Have we been duped by AlphaGo?

[Bill Spight]

I don't endorse a pessimistic line on such matters

Oh, I'm not pessimistic. Not at all.

[xed_over]

But one of the reasons I keep responding to this thread (other than my stubbornness) is that I get a little angry and worked up with the tone regarding modern time limits. Calling them "Mickey Mouse" time limits, and "full of mistakes", and things like that seem to belittle something that I respect.

I think its perfectly fine to express your views on the subject, for as long as it takes to get your points across (within reason, of course), even though there may never be a consensus.

First, let me ask, just to be clear, what it is specifically that you respect: the shorter time limits? or the great game of baduk in general? or maybe its Mickey Mouse himself?

Because I believe Mr. Fairbairn has great respect for the game of Go as well. But it sounds to me like he is offended by the shorter time limits, believing those to be ruining the game he respects so much (forgive me if I'm putting words in his mouth, but this is my impression so far).

erislover makes a good point about the different time limits probably not changing the outcome of the matches, but I don't believe that's really at issue. I hear the issue being the resulting quality of the play as a factor of time allowed to consider one's next move.

So it sounds to me like you're offended that John is offended? That's just weird.

I agree with you (again, apologies if I'm putting words in your mouth now) that I also think that Mr Fairbairn might be a little oversensitive to the time limit issue, and I also don't like it when he refers to that particular issue as "Mickey Mouse time limits", but otherwise I tend to see his points and kinda agree with him about the probable lower quality of the resulting games.

Maybe we should all take a step back and reconsider what we're each being offended about?

[Kirby]

I agree with you (again, apologies if I'm putting words in your mouth now) that I also think that Mr Fairbairn might be a little oversensitive to the time limit issue, and I also don't like it when he refers to that particular issue as "Mickey Mouse time limits", but otherwise I tend to see his points and kinda agree with him about the probable lower quality of the resulting games.

1. I don't agree that there is a "probable lower quality" in the games. I think they are of high quality. I think that quality in Go has improved throughout history. While I agree that some amount of thinking time is required to produce a high quality game, I think that 3 hours is quite a bit of time for a pro.

And saying that a 3 hour time limit is only 3 hours is selling the game short. Pros have studied for thousands of hours before the game even started, in order to prepare themselves. Look at a 3 hour game played today by a pro, and you are looking at thousands of hours of preparation that have gone into playing a high-quality game.

Yes, pros make mistakes under short time limits. But to say that the quality of the game is lower than a game played under increased time limits for a *particular* time limit is pure speculation. The only real evidence that we have are game reviews, but those are only real data points if they are reviewed with equal vigor by the same individual.

To elaborate more on the speculation bit, it is possible that it's the case that increased time universally results in increased game quality. Maybe if you graph it, it looks like this:



With more thinking time, there must be more time to fix mistakes, right? This seems to be a common idea.

But it's also possible that increased time doesn't follow that trend. For example, maybe it's like this:



Extra thinking time helps, but only to a certain degree. After that, the extra benefit of extra time isn't useful.

But a third possibility is also there:



Extra thinking time helps until a point... But then it may actually hurt your game quality - maybe it makes you weary. Or maybe you overthink something. Or maybe something totally different.

Who knows? Not any of us here on the forum.

2. Now, let's say that you assume one of the hypothetical trends above. Even if that's the case, you can note that up until now in this discussion, I didn't label any of the specific time limits on the graphs. Therefore, who is to say that a 2 or 3 hour time limit is not enough? Does 2 hours come before the hump in the last graph? After? What about 6 hours?

We don't know any of this objectively without relying on pro commentary. But pro commentary does not provide an objective view toward the quality of the game.:
* A pro commentator from 50 years ago may commentate based on the trends or what was known 50 years ago.
* Maybe the commentary is intended for a particular audience.
* Maybe the pro commentator missed nuances of the game that the players didn't.
* Maybe the commentator wanted to be nice.
* Maybe the commentator wanted to sell books.
* ...
the list goes on.

3. From the above, my strong view is that the "best" time settings are somewhat of a religious matter. Some people think 3 hours is enough. Some people think you need 6 hours for a high quality game. Some people think you need several days. Some people don't care.

4. Because this is subjective, I object to the laid-back criticism of game quality of what I believe to be high quality games. By calling modern time limits "Mickey Mouse time limits", and by somehow equating modern games with blitz, JF has achieved the apparent objective of leading you and others to believe that there is a "probable lower quality" in today's games.

This is what I am offended with.

First, let me ask, just to be clear, what it is specifically that you respect:

I respect the quality of modern games, the effort that those playing the games have put into them, as well as the effort and sacrifice today's pros have given in order to produce such high quality games.

