Annotating move times

[SmoothOper]

Why don't the pro records annotate move times? I think these would be somewhat interesting and would add insight into how they are thinking.

[oren]

For important title games in Japan they do.

[Phoenix]

I loved watching 'live' pro tournament and title matches on IGS. I would go nuts trying to figure out what they were spending so much time thinking about! Sometimes they would slowly play into a joseki, then halfway through a straight variation, one of them would stop to think for 15-20 minutes, then play the joseki move!

I find it fascinating to watch where they play very fast and where they really take their time. It's helped me identify key parts of my own games and manage my time in a more logical and effective way.

That being said, I don't think I could put myself through watching one of Cho Chikun's games live.

[SmoothOper]

I loved watching 'live' pro tournament and title matches on IGS. I would go nuts trying to figure out what they were spending so much time thinking about! Sometimes they would slowly play into a joseki, then halfway through a straight variation, one of them would stop to think for 15-20 minutes, then play the joseki move!

I find it fascinating to watch where they play very fast and where they really take their time. It's helped me identify key parts of my own games and manage my time in a more logical and effective way.

I just want to know where they spent their time, though the occasional live feed is interesting.

That being said, I don't think I could put myself through watching one of Cho Chikun's games live.

This is exactly why I want to know, because flipping through Cho Chikun's(and also Go Seigen) games is kind of like watching a grasshopper on a hot skillet. He is all over the place, but in reality they are pretty slow. I think knowing where they spend there time thinking would give insight into which moves are interesting and difficult, where there are potential options.

[snorri]

The best thing is to sit next to the pro while they are playing, so if you get a chance to be a game recorder, take it.

2nd best may be watching video broadcasts on Baduk TV or the like.

In both of those cases, you can at least see when the players seem to be thinking and when they are, for example, going to the bathroom. (Not that the players aren't thinking when they are in the bathroom. I've figured out some of my best moves in the bathroom.)

It's actually amazing to watch a top pro rifle through the bowl, pick up a stone, almost play the stone, put it back in the bowl, then repeat 3 times before playing. It makes me feel they are human after all. Well, at least until they play something that shows they are closer to freaks of nature.

I like watching games that are closer to the time limits I play myself in tournaments or online. That way the time management is more relevant.

I feel one has to take those hints with a grain of salt, though. I know I can't play a ko fight or the endgame as fast as a much stronger player. Trying to emulate that is foolish. I also know that I can't make as good a use of my time in the opening. Unless I'm estimating the score, the time I can concentrate on a single move is limited. I think maybe I've played too much blitz and I have a hard time planting my butt down for more than 5 minutes in a critical situation. Somewhere around 2-3 minutes I lose track of what I'm trying to read out and often start over. The same thing happens with go problems. If I can't solve one in 2-3 minutes, the chances of my solving it with more time are pretty slim. I have a modest goal of getting to the point where in critical positions I can think for 5 minutes productively. That will do for the time limits I typically play. Any suggestions would be appreciated.

[oren]

2nd best may be watching video broadcasts on Baduk TV or the like.

I'm not sure that works so well since almost every game on BadukTV is played in byoyomi right away. Many of the big matches with some main time air delayed and get fast forwarded to live with commentary. You can see where they use their cards to get some extra time, but it's not like Japanese title matches with an hour or two on a single move. Now with Niconico you can see where they spend it during Honinbo. I have some books with title matches and they don't show time for every move but do show time used after a sequence.

[Bantari]

Not sure, but do the current servers/software capture time when you enter/play game? This should be a really simple thing to do programmatically, and this data can then be used if desired. Might be a nice feature to implement in the next server or sgf editor.

[xed_over]

Not sure, but do the current servers/software capture time when you enter/play game? This should be a really simple thing to do programmatically, and this data can then be used if desired. Might be a nice feature to implement in the next server or sgf editor.

KGS does.

[oren]

KGS does.

During main time. I thought it would be interesting to know how much time is used in byoyomi. With pros, it's less interesting since as you know, they often wait until the last second.

