How to learn to count during midgame

emeraldemon · **#21**

Clearly this calls for a Magicwand-RobertJasiek malkovich!

p2501 · **#22**

RobertJasiek wrote:

(Where "weak" = German 5d.)

@Magicwand
http://www.europeangodatabase.eu/EGD/Pl ... y=10213203

RobertJasiek wrote:

An alternative theory is that I am one of the greatest victims of the available teaching by examples literature especially in case of the life and death topic. If (advanced) analytical life and death solution literature were already available and I would not need to write it for the sake of giving myself the chance to read it, then I would have improved incredibly much faster at life and death problem solving, because I am very good at learning from analytical teaching material. Instead, I am struggling to find our everything by my own, ...

I don't know what to say :scratch:

Do you really mean that?

RobertJasiek · **#23**

p2501 wrote:

http://www.europeangodatabase.eu/EGD/Pl ... y=10213203

The "declared rank" graph shows in 2011 what the French congress organisers declared, not what I declared.

Quote:

Do you really mean that?

Sure. (The major problem is that almost all literature available on life and death is teaching by examples only, sometimes teaching by examples combined with sorting of problems by some low level type. What is almost completely missing is generalising, example-independent analytical teaching of the move sequences and involved decisions.)

speedchase · **#24**

RobertJasiek wrote:

Sure. (The major problem is that almost all literature available on life and death is teaching by examples only, sometimes teaching by examples combined with sorting of problems by some low level type. What is almost completely missing is generalising, example-independent analytical teaching of the move sequences and involved decisions.)

The first chapter of Tesuji?

RobertJasiek · **#25**

speedchase wrote:

The first chapter of Tesuji?

This (and in more detail the Reading chapter of First Fundamentals) is about what reading is, but not about how to solve classes of LD problems.

Mef · **#26**

Magicwand wrote:

you only need two step:

Step 1:
count what is definit resolved territory.

Step 2:
for unresolved territory, make an estimate on howmany points you or opponent need to cashout in order to make the game balanced.
you must factor thickness, weakness, aji, etc.

very simple and surprisingly accurate.

I think this method is quite practical and more or less what I try to do. The whole purpose of counting is to devise your strategy going forward in the game. If you count current solid points, you have a picture of what you need to do to win the game. If what you need to do seems fairly straightforward, you are probably ahead and can play a simple game to victory. If what you need to do is challenging, you are probably behind and need to complicated things. When you are actually playing the game, it's more important to know what you need to do to win, rather than know if you are ahead or behind.

Magicwand wrote:

I suggest everyone to study endgame before counting. you must have basic knowledge on endgame in order to apply counting.

I would disagree here, if only because I think counting is something everyone can start doing, and the more practice they get the better they will be at it. Studying endgame will certainly help you improve your counting abilities (because you will learn what is truly solid), but I wouldn't let not having studied endgame stop you from counting.

And for the side discussion with Robert (hidden for the benefit of those who wish to avoid it):

RobertJasiek · **#27**

Mef wrote:

area scoring

It is not a scoring system argument.

Quote:

it is trivial to extend that to an accurate count on a prisoner board without paying respect to prisoners.

It is not trivial (but becomes obvious once one sees it explained carefully for the first time), I will show a bit of it later:)

Hint: more than one sequence must be considered. Each has its prisoners.

Mef · **#28**

RobertJasiek wrote:

It is not trivial (but becomes obvious once one sees it explained carefully for the first time), I will show a bit of it later:)

No, it is trivial. You can accurately count the board under area scoring without tracking prisoners. In middlegame, the area score and territory score will yield the same result. Therefore you can accurately count the game without keeping track of prisoners.

RobertJasiek · **#29**

Ah, now I get it what you want to discuss. Yes, with a pure area count, prisoners are not needed. Which experience do you have with speed during the middle game?

Phelan · **#30**

otenki wrote:

Hi guys,

I think it is very important to be able to count well during midgame in order to have ideas on how to play next.
Howhever I have never managed to do this very well.
I think the reason that I have never done this very well is because I always have done this by just "feeling" how big things are.

