Learning through imitation: replay games to get stronger?

[tekesta]

A couple of days ago I had the following idea. Is it possible for a novice to improve at Go for the first 3 months just by replaying pro games in addition to playing actual games? Hold off on puzzles until after the 3 months have passed; by this time the novice will have acquired enough background information to be able to make educated guesses when solving tsumego and other Go puzzles.

[Bill Spight]

A couple of days ago I had the following idea. Is it possible for a novice to improve at Go for the first 3 months just by replaying pro games in addition to playing actual games? Hold off on puzzles until after the 3 months have passed; by this time the novice will have acquired enough background information to be able to make educated guesses when solving tsumego and other Go puzzles.

I don't know about replaying pro games, but I did not tackle go problems until I had been playing go for almost a year, by which time I was 4 kyu (maybe an AGA 3 kyu today, I dunno. Probably a Japanese 2 kyu now.) I had the English translation of Korschelt, which has problems in it, but they are above beginner level. Except for the first problem, which I did solve. (I checked, and I had marked it solved.) I did play over a few pro games, but since there was a single diagram per game, I found that tedious. I did study the endgame and opening diagrams, since they were not too crowded.

[RobertJasiek]

Probably, for a newbie, it can help a bit, but IMO it helps MUCH more if he gets club advice, reads good beginner books or takes lessons because these sources (should) know what is relevant for his understanding.

[leichtloeslich]

Say that playing a move happens in two stages:

• gather candidate moves
• read them out

(these stages of course aren't strictly seperated, for example reading out your first guess might lead you to finding a better first move, then read that out etc.)

While replaying pro games can certainly help with the first, I don't think it helps much with the second stage, which is the meat of your playing ability.
(It won't help a 20k much to know about fancy clamps if he puts himself into atari on every 3rd move.)

And the idea that you need pro-games to be able to solve tsumego is kind of outlandish.
I bet every dan-player has done a few ishi-no-shita problems, yet most haven't seen that sort of tesuji occur in any of the games they played/observed.

[John Fairbairn]

I think the point of playing over pro games is often missed. As with many other controversial things, too often it is dismissed and then reasons are sought to justify that dismissal. A more open-minded approach can be useful, although the guidance of a commentary is often also needed.

E.g.

(;AB[fc][pc][qd][qe][pe][pf][ic][qq][pq][np][nn][dq][cp][co][cm][hq]AW[ng][qp][pl]
[pn][dc][ce][oc][eq][en][do][od][lc][pg][qg][qf][dr]LB[dr:A]AP[MultiGo:4.2.1]SZ[19]

(;B[gq];W[fp];B[ho];W[cr];B[fr];W[er];B[om])
(;B[cr];W[dp];B[cq];W[ho]))

One reason for looking at pro games (or josekis) is to see moves that you would not normally think of, so that they swim into your consciousness in future. In the position above, the last move A usually elicits the reply shown in the variation (2), and in fact it can be called a knee-jerk sequence among amateurs.

Black here, Miyasaka Shinji 5-dan against Segoe Kensaku 6-dan, did not like the White cap that results after the usual sequence, so he played a very unusual alternative (1). At first I thought it was a misprint!

That move drew a comment from Honinbo Shusai. What do you think he said? It might be useful to remember that there was no komi then.

Shusai said: "For Black 33 [= 1], playing at 36 [= 2] as usual is more reasonable."

In those days (1925) Kido usually expanded Shusai's elliptical comments to tell us what he really meant. In this case they explained about White disliking the cap but said the result could not be deemed favourable to Black, and said that the more reasonable attitude for a Black player was to play in the usual way.

Even after the extra explanation, I am sure there will be many westerners, not just RJ, who feel frustrated that no-one has explained what is actually wrong with Black's result. But therein lies the point. The westerner, typically thinking analytically, wants to know how many chunks of meaty territory or how many croutons of thickness each side has acquired so he can make a static comparison. Orientals, while never eschewing static analysis where it can be useful, tend to look down on that approach in environments with too many unknowns, and prefer instead a more dynamic synthesis. They often refer to this as a time-space continuum, and it is best exemplified for many of them in the Book of Changes (Go Seigen's favourite book), but it is also exemplified by Kido's comment here. What is more important (given that the usual play does not give an obviously bad result) is Black's correct attitude, which has the great advantage that it can last for the whole game (time+space), rather than just one move (space). Clearly this can also be seen as a recipe, if misused, for slack play, but Shusai saw it, rather than crude counting, as the starting point, and as advice a Meijin thought worth passing onto a very high dan and senior pupil.

