Tips for self-analysis?

terraform · #1

Hey guys, didn't really know where to post this question, but this seems like as good a place as any. =) I was wondering if anyone had any tips for reviewing your games when you don't have time directly after you play, so you have to come back to it after a few hours. I should note that I like to write my notes down while I stare mindlessly at the game record. I also play it over on the board from the record, but I spend most of my time away from my board with a notebook, so the aforementioned notetaking is my best mode of study.

What I do currently:

-Analyze the flow of the game during different segments and try to keep a running count of area, then try to figure out why the flow changed when it did
-For some moves, try to find 2-3 alternates, and engage in a discussion about the pros/cons/intent of each and pretend I know what I'm doing
-For complicated fights that I lose, carefully keep track of liberties and see where I went wrong
-For sacrifices, count the score prior to, after, and along the alternate route for the sacrifice
-Face palm ridiculously stupid moves that seemed okay during the game
-For every move, question whether to tenuki or respond

Some specific questions: for point 1, what moves should I consider alternates for; what aspects of the game did I miss?

Joaz Banbeck · #2

I gotta say that I admire your determination, but the path of improvement purely by self-analysis is a long and arduous one. Either you have to be an intuitive genius and see the underlying nature of your mistakes ( "Ahhh...32 moves ago I played too low, that's why my group died! The next 31 moves were flawless." ) or you have to brute force the whole issue ( for which the human brain is too slow by a dozen orders of magnitude )
You can, of course, do self-analysis for local situations ( maybe a joseki where the relevant moves are something like 3**5 ) but for the whole game it is just not practical.

It is far better to get advice from stronger players. Post a game in the 'game analysis' sub-forum of the 'improve your game' forum. There you will get the condensed wisdom of past geniuses, and a few good guesses from the rest of us.

Dusk Eagle · #3

In no particular order, here's a non-exhaustive list of stuff I focus on when reviewing one of my own games:

Moves where I was uncertain where to play during the game.
Sequences that went poorly for me.
Moves where I had sente but stupidly refused to count during the game to know what to do with it. (This may be beyond a 15k level, but it couldn't hurt to try to count to see who's ahead at some point during your review, if not during your game itself).
Opponent's shapes that seemed to have aji which I didn't exploit well, and what I could have done to use that aji better.
Areas of my own which had too much bad aji, and how I could've avoided those shapes.

Edit:

Quote:

( "Ahhh...32 moves ago I played too low, that's why my group died! The next 31 moves were flawless." )

This is a bit ridiculous. If you end up with a problem with one of your groups, you can probably find times where you could've played better to avoid it. If you did end up playing too low one move, it seems highly unlikely you could play 31 flawless moves in a row (=62 moves in-game) and end up dying (unless you are a professional I guess). If it's not immediately obvious that one move or one neglected move cost you severely, then I would wager that there's more than one bad move which when combined together has given you problems. Being able to spot at least a couple of these bad moves during a self-review shouldn't be hard.

I have never reviewed a game of mine which I lost and not found at least one obvious flaw in my play. I actually think that if you're not willing to at least attempt to self-review a game, you won't get much from having that game reviewed by others. You need to put effort in in order to improve.

terraform · #4

I try to do what Dusk Eagle mentioned. If I can find one flaw that I know is a flaw, I'm happy with the review, and the rest is purely exercise. Since I've started, I've noticed that a few of the same mistakes keep showing up in my reviews, and I find myself going, "Whoa, hold on, remember what happened last time?" during games more recently.

In reviews I can also think of crazy, ridiculous, ludicrous moves and not be punished for them. =) If one happens to work and I see the situation come up in a game, I can go, "I think I've seen this before... I think I did something like this..."

'Course I'm wrong 99% of the time. But I think a 1% improvement is worth the work.

You gave me some ideas about my own reviews Dusk, thanks. I also didn't think about aji. I don't know much about that type of... property (?) so I've got a lot of studying to do in order to study better. Mwahaha.

@Joaz: I agree that fully analyzing an entire game isn't at all helpful in its entirety, but I feel like there's something to be had. For now there's also the plus of having something to do when I get bored at work or the professor is droning on about "what is art". Anywho, I do plan on asking for review help from L19 members, but only after I've contributed to some reviews myself. =) I usually submit to GTL but I haven't had games that were super close recently, like the FAQ says.

Thanks to you both for your input! It's really helped to hear other people's views on this kind of thing.

Mark356 · #5

I usually go over games right after I play them, but I actually prefer going over them a few hours after. Right after, you're still too close to it, and the moves you played or thought of still seem like the only moves. Letting it sit for a few hours is kind of like backing away from the canvas and looking at the painting from across the room.

When I look over a game, I more or less do what Dusk Eagle does too. If you lose a group, usually you can find some good reasons for it. (Yeah, sometimes it is because I played too low 30 moves before.) You can usually find the point at which it got cut off or when its eyespace got destroyed. In fact, I also like to look for my opponent's mistakes, too: could they have offered more resistance at this point? Often they let me make a big kill, and if that happens I like to go over it and see if they had any choice in the matter. I also like to apply the same process to other parts of the game. Say my opponent has a wall that they convert into 40 points of territory. Why did I let my opponent get a wall like that in the first place, and did I get adequate compensation? If not, I go over the line of play that led to that wall from the beginning again and see what my other options were.

snorri · #6

I think it may be too amibitious to try to self-analyze to that degree at your level. (Or maybe at any level. Even pros analyze in groups.)

When I get may games reviewed by a stonger player, they always find some moves that I wouldn't have considered ever. Some moves I thought were fine in my self-review, they will consider terrible, and others that I worry about they will say are not mistakes.

Guo Juan has said that you should only try to learn a couple of things each game, anyway. So maybe go through the game and pick something that you'd like to study more. Maybe there's a joseki that you're always messing up or one you wanted to play but were too timid to try. Learn more about it. There could a life-and-death position and you made a mistake that was beneath your level, so you might consider playing through the variations on it a few times to make sure you don't make that same mistake again.

But I wouldn't spend a lot of time studying positions you don't understand without the help of a stronger player. That doesn't lead anywhere. When you're analyzing by yourself, you've just playing against yourself, and that's not sufficient resistance to improve.

Even if you find a few things that you're sure about and can understand, you have to honest with yourself about whether you are going to remember it all. It may be best to focus on a small number of things you have a really good chance of remembering and being able to use again in future games.

daal · #7

While the suggestions to consult a stronger player who can recognize your blind spots certainly make sense, making the effort to analyze one's own games with such depth and intensity can't be bad.

Chew Terr · #8

I highly recommend self-analysis. As a matter of fact, if you self-analyze before you have your games analyzed by someone else, you'll likely get more out of the lesson. In the lessons I take, I was told to pick a few games to go over, and find a few mistakes for myself before the lesson itself. This way, you already know the obvious things, so the reviewer's time can be better spent. Additionally, you're more familiar with the game, so corrections more easily stick in your mind. Good job, keep it up.

Andd · #9

Just a small tip, when reviewing by yourself it's easy to get into a click frenzy, and do all let an sgf editor do all your reading for you. My experience is that my best reviews are situational. I think it is more beneficial to review only area of the game where you feel you faltered at a time. Look for alternative moves, count the score, evaluate influence, but most importantly, read for yourself. Through self review it is hard to teach ourselves things we do not know yet, but it is perfect and I think extremely beneficial as practice for training mindset during games.

Tips for self-analysis?

Who is online