How many points do players lose in the endgame versus perfect play at different levels (e.g. pro, 5d, 1k, 5k, 10k)?
Anybody has at least rough estimates?

If somebody has a link to a worked through endgame (of non-pro) game, I think it would be relevant.
What if we restrict ourselves to endgame after the values of the moves is smaller than 10 or 4?

Thanks

4d:

When I was 4d, I asked Saijo Masataka 8p for how many points I lost during the endgame in one of my tournament games. He reflected for a few minutes, then said: 7 points.

1d ~ 2d:

Antti Törmänen guesses ca. 30 points by his experience (see his endgame video). I think this is reasonable but have not tried to verify it yet.

5k:

I analysed a game and concluded ca. 80 points. You can find it somewhere. Critics objected that I would have counted same intersections more than once arguing that one could lose each intersection only once. However, endgame does not work like this. Rather, at each moment, a mistake loses points immediately.

Kyu:

Antti Törmänen guesses that 100+ points are often lost.

In my experience from teaching games against my pupils and from watching lots of kyu games, I think that 100+ lost points are the norm for weaker than intermediate kyus, surely for players around 10k. In particular, they tend to make several mistakes of which each roughly equals a pass, which equals ca. 14 lost points. (Superfluous reinforcements / invasions lose even 1 point more than a pass.) So DDKs would lose 100 points already by their largest endgame mistakes. Many more points are lost by intermediate and small endgame mistakes.

Suppose a player makes ca. 80 endgame moves and each only loses 1/2 on average. Already this would amount to 40 points or ca. 3 ranks. Reality is that kyus do also make lots of intermediate mistakes, of which each loses a few to several points.

Dans in fast games:

They make occasional kyu mistakes.

I once did a little experiment on KGS where I (maybe 3d at time) played the endgame of a game I picked as black vs a weaker white player (maybe 5k or so I forget exactly), and then we swapped colours and did it again. The difference was about 20 points iirc. And this is with the weaker player being able to use the information of how I played in the first game to help them in the second.

From a different angle, I roughly estimated top pro levels to drop around 50 pts in a whole game. An 1d may drop 14*8 more, and since the majority of these points likely lost before endgame, so you can take a fraction (~10%?).

A lot depends upon time limits. It is said that top players in the 19th century could play the endgame almost perfectly. OC, with no time limits they had all the time in the world to think about the endgame. In the 20th century, with each player having 10 hours to play a game, I have noted that top players made errors when plays gained 3 pts. or more. At today's time limits I don't know.

However, I have participated in organizing and running endgame pro tournaments with set endgame positions where each player got 1 hour. It was not unusual for some pros to score 6 pts. worse than others, much less versus perfect play. And on one occasion the pros burst out laughing. Rui Naiwei had killed one of her opponent's groups.

My educated guess is that at today's time limits pros lose around 5 pts. to par in the endgame, more or less.

In the typical endgame perfect play gains around 100 pts. That's pretty much the limit for losses to par by incorrect choice of play. However, you can lose more than that by making losing plays. Amateurs do that, and so do pros, on occasion, like Rui Naiwei's opponent did. Adjusting for making losing plays, around 30 pts. for shodan sounds about right to me.



Since perfect play gains around 800 - 900 pts. per game, losing only 50 pts. to par is quite good.

To the reader: How can perfect play gain more than 361 pts. in a game? Because there is a back and forth. If the first play gains X points and the second play also gains X points, you are back to 0, no matter how large X is. The difference between how much successive plays gain is, on average, much less than 1 pt.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

What is the reference point for this figure? (I would normally choose perfect play to be my reference point, and then perfect play would gain around 0 points. ) Do we treat every position in the game as if either player might have the move, and then look at gain with respect to that (as we do with miai counting)? That would seem a bit odd to me (what would the total number really signify?) but maybe I need to think about it some more.

At the start of the endgame plays gain around 5 pts. (10 pts. deiri. See many endgame texts.) We may estimate the gain of an endgame play as 2.5 ± 2.5. In the endgame each player play roughly 40 moves (within wide limits). Thay yields a very rough estimate of 100 pts. gain per player in the endgame.

Edit: We can ignore nearly all the moves in a ko fight, because taking and taking back the ko gains nothing locally, and the same goes, as a rule, for playing and answering a ko threat. IOW, we can consider a ko fight before it ends as a long sente sequence.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

If we only consider the endgame phase and not boundary plays before, a scored game can often have more than 50% of its moves in what I perceive as the endgame phase. Therefore, I would consider circa 150 - 200 moves (half of them per player).

Given, say, 2 hours of study time in a day, what percentage should go toward endgame study to optimize for winning more games?

This question is vague, because some people may comprehend endgame concepts more quickly than others, relative to comprehending, e.g., life and death problems. But I'm still curious.