Whether or not you choose to accept such advice yourself, it should be thought-provoking, and you won't get that sort of stimulus from tsume-go or tesuji problems. You don't really get it from replaying pro games alone, of course - the extra comment seems essential here. But at least you would never get the chance to think about Black 1 at all unless you played over this game. White won BTW, and took away much of Black's lower moyo, an action which immediately prompted Black's resignation.

[RobertJasiek]

One reason for looking at pro games (or josekis) is to see moves that you would not normally think of, so that they swim into your consciousness in future.

Yes, but the beginner must become aware of the basic moves first.

not just RJ [...] would never get the chance to think about Black 1 at all unless you played over this game.

Eh, Black 1 is a kind of move I learnt by myself when I loved attacking regardless of how bad the shapes;)

instead a more dynamic

Both static and dynamic aspects should be considered when analysing a position. (Both are difficult for beginners, who can hardly learn them on their own just by replaying pro games.)

[Knotwilg]

When I started playing table tennis competitively 1,5 years ago, I also watched a lot of pro movies. It's fun, it's spectacular, it's inspiring.

A while ago a strong player watched me and said: "your serve looks like you copied it from a video or something, but there's not much going on". I had copied the body weight shift, some arm motion and a lot was going on in my head that didn't transfer into the mechanics (loads of spin - NOT, astute disguise - NOT).

It turned out there were some aspects that I was not copying at all: backswing, keeping the bat flat, wrist action, keeping the contact thin, using the tip of the bat for extra momentum ... all details of serve action that I either had not noticed, or thought I executed but didn't.

I needed to go back and train these aspects first and each one in isolation. Now my serves didn't look like professional serves at all. Now I was really training.

[snorri]

I guess there are people who think pro games obscure basic moves and those who think they exemplify them.

[oren]

A couple of days ago I had the following idea. Is it possible for a novice to improve at Go for the first 3 months just by replaying pro games in addition to playing actual games? Hold off on puzzles until after the 3 months have passed; by this time the novice will have acquired enough background information to be able to make educated guesses when solving tsumego and other Go puzzles.

I would say no. It would be incredibly boring, and not the most effective use of time.

[Bill Spight]

I guess there are people who think pro games obscure basic moves and those who think they exemplify them.

They do both. Most of a pro game is below the surface.

[Bill Spight]

As for learning by imitation, I think that a lot of people do not understand it or its power. After all, we all learned our native tongues by imitation. That is quite a feat! We just don't remember how we did it.

At a certain point what seems to be the most frequent word in a child's vocabulary. Why? Kids wonder about things. They ask why. The make guesses. That's all part of imitation (Edit: at least, when they wonder why by themselves). There is surface imitation, monkey see monkey do, but there is also deeper imitation, which generalizes and makes distinctions. Both are learning by observation, which is the definition of imitation.

[xed_over]

A couple of days ago I had the following idea. Is it possible for a novice to improve at Go for the first 3 months just by replaying pro games in addition to playing actual games? Hold off on puzzles until after the 3 months have passed; by this time the novice will have acquired enough background information to be able to make educated guesses when solving tsumego and other Go puzzles.

I would say no. It would be incredibly boring, and not the most effective use of time.

I disagree. I think it is possible. And I find it quite enjoyable. I claim that a significant portion of my early improvement was using this method.

[oren]

I disagree. I think it is possible. And I find it quite enjoyable. I claim that a significant portion of my early improvement was using this method.

Could you imagine telling your kids at a young age to play through three months of game recordings and only after that they could play a game?

I know playing capture go regularly with a kid, I couldn't see him wanting to do that for three months. I'm just a bit skeptical.

[Bill Spight]

Could you imagine telling your kids at a young age to play through three months of game recordings and only after that they could play a game?

Who is advocating that? Tekesta is only saying not to do problems for three months.

Is it possible for a novice to improve at Go for the first 3 months just by replaying pro games in addition to playing actual games?

[oren]

Could you imagine telling your kids at a young age to play through three months of game recordings and only after that they could play a game?

Who is advocating that? Tekesta is only saying not to do problems for three months.

Is it possible for a novice to improve at Go for the first 3 months just by replaying pro games in addition to playing actual games?

Sorry, then, I misread. I'm guessing most novices just do the play games and never touch problems.