Have we been duped by AlphaGo?

[Bantari]

The only thing I have is my gut feeling that the 3 hours are not enough and modern time controls are too short.

Yes. And my feeling is that references here on the forum to "Mickey Mouse time limits", etc., may contribute to your gut feeling

No, they do not.
But common sense does.

[Bantari]

Otherwise, you could apply the same strategy in evenly-timed games.

You can still do this, since people manage their time differently (e.g. Go/Kitani).

Sure, but still the underlying idea is that one side has less time left on the clock than the other.
And so in complicated situations, the side with less time will likely make more mistakes.
So the side with more time creates complications to benefit from the time difference.

Or am I missing something and this kind of "strategy" works completely differently.

[Bantari]

But the underlying assumption of such strategy would be that the player with shorter time would make mistakes which he/she would not make if the times were equal. Q.E.D.

What was to be demonstrated to me is that such mistakes impacted game quality in some non-tautological way (where we don't just define game quality to be exactly the inverse of the number of mistakes^*). For sure, this is my general disagreement.

I think it is not just the number of mistakes but also their seriousness. If you make 5 big blunders in a fast game, and 5 very subtle mistakes in slow game, I would classify the slower game to be of higher quality. But you are right - we might want to define it first.

Still, this is not a mathematical thesis that we are writing, just a discussion about personal opinions.

The premise is that, in general, given shorter time limits a player will make more mistakes, or more severe mistakes, or both. We already admitted that there might be certain limits beyond which this does not apply. For example: the game can be of the same quality if the time is 2 sec or 4 sec. On the other end, extending time limits from 60 hours to 80 hours might not matter neither.

I am not sure how you see it, to be honest.
What I am claiming that, taking it to the extreme, a 2 min game will be of lower quality than 2 hour game.
If you aagree with that, then we are already in agreement about the principle, we just have to agree on the exact tipping points for the time-vs-quality curve.

If you disagree, I would love to hear your rationale for that.

[Uberdude]

If pros didn't think they could play better with more time they wouldn't play time tesujis (playing a sente move, wasting a ko threat so as not to use up a byo yomi period but basically get double the time to think about the important move).

[Kirby]

The only thing I have is my gut feeling that the 3 hours are not enough and modern time controls are too short.

Yes. And my feeling is that references here on the forum to "Mickey Mouse time limits", etc., may contribute to your gut feeling

No, they do not.
But common sense does.

Don't know how it's "common sense" that 3 hours isn't enough time for a quality game, but ok.

[erislover]

The premise is that, in general, given shorter time limits a player will make more mistakes, or more severe mistakes, or both. We already admitted that there might be certain limits beyond which this does not apply. For example: the game can be of the same quality if the time is 2 sec or 4 sec. On the other end, extending time limits from 60 hours to 80 hours might not matter neither.

I am not sure how you see it, to be honest.

That the quantity of mistakes is not correlated with game quality. If it were so, I would never have made it beyond 30kyu.

[seigenblues]

I find it astounding how much this thread is discussing whether or not the *human* will play better, and very little is spent on determining if the *computer* will play better. Whether a pro plays slightly stronger or not has to be balanced by whether the computer will play equally, more, or less strongly in the same time period.

I will argue that a human might play slightly better given a day vs. 10s, but a computer will play SUBSTANTIALLY better -- enough that a reasonably fair matchup will become increasingly lopsided in favor of the computer as its given longer and longer to think.

I think it's probably not a stretch to say that it will absolutely play better, given longer time. There are some diminishing returns, but it will *absolutely* play better.

My reasoning is as follows: A purely random MCTS can be proven to (incredibly slowly) converge on the right answer. the "incredibly slowly" is so incredibly slow, that programmers try to speed it up by giving it pattern-based playouts. Essentially, programmers trade off the certainty that it won't miss any successful lines for convergence speed. One has to only 'watch' an AI like pachi work through its initial terrible moves until it finds a decent one to realize that the vaunted speed of computers is *not* at the 1s-10s timescale, and that sufficient playouts are required for a computer to converge on a 'good' move.

humans do an incredible amount of pruning right off the bat.

I would absolutely bet on LSD in a 5s game vs AG. (or at least, the AG of 3 months ago

[Kirby]

I find it astounding how much this thread is discussing whether or not the *human* will play better, and very little is spent on determining if the *computer* will play better.

You underestimate our lack of focus! The real question is, given sufficient thinking time, can we produce higher quality posts?

My bet is that computers can do better.

[dfan]

I find it astounding how much this thread is discussing whether or not the *human* will play better, and very little is spent on determining if the *computer* will play better.

That's because the question of whether humans play better when given more time is a highly charged one on this forum, so when the topic comes up it tends to dominate the discussion.

