so you will have an idea of what your opponent is (or should be) trying to accomplish. There a bunch of Opening Explorer type apps that will let you save a few of your favorite openings and walk through the different variations. Practice not just going through the variations but playing and analyzing the positions these openings create.

3. Run your games through an analysis program or go over them with a stronger player than you. As I mentioned before, frequently when starting out you will win or lose games based on tactics. Having a program or stronger player point out these opportunities in YOUR OWN games is a very effective way to learn to recognize them in the future. Any time you play a close, hard fought game where you arent quite sure what went wrong, you should be looking to run that through an engine or review with a strong player to find the turning point(s)

4. Play. Just like best way learning to program is to program / build stuff, playing more games will give you that experience to quickly recognize and evaluate a position. Also, one of the keys learning to program is not getting frustrated when nothing works when you start out. Same principle applies to chess :)


I have a similar USCF rating and I agree with everything you said. I would also add, when buying chess books, get books of the form "My best games" of top players and go over them. Try the side variations in those books as much as possible in your head. Do not buy opening books. Instead go over master games whose common openings match yours.


Very good points.

5. Read "Chess Middlegame Planning" by Peter Romanovsky. There are a LOTS of books on openings and end-game, but very few on middlegame. Among them, this is probably the best one. Truly recommended.

[0] https://www.amazon.com/Chess-Middlegame-Planning-Peter-Roman...


This is extremely interesting. I've been searching for something similar for years.

Disclosure: I'm a class B player.


From your perspective, what was the most interesting and surprising thing discussed in this article?


This is very interesting. I learned chess almost totally by deconstruction (starting with a well worn copy of "Bobby Fisher Teaches Chess") and then practice. I was a passable chess player since I had an excellent memory, but got sort of turned off both by the community at the time (mostly a reflection of the chess players regionally, not in general) and then by feeling that the problem had been solved. It felt like a race of who could memorize the most games. This shows that a much more systemic approach is really occurring. It doesn't do anything for the feeling that chess is solved, however, since a computer can memorize more games than me.

That said, I feel it's interesting to see how much this reflects training for other sports. Playing soccer is as much about deconstructing older games, players, then systematically attempting to anticipate their actions and your responses in a variety of contexts (i.e. a search tree). Much of this is instinctive in good players, but the same process is taking place.


> In this chapter, we have in mind a player who has already mastered the basics of chess (with a level of, say, 1800 Elo)

Uggghhhh... 1800 is the "basics of chess" ?? Way to make me feel bad about my 1500 rank at Chess.com


The author was most likely talking about a FIDE rating. I think you'll find pretty much every on-line site has people rated well above their FIDE ratings. This is partly because people who pay money to join FIDE and pay money to play in tournaments take it much more seriously.


That just makes it worse, doesn't it?


1800 is _mastering_ the basics of chess. That seems about right to me.


Spellchecker would be good; findinds in the first paragraph, tut.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: