Analyze Your Games

With the new year came a new study plan. Besides reviewing master games, solving a lot of puzzles, and playing as much as I can, I will be focusing on analyzing all my lost and drawn over-the-board (OTB) games.

I have developed a decent habit of reviewing my slow games. For instance, in 2014 I played 112 OTB games and another 60 chess.com games. I reviewed 90 of those games in some way.

Many of the games, particularly from OTB tournaments, I reviewed with my chess coach, who I meet with for several hours a month. With him, we review why I made certain poor moves, what was the root cause of the moves, and then how to fix those root causes. For instance, maybe I didn't apply a chess principle to a middlegame position. Or, maybe I moved without considering all my opponent's (or my own) forcing moves.

I also analyze my games with chess friends. In these sessions, whether live or in an online forum, other people give me ideas about the games. These can be very useful ideas for me to explore later, even if the chess friends are not always significantly stronger players.

Finally, I spend a lot of time analyzing games by myself. Many trainers and grandmasters have the opinion that analyzing one's own games is the single most valuable thing one can do to improve. In the past when I have analyzed my own games I often ran the computer engine over it, looking for missed tactical shots and for positional moves where one side or the other lost a quarter point or more in the evaluation.

This is very helpful. But, unless it is a tactical shot, the computer can't tell us why our moves are bad. I can pinpoint the weak moves, though, and try to understand why they are weak and why I made the move, similar to how my chess coach quizzes me on my moves. But without his expertise to give me a definitive answer as to why, I have to make the best guess that I can.

Over the past few months, I've taken this a step further. I've been refraining from running the computer over the game. Well, that's not entirely true. I do tend to run the computer over the opening moves, when the moves leave my repertoire. At that point in my review, I look up the new moves in my opening books in order to add the correct move to my repertoire tree. If I cannot find the position in my books, I'll look in my master database and also quiz the computer to find out what I should have played and what I plan to play in the future.

But other than that opening work, I am currently tending to avoid turning the computer onto my games initially. I am taking my time and analyzing the game for an hour or two, writing comments and possible alternative candidate moves and variations. I also try to add my own evaluation of each variation I look at.

Then I sit on the game for a while. I might show it to some chess friends or my chess coach. Then, I'll go back to the game a few days or weeks later and see the game again from a fresh perspective and with some insight, hopefully, from at least one other player. I look for more moves and analyze the variations again, even deeper. I rethink the evaluations.

Then, if I feel like it, I'll turn the computer loose on the game and see what it says and compare how it's ideas match what I had already come up with. Are there any big surprises? Any tactical shots I missed? Why did I miss those after reviewing the game several times? Can I figure out why the computer's evaluations differ from mine? These questions now give me another reason to go over the game and try to integrate my previous ideas or misconceptions with the computer's "truth" about the position.

It is hard work, exhausting. It was so much easier in 2014 when the bulk of my training was mindlessly (yes, pretty much) going through thousands of tactical problems on Tactics Trainer. Now I have to sit and come up with ideas and ask myself tough questions, like, why was I thinking incorrectly, why didn't I apply a principle I knew about, or why did I fail to do a blunder check in a certain position.

But it is also rewarding to feel like I am searching for the truth in my games. IM Jeremy Silman writes in How to Reassess Your Chess 4th Edition, "When you're playing a game you should generally be able to lecture on your position at any given time."

What Silman means is that you should understand the ins and outs of each position that your game reaches and you should understand what the ideas are for each side in those positions. Then, you should only make moves that are logically congruent with those ideas.

When I am analyzing my own games, this is a great exercise. Now that I am off the clock so to speak I have an endless amount of time to understand the different positions and decide if my moves were logical given how I now understand those positions. It doesn't matter so much if my understanding is absolutely correct or deep, but given whatever my level of understanding is, can I come up with logical variations? Why didn't I during the game?

At least for the next few months I plan to continue this focus on deeply analyzing my games. I recommend you give it a try too.