I have limited time, and I have a hard time justifying studying endgame that much.

_________________
be immersed

It's supposed to be easy:

1) Know the approximate value of the boundary moves.
2) Don't delay your sente moves for too long.
3) Accept and work the concept of mutual destruction.

Cheers,
Vesa

Kirby, suppose all your skills have been equal. Study as much endgame as life + death. Spend circa 20% on the endgame. For your 2h/day, that is 24 minutes. To ease your learning, spend half of that on value-less decisions and half on evaluation. Vesa has mentioned a fraction of topics of value-less decisions but there are many more. When your knowledge and skill of them will be developed, spend most on evaluation. IIRC your rank, this may come sooner than you wish. Dans need to study more evaluation than kyus.

Combine endgame and tsumego. Both are matters of technique. Both are concerned with best play in a limited area, where the task is for the most part tractable.

When reviewing your own games try to find the game losing mistake. That will often come in the endgame.

As a dan player you should play the small endgame almost perfectly, without having to take much time per play.

Most books or sets that cover the whole game devote around 20% of their pages to the endgame. If you followed that general rule, you might devote one study day per week to the endgame.

In my own case, I learned the endgame as a kyu player via Takagawa's Go Reader (one book out of five) and Sakata's Killer of Go series (one book covered both tsumego and yose). As a dan player I spent a couple of years studying 15 hours per week, at that time I did not study the endgame per se, but I always included it in reviewing both pro games and my own.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

I think maybe I don't know when the endgame starts...

If it's as soon as you start loosing hundreds of points then, for me, on move 1.

Take care

_________________
一碁一会

Does anyone have some numbers that only count each local situation once?

I get why the numbers are as high as they are but it makes it hard for me to grasp the difference for the ranks like this.

The word "endgame" is not defined precisely so opinions on its start can vary.

Endgame / boundary play can occur from the start of the game. The endgame phase only consists of such moves. However, during a stage first appearing as the endgame phase, play can become hotter again and middle game fighting or killing might be resumed. Otherwise, we might consider the phase between middle game and game end when life statuses are settled. Even so, non-essential stones can be sacrificed, or offered as a sacrifice, also during the endgame phase.

I consider the start of the endgame phase when first endgame considerations become dominating. The stronger one is at them, the sooner one might perceive its start.

Some people say that the endgame starts when the life-and-death status of each group is determined. Even more confusing is that there are differences between endgame and yose, ask John Fairbairn about that.

Moreover, the question in this thread is a bit ambiguous. Let me rephrase it in two different ways.

1) Let X be the sum of points that I lose during the endgame (thus, if I am black and if endgame starts at move 150, then X is the number of points lost at move 151 + number of points lost at move 153 +...). Estimate A=the median value of X.

2) Determine the smallest integer B such that in a game where I lead by B points at the start of the endgame, the probability that I finish the game with a victory against a perfect player exceeds 50%.

I understand that Robert Jasiek gave an estimate of A, while Uberdude's experiment gives some hints on the value of B.

However, A and B are a priori different, my guess is that A>B.

Local situations change and exchanges can occur. Maybe you better ask for "counting each intersection only once for both players (or for each player), even if different mistakes affect same intersections several times"?

However, how would you want to do that? Not playing in one local endgame but playing in a different local endgame elsewhere must be considered together to assess sizes of mistakes. At different moves, different combinations of local endgames are involved.

As has been pointed out, the boundary between different phases of the game can be fuzzy. In the endgame the board breaks up into local regions of play. OC, ko fights can cover the whole board, because of the possibility of ignoring ko threats. Ko fights aside, the endgame typically starts when large scale fights end.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

Suppose, as an example, that there are 6 endgame plays left, that gain 6, 5, 4, 3, 2, and 1 point. I play first. With perfect play I gain 3 points: 6 - 5 + 4 - 3 + 2 - 1 = 3. Now let me make the worst move choices I can; then I lose 9 points: 1 - 6 + 2 - 5 + 3 - 4 = -9. (I could make better choices with the same result: 3 - 6 + 2 - 5 + 1 - 4 = -9. )

Versus perfect play by me, I lose 12 points. I could have played 12 points better. I think that this is the number that Antti, Robert, and I are talking about.

Versus perfect play by the opponent, I lose 9 points. Had I been leading by 9 points I would have got jigo.

-----

Earier I said that the largest endgame plays typically gain about 5 points. 6 points is close enough for illustrative purposes. Now let's multiply the number of plays of each size by 11, so that there are 66 plays. Then my loss versus perfect play by me would be 102 points, and my loss on the board would be 99 points. This example is close enough to real life that you can see how someone could lose 100 points in the endgame, just by poor choice of moves.

Edited for correctness.

You can also see how, with a lot of plays, there is little difference between the two measures of loss.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.