(We can talk about it offline if you'd like - I will probably be at the SGC sometime this week to drop off some equipment. While I still believe in the message I am trying to convey, I think you are correct that there will probably never be a consensus on this issue.)

[erislover]

erislover makes a good point about the different time limits probably not changing the outcome of the matches, but I don't believe that's really at issue. I hear the issue being the resulting quality of the play as a factor of time allowed to consider one's next move.

Myungwan Kim: https://youtu.be/1jrZ16L8JVU?t=1h4m58s

If it's only 20s, that's pretty amazing how fast that alphago can read. A professional player, surprisingly, whether they have one minute, or ten minutes, or even eight hours, the difference [in how well they read] is pretty small.... A human can read, for example, 50 moves, but even if they have a lot of time, it cannot increase meaningful variations.

[Bantari]

I should first note that I am replying again to this stupid thread because I have little self-control, and I am addicted to replying when I disagree with someone. It is not a wise decision, but it is Friday... And when have I ever been wise?

I feel your pain, brother.

What I know is that:
I make better moves when I have more time. I also make less mistakes. And the mistakes I make are less stupid. This is a fact.

1. You assume you make less mistakes. I find it hard to believe that you can conclude it as fact. For example, maybe you notice "obvious" mistakes during your review, but miss subtle ones that might be harder to find.

I see it as a hard fact. I have the ability to objectively analyse my games. And while I agree with you that in slow games I might make more subtle mistakes than in fast games, I make more obviouse mistakes in fast games for sure. These two facts alone firmly support my statement. And they *are* facts.

The same you can read in Chess, if you care - slower games make for higher quality games.

And anyways - I don't get your argument. Go is a thinking game. It seems logical to me that the more time you have to think the better moves you come up with.

At our level, as amateurs, this might not always be so for various reasons - mental stamina, boredom, overthinking... But we are talking about pros here.

2. The assumption that increased time is linearly related to increased game quality is unfounded.

Linearly? I have no clue.
But it is connected somehow for sure. Otherwise - why not just play 5 min games and have the whole tournament over in half a day?

I am really not sure how you can honestly argue that thinking time has no influence on quality of play. Of course there is some cuto-off beyond which this does not apply, but I seriously doubt that 2 hour per game is it for a pro. Do you have any indication that it is?

Personally, I do *not* know it is a fact, but I also get the feeling that I play a better game if I have 30 or 40 minutes of time, compared to when I play a 10-second blitz game. But it does not follow from this that a 6 hour time limit will result in a better quality game than a 40 minute time limit. And in fact, I may overthink the situation and play worse.

*You* might. Do you think a pro would as well?

One of the points is - and it is an important one - that you don't *have* to use all the time. But... given more time, you remove the element of time pressure. In 40 min games, to follow your example, you might find yourself in a situation in which you need more thinking time on a move but knowing you only have 40 minutes you rush your decision - just in case you need more thinking time again later on. In 2 hour game you will not rush it - even if ultimately you might also only spend 40 min on the game.

The difference is the comfort of knowing you have the time if you need it. Otherwise, you still play at the speed which is most comfortable for you, on average.

Based on the last part of your comment, you seem to agree with #2, above.

Correct.

Although - there is, of course, a limit. I think that given my average "comfort speed"of 30 min per game, it would not make much difference to me if I had 2 hours or 4 hours per game. I would run out of things to think about in both cases - I am simply not strong enough to need so much time to think. But my assumption is that the pros are stronger, they have more things to think about, can calculate deeper, and can actually use the time which I cannot.

So - I don't know if you would see quality difference between 2 day games and 3 day games, for example. But I think you would see difference between 2 hour games and 2 day games.

Do you have any indication or proof that the same principle that applies to us, mere mortals, somehow does not apply to the pros?

No, and I don't think it applies to us, either. See #2, above.[/q]

And I think it does. I gave you my arguments. I assume you disagree and you think that your personal play is the same between 2 hour game and 4 hour game. It might be so... but since mine is not, we cannot really generalize. I would say a much more general (and safer) assumption would be that more time at least *can* (and often *does*) translate into better quality play. Although sometimes it does not, like in your case.

[Bantari]

I agree with you about the diagrams you provided. I think that, for a pro, the diagram will look like the one in the middle. For an amateur, it will look like the 3rd one.

But the issue here is not how the graph looks, but where exactly is the flattening/tipping point of the curve.

All I know for certain is my own tipping point - which is probably somewhere around 1-2 hours beyond which my game degenerates into silliness.

I guess what we disagree about is the exact point at which the curve flattens for a pro. You seem to assume that for a pro it would flatten at around 3 hours per game - or at least there is no "proof" that it is otherwise. I assume that it would flatten much later than that, and that a difference between 3 hours and 6 hours would have a noticable effect. Although I also have no proof of that.