[xed_over]

I think it's probably not a stretch to say that it will absolutely play better, given longer time. There are some diminishing returns, but it will *absolutely* play better.

I thought I suggested that earlier in this thread (or was it a different thread), and got shot down.

[Bantari]

Yes. And my feeling is that references here on the forum to "Mickey Mouse time limits", etc., may contribute to your gut feeling

No, they do not.
But common sense does.

Don't know how it's "common sense" that 3 hours isn't enough time for a quality game, but ok.

Again, you persist arguing right after we agreed on something.

Its not the point that 3 hours is too short for quality game. Pros play quality go on most time limits - I would even call a half hour pro game a quality game, especially compared to my play. The point is that 6 hour game *might* help create a game of even higher quality. I am really not sure why you are so vehemently against this idea.

You really cannot imagine that, given 3 extra hours of thinking time, a pro might be able to use this extra time to produce better and deeper moves?
Well, ok then. We just have to disagree on that.

[Bantari]

The premise is that, in general, given shorter time limits a player will make more mistakes, or more severe mistakes, or both. We already admitted that there might be certain limits beyond which this does not apply. For example: the game can be of the same quality if the time is 2 sec or 4 sec. On the other end, extending time limits from 60 hours to 80 hours might not matter neither.

I am not sure how you see it, to be honest.

That the quantity of mistakes is not correlated with game quality. If it were so, I would never have made it beyond 30kyu.

Really? So you would call a game with plenty of mistakes and a game without any mistakes to be of the same quality?

I do not try to be mean here, but have you made it beyond 30 kyu?
If so, have you accoplished this my lowering the amount and severity of your mistakes? Or have you become stronger by making more and more severe mistakes? And when you improve, is that by trying to make more mistakes or less mistakes?

This really starts to feel like a Fellini movie here...

[erislover]

Really? So you would call a game with plenty of mistakes and a game without any mistakes to be of the same quality?

Sometimes, sure. I enjoy the game of go. Moves may be bad for boring reasons, or really interesting reasons. Mistakes just have little to do with quality.

If so, have you accoplished this my lowering the amount and severity of your mistakes?

Yes, I have grown stronger by making fewer mistakes, or at least, fewer mistakes of certain kinds. But some majority of my games are such that I found them to be quality games, or else I would not continue playing. If I disliked all my games because of my mistakes, why would I continue to play for all these years? If you don't like pickles, you don't eat pickles. I mean, if micky-mouse-timing is causing a drop in the quality of professional games, I have to wonder what enjoyment amateurs possibly could get of their own games.

[Kirby]

The point is that 6 hour game *might* help create a game of even higher quality. I am really not sure why you are so vehemently against this idea.

Again, I am not against the idea - especially when you say "might". What I am against is the automatic assumption that this must be true, and the implication that modern games are full of mistakes. I've changed my avatar to Mickey Mouse to try to get over the name calling.

Please remember: I am not dissing 6 hour games, nor am I calling those time limits "Mickey Mouse time limits". I am not fighting so much for the superiority of 3 hour games so much as fighting against their supposed inferiority.

[Kirby]

I received an email from David Ormerod of GoGameGuru today, which may be more relevant to the specifics of the time settings used in the AlphaGo vs. Lee Sedol match. He is a little busy with his baby at the moment, but indicated that it would be alright to quote him.

It appears that Younggil and David were concerned back in February that the members of the Korean camp weren't taking the match seriously enough, and might not be briefing Lee Sedol as well as they could. They contacted Lee Sedol's sister, Sena (she used to live in Sydney, incidentally), and requested that she pass on information about computer Go and to have Sedol consult with a computer expert. David indicated that Sena later explained the reasoning that Lee Sedol had in selecting the two hour time limit.

Here's a quote from David's email to me:

Anyway, I don't think anyone has mentioned that longer time in a human vs computer match fatigues the human, but not the computer. This was a key factor in Lee's decision to choose 2 hours each instead of a longer time limit. He was worried that with both players using 3 hours or more he would get too tired and the computer would gain an advantage -- as you know, the match with Gu was 4 hours. He may have miscalculated this, but he thought he could play well enough with 2 hours and not get too tired.

So while it is possible that Lee Sedol miscalculated the optimal time settings for the match, it appears that he believes that fatigue would affect the quality of his play, whereas it would not affect the computer. With a longer time setting, like with his match against Gu, presumably he felt that his game quality would be worse with the increased time.

Of course, this is not any sort of definitive proof as to the best time setting, but it at least provides some evidence for this theory:


If nothing else, I'd think that we can see that game quality is not so simple to analyze, just by looking at the time limits alone.