[John Fairbairn]

Why don't the pro records annotate move times? I think these would be somewhat interesting and would add insight into how they are thinking.

Many aspects here.

1. Only games with recorders have move times recorded. Typically these are title matches. Such games are rarely if ever presented with all the move times, but long ponders are usually remarked on. There have, however, been cases, such as Hiseikai games, where every minute of every move has been published. Most games have been recorded by pros from memory, which explains not just the lack of times but also why often two different records appear.

2. In modern Mickey Mouse games, even though the use of computers makes recording time easy, it hardly tells you anything significant whether a pro took 10 seconds or 8 seconds.

3. Even in games with long time allowances, the information can be deceptive. The shogi champion Yonenaga Kunio once told me of a case where he sat for over an hour before making a move. When the reporter later reverently asked him what deep schemes he had been thinking about, he said, "Nothing. It's just that there's an umbrella stuck in the middle of the garden out there and I was wondering how the hell it came to be there." There is also some manoeuvering when it comes to close to deciding who gets to seal a move, and a pro will often play slowly after a good lunch to give blood time to get back to his brain. There have even been cases of nervous pros who spend a long time on frequent visits to the loo. In older games with long time limits it was also common for some players to try to win on stamina rather than go technique (usually because final sessions continue through the night without adjournment).

4. All that said, in cases where long thought is genuine, in my experience it is nearly always the case that the pro sees the move he wants to play almost instantly. The extra time goes not on discovery but on checking and evaluating. However, in a Go-Karigane game I have just done, K spent over 5 hours of his 14 hours allotment on just three moves and ended up playing ordinary moves. He had been looking for brilliancies in a difficult situation but failed.

Personally I've found the only consistently useful time data to be the total times used. On that basis you can see that Go Seigen nearly always plays much faster than his opponent, and on the rare occasions he matches or exceeds them, you can be sure he was in trouble in that game. But the commentary tells you that anyway.

[Kirby]

\
2. In modern Mickey Mouse games, even though the use of computers makes recording time easy, it hardly tells you anything significant whether a pro took 10 seconds or 8 seconds.
...

Exaggeration, Mr. MickyMouse! There are many modern pro games that aren't this fast. I've watched them.

[Bantari]

Personally I've found the only consistently useful time data to be the total times used. On that basis you can see that Go Seigen nearly always plays much faster than his opponent, and on the rare occasions he matches or exceeds them, you can be sure he was in trouble in that game. But the commentary tells you that anyway.

Not sure I agree with most of what you write there, for various reasons. The main of them being that: unless we assume the pros use their time willy-nilly and with total disregard, the longer intervals when they think more do tell us something, regardless of how seemingly 'ordinary' the moves they make are.

But this aside, it all reminds me of what I read once about chess - the research into Fisher, if memory serves. It concerned something called 'time graphs' and from what I understand was the diagram of the time players spent thinking about their moves. I think the consensus was that for the very strongest players at the time (Fisher, Karpov, and then Kasparov) there were certain similarities in these graphs which were missing for weaker players. The whole idea was controversial at the time, and I have no clue where the research is now, but still... interesting.

Back to Go pros - if nothing else, I would find it interesting if there are certain types of positions in which pros tend to spend more time thinking... and if so, is it a general trend, or does it vary from pro to pro. Or is there no pattern whatsoever. Again - regardless of what moves they end up making in the end.

And this would also be interesting in the MickyMouse games - time in such games being an even more precious commodity, so its usage might be more telling. Or not.

I wonder if anybody ever looked into that in Go.

[hyperpape]

Everything John says fits with what I have read about highly talented people in various intellectual fields: 16 hours of concentrated focus over two days is just hard, and may even be counter-productive (for example: http://www.johndcook.com/blog/2012/08/27/work-or-rest/). The ability to manage your time over a long match so that you get your best thinking done when you need it is probably a skill, but we shouldn't think that means spending every second in deep thought.

[SmoothOper]

Without the data it's mostly hypothesis, however one could conjecture that longer times would be more difficult to assess where or what they were/thinking about. If you were to analyze them they might even be thrown out as outliers.