I guess without asking some pros, we really cannot easily decide who is right. Anecdotal evidence provided seems to indicate pros side with me rather than with you here. At least with respect to the quality of the game - which is different from winning chances.

So I keep thinking of my idea of asking two pros to play a game with different time limits - one has 3 hours the other 6 hours. If money was at stake, do you think this would be a fair contest? I think not. According to what you say, you would call it fair. It is hard for me to reconcile that.

[erislover]

I guess what we disagree about is the exact point at which the curve flattens for a pro. You seem to assume that for a pro it would flatten at around 3 hours per game...

So I keep thinking of my idea of asking two pros to play a game with different time limits - one has 3 hours the other 6 hours. If money was at stake, do you think this would be a fair contest? I think not. According to what you say, you would call it fair. It is hard for me to reconcile that.

Longer time would favor one player. How does this relate to "game quality"? The player with more time may simply steer the game into a direction to exploit the time advantage as a metastrategy. I have a hard time understanding how this does or doesn't affect game quality. I mean, timekeeping is not a rule of go, it is an artifact of the practice of playing the game. The Great Go Player in the Sky would win under blitz conditions, but since her opponent may be a fallible human, the game quality could still be quite low, no matter how much time was given.

I don't know how to judge "game quality." The purported measure in this thread is number of mistakes, but this is only justifiable with the rest of OP's comments if we assume some correlation between novel, instructive play and mistakes under time pressure. This correlation is spurious on its own. You need only consider the advice to DDKs to use their influence for fighting more—their walls have died before and they still feel the sting of that, but we would say it was not the single outcome of that game that mattered, you can still trust me on this, please use your walls for fighting and not enclosing territory; which is to say, the quality of the game is not a reflection on the quality of any particular sequence, result, or application of a concept. If it were so, instruction would be literally impossible.

So what we are left with is the notion that the number of mistakes is itself some kind of measure of game quality. I simply can't agree with that. I have my own games that I found very exciting to review, and pro games that were horribly boring to review, and even though there be none who would say my game had a higher "quality" than the pro game in terms of some gimped notion of perfect play, it doesn't stop me from loving that game and going back to it sometimes to feel some measure of pride at how well I executed a plan—even when part of that execution depended on my opponent's mistakes which they faithfully made. How could amateurs ever muster the psychological force needed to play the next game, being of such low quality?

[Kirby]

Bantari, I don't feel like we really disagree that much. I am not making the claim that 3 hours is optimal time for a pro to play. Rather, I claim that we don't know what the optimal time is to produce high game quality for a pro game.

My objection is to the belittling of modern time settings, since I still have faith that the current time setup is able to yield games of high quality. I don't feel that there is sufficient evidence that games from long ago with longer time periods produced higher quality games.

[dfan]

Regarding graph #3, this comment by Kap on the subject of Crazy Stone's blunder in its game with Hajin Lee may be of interest.

[Bantari]

I guess what we disagree about is the exact point at which the curve flattens for a pro. You seem to assume that for a pro it would flatten at around 3 hours per game...

So I keep thinking of my idea of asking two pros to play a game with different time limits - one has 3 hours the other 6 hours. If money was at stake, do you think this would be a fair contest? I think not. According to what you say, you would call it fair. It is hard for me to reconcile that.

Longer time would favor one player.How does this relate to "game quality"? The player with more time may simply steer the game into a direction to exploit the time advantage as a metastrategy.

Well, ok.
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.

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

[erislover]

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).

[Bantari]

Bantari, I don't feel like we really disagree that much. I am not making the claim that 3 hours is optimal time for a pro to play. Rather, I claim that we don't know what the optimal time is to produce high game quality for a pro game.

My objection is to the belittling of modern time settings, since I still have faith that the current time setup is able to yield games of high quality. I don't feel that there is sufficient evidence that games from long ago with longer time periods produced higher quality games.

Well, I can't really disagree with that.
The only thing I have is my gut feeling that the 3 hours are not enough and modern time controls are too short.

All in all, I find it more interesting than a point of contention.

[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

[Fedya]

I see it as a hard fact. I have the ability to objectively analyse my games. And while I agree with you that in slow games I might make more subtle mistakes than in fast games, I make more obviouse mistakes in fast games for sure. These two facts alone firmly support my statement. And they *are* facts.

I'm reminded of the episode of The Simpsons in which Homer changes his name to Max Power:

Homer: There are three ways to do things: the right way, the wrong way, and the Max Power way!
Lisa: Isn't that the wrong way?
Homer: Yes, but faster!

[erislover]

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.

*_{Math fan alert: presumably in a non-field like the extended reals where we can divide by zero.}