Annotating games

A recent email:

I've never annotated a game. Could be interesting. Perhaps you could
send me a game and I'll try to annotate it without computer. Might show
you my thinking.

The games that are most valuable to annotate are your own games, (but maybe in future it might be a good exercise to look at somebody else's). I think it's a good discipline to look at all of your serious games at least briefly after the game with your opponent, then again at home with some software, and record your thoughts.

I'm happy for people to use a computer to reveal any tactical points. The aim is to get you to show yourself what you were thinking about during the game, what you think with hindsight, and, from any mismatch, see if there's anything you can learn from or smarten up in future.

Never annotated before? You do it how you like, but the way I do it is: Fritz looks over my shoulder as I trot through the moves of a game, giving me a running assessment of the balance of the position (+0.31,+0.2, +0.2, +0.3, 0.0, -0.87). If you plotted this assessment on a chart, it would wander up and down around 0.0, taking an occasional lurch as one side or the other missed something, maybe settling at a new level after a particularly classy or duff move. A good annotation would comment on the assessments while they are stable (The features of the position are... which is balanced or one side is better because...) and the lurches (This makes Black's position worse because... Here White overlooks...) and the mismatches (At the time I thought X and my opponent thought Y but really Z) and give some sort of overview of the game (White did this well, Black did that well but this not so well... I've made this mistake before, so... Lessons for next time might be...).

The point of doing an annotation for training purposes is to compare what happened with what should have happened (the computer can do) but also to compare what you were thinking with what you should have been thinking. [I've got a bunch of my own games with my own comments here if you want to see what sort of mess I make of this task.]

I usually feel free to ignore my computer when it disagrees with me about exactly which side of 0.0 we are, except when the difference starts to get to be the size of half a pawn, when I worry that I've mis-assessed the position. [Either it is or I am bad at positional sacrifices; I played a pawn sac against Ivor on Saturday which I thought was just killing, but my computer reckoned I was losing all the way through.]

And always remember, you're trying to find the turning points, and not just noting mistakes but trying to find improvements. While you're digging deeper, finding better moves for yourself or your opponent I have hopes that you're learning. When you're just describing what's happening and not pointing out better moves I worry that you're not getting the most out of the exercise.

Jacob Aagaard throws in a how-to guide for annotating your own games in his Excelling at Positional Chess. He describes a number of different levels of work, depending on your capacity for work:

Write down three things you have learned from the game.
Always write down the time you spent during the game
Write down the critical moments of the game, the things you saw during the game and what you think went wrong. Do this the same evening.
Analyse the game yourself. Only when finished should you refer to [computers].
Check for structural assistance [from your computer] to gain additional insight. "
[This is a database search based on pawn structure and maybe ECO codes, to find similar games and find out how other people handled the position from either side.]
"Tournament reports and diagnosis of weakness. ...Make a list of all my mistakes from my games, and describe them." [The idea is to find your typical mistakes, things that you often get wrong.]
Training based on tournament reports. ... For every weakness there is a remedy."