Thursday, January 08, 2009
The Long Answer
In the comments to my last post, Tempo asked me what my shortcomings are that prevent further improvement, and I replied: "The short answer is that it probably poor selection of candidates and/or limited ability to visualize positions 5 or 6 ply down a forcing sequence. That's why I'm trying the Hertan and Dvoretsky books. The long answer will be a blog post in a day or two." I guess this is that post.
The long answer is that every part of my game seems weak, to the extent that everything I've done to improve has resulted in finding new ways to lose.
In a post last March that I've referred to recently, I listed all the things I did in the last couple of years. I concluded, "a hundred point difference could be chance", and I have begun to think that it was. Also, you could argue that it wasn't a hundred points, either--my rating only moved 50 points, the other 50 was a combination of bad performance when my rating was crashing down and good performance when it was recovering.
I worked my opening repertoire, partly because I'd noticed that in games where I used a lot of time in the first 10 moves, I did poorly even when I got a good position. That "worked"--I rarely get out of book early now, and when I do it's just as likely that it's because I'm refuting bad play by my opponent as forgetting my lines, and so my score in those games is just as good as my score in games where I don't use up a lot of time in the opening. But my rating overall doesn't move.
I did a lot of work on tactics, and I can see that I make fewer outright blunders than I did. By my rating doesn't move--I miss tactics several moves down a variation instead in the first couple of moves.
I worked on visualizing positions, and I can play through games blindfold pretty well--but I don't seem to see the possible continuations in the positions I visualize.
And so on. I do things to improve certain perceived problems, and those problems get a but better--and something else causes me to lose anyway, so my rating doesn't move.
Maybe I'm just grumpy. Maybe having a rating 50 points higher than a couple of years back indicates a real (but slow) improvement.
Anyway. I'll be working on visualization and especially on candidate selection, because those seem like obvious problems. I can point at places in my games where I didn't consider the strongest move at all, especially a few moves down the tree of variations. We'll see how that works.
You can't see what's in your blind spot, so you have to look for circumstancial evidence to find out what's there.
The times that I did something that was quite alien to me, I learned the most and improved the most.
Since you are not able find it yourself, the input must come from outside. Only somebody else can bring something forward that isn't obvious for you. So which author writes the most incomprehensible in your eyes? :)
I did notice an odd suggestion in Hertan's book--he recommends solving composed mates-in-two. I've never heard anyone else claim that would help your game, and a number of writers say composed problems are worthless for improvement.
Since it seems like a strange idea, I'll try it. Of course, with my Gd-given powers of rationalization, I can already come up with an argument for why it might help...