Average value of a move

[skydyr]

And on the other hand, you can see a horribly embarrassing game of mine where W made a monkey jump to add an eye!

Didn't white already have two eyes at O18 and P12?

[Uberdude]

Didn't white already have two eyes at O18 and P12?

No, o18 is a false eye.

[skydyr]

Didn't white already have two eyes at O18 and P12?

No, o18 is a false eye.

Ah. I can tell this is true because I seem to be looking at it through my own false eye

[Bill Spight]

This is something I have been puzzling over.

What is the average value of a move, as the game progresses?

Nobody knows.

We now think that the first move gains around 14 points. But 50 years ago (and for a long time before) we thought that it gained around 10 points.

My first line of thinking was that komi is 6.5, and an endgame move (around move 200) might be 1 or 2, so you could draw a straight line between those points. But that doesn’t sound right (too low).

It is too low. Komi does not equal the size of a play. But even if you drew a straight line between 14 points at move 1 and 0 points at move 240 or so, you would surely not even be close for much of the game. As others have pointed out, the size of a move increases for sustained periods during middle game battles.

My reason for wanting to know is that, say on move 100 I find a move worth 5 points, should I be pleased to play it without a lot more thought, or spend the time looking for something better?

First, I don't know what you mean by a move worth 5 points. For one thing, there are two main ways of valuing points. One is by how much a move gains on average, the other is how much a move gains for one player plus how much a move gains for the other. The 14 point value uses the first way. For instance, if all a move does is to capture 5 stones that the opponent could save in one move, it is worth 5 points by the first way, 10 points by the second. And you may have yet a different idea in mind.

Second, at move 100 you are very likely to be in the middle game, and the moves will be worth much more than 5 points by either way.

Third, of course you should spend some time looking for something better, even if it is only a few seconds.

[Bill Spight]

Let me come at it from a different angle.

Suppose an opportunity for a monkey jump arises in the first 20 moves. A monkey jump is usually worth around 8 points, I think?

Well, obviously you would not want to play it straightaway.

Roughly when in the game would you expect to start considering it as a serious option? (Or blocking it, if you are the one who left it open)

The size of the monkey jump depends upon the surroundings. Sometimes it is sente, sometimes it is not.

If you want to estimate the move number to play it, my first thought is that that is a question for the database guys. This page on Sensei's Library is relevant. http://senseis.xmp.net/?PointPopularityByMoveNumber

My interpretation is that the popularity of plays in the center at move 105 indicates the middle game. There is not much territory in the center, so I expect that plays there are made to attack and defend groups. The fact that play shifts back to the second line at move 142 indicates the large endgame stage. (That is later than John Fairbairn reports. That could be because the pros do not actually note the number of the move when they start playing the endgame -- who does? --, and are misestimating it, or it could be because the largest endgame plays are often in the center. Edit: It could also be because large yose plays may be made in the middle game, or even the opening.) By move 199 play starts moving to the first line. Most of those plays are not monkey jumps, and are smaller. So monkey jumps would probably be played earlier.

Based on that one page, my guess for the typical move range for a monkey jump would be between move 140 or so and move 180 or so, depending upon the surroundings.

[Bill Spight]

And on the other hand, you can see a horribly embarrassing game of mine where W made a monkey jump to add an eye!

Too bad White didn't resign after move 25.

[EdLee]

I know, but playing a game in online timeframes, and with 200 empty points to evaluate in a typical middle game position, a rule of thumb comes in handy .

Umm... actually, no. I'd say it's not handy; it's detrimental to your Go, IMO.

[PeterPeter]

I know, but playing a game in online timeframes, and with 200 empty points to evaluate in a typical middle game position, a rule of thumb comes in handy .

Umm... actually, no. I'd say it's not handy; it's detrimental to your Go, IMO.

Yes, I think you might be right.

[Phoenix]

The average value of a move doesn't help much as a guideline. For example when I play a slow, solid move, I know for a fact that later on in the game the stone that made no territory whatsoever is going to help me make lots of territory later.

Calculating the average value of a move as the game progresses leaves you blind to what's going on on the board. Stones have function, groups have strength. You're making plans and so is your opponent. That's what matters.

At the end of a close game, each stone you played will have returned a value of around 0.5 points anyway. There's enough to worry about without trying to find out whether or not you've met quota every move. If you really are worried, do what the pros do: count relative territory every chance you get.