So lately it has been happening a lot that I loose games which I thought I was leading and vice versa.
I want to learn to count well but not loose too much time with it.
How did you learn to count well, and did this have impact on your game ?
Do you use some kind of guessing system, acurate counting, pair counting ?

Any tips welcome!

Thanks,
Otenki

There's an overview of several methods used in Senseis(where else?

). I, like xed_over, really recommend DrStraw's counting lessons.

RobertJasiek · **#31**

I promised an example for prisoners, so here it is. For the sake of simplicity, I let it look like an endgame position. Just imagine the local shape to be part of a middle game position.

INITIAL POSITION

Click Here To Show Diagram Code: [go]$$B Initial prisoner difference = 100 $$ ------------------------------------------- $$ | X . X . . . . . . . . . . . . . . O . O | $$ | . X X X X X O . . . . . . X O O O O O . | $$ | X X X O O O O . . . . . . X X X X O O O | $$ | . X O O . O . X . . . . O . X . X X O . | $$ | . X O . O O . X . . . . O . X X . X O . | $$ | . X O O O . X . . O X . . O . X X X O . | $$ | . O O . . X . . . O X . . . O . . X X . | $$ | . . . X X . . O . O X . X . . O O . . . | $$ -------------------------------------------[/go]

POSITIONAL JUDGEMENT I - MISTAKE

Click Here To Show Diagram Code: [go]$$W Sequence for determining Black's territory $$ ------------------------------------------- $$ | X . X . . 4 3 . . . . . . . . . . O . O | $$ | . X X X X X O . . . . . . X O O O O O . | $$ | X X X O O O O . . . . . . X X X X O O O | $$ | . X O O . O . X . . . . O . X . X X O . | $$ | . X O . O O . X . . . . O . X X . X O . | $$ | 2 X O O O . X . . O X . . O . X X X O . | $$ | 1 O O . . X . . . O X . . . O . . X X . | $$ | . . . X X . . O . O X . X . . O O . . . | $$ -------------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$B Black's territory, current prisoner difference 100 $$ ------------------------------------------- $$ | X C X C C X O . . . . . . . . . . O . O | $$ | C X X X X X O . . . . . . X O O O O O . | $$ | X X X O O O O . . . . . . X X X X O O O | $$ | C X O O . O . X . . . . O . X C X X O . | $$ | C X O . O O . X . . . . O . X X C X O . | $$ | X X O O O . X . . O X . . O . X X X O . | $$ | O O O . . X . . . O X . . . O . . X X . | $$ | . . . X X . . O . O X . X . . O O . . . | $$ -------------------------------------------[/go]

Black's territory including the prisoners is

8 + 100 = 108 points.

Click Here To Show Diagram Code: [go]$$B Sequence for determining White's territory $$ ------------------------------------------- $$ | X . X . . . . . . . . . . 3 4 . . O . O | $$ | . X X X X X O . . . . . . X O O O O O . | $$ | X X X O O O O . . . . . . X X X X O O O | $$ | . X O O . O . X . . . . O . X . X X O . | $$ | . X O . O O . X . . . . O . X X . X O . | $$ | . X O O O . X . . O X . . O . X X X O 2 | $$ | . O O . . X . . . O X . . . O . . X X 1 | $$ | . . . X X . . O . O X . X . . O O . . . | $$ -------------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$B White's territory, current prisoner difference 100 $$ ------------------------------------------- $$ | X . X . . . . . . . . . . X O C C O C O | $$ | . X X X X X O . . . . . . X O O O O O C | $$ | X X X O O O O . . . . . . X X X X O O O | $$ | . X O O C O . X . . . . O . X . X X O C | $$ | . X O C O O . X . . . . O . X X . X O C | $$ | . X O O O . X . . O X . . O . X X X O O | $$ | . O O . . X . . . O X . . . O . . X X X | $$ | . . . X X . . O . O X . X . . O O . . . | $$ -------------------------------------------[/go]

Including the prisoners, White's territory is

8 + 100 = 108 points.

The territory count is Black's minus White's value:

108 - 108 = 0 points.

POSITIONAL JUDGEMENT II - CORRECT

Click Here To Show Diagram Code: [go]$$W Sequence for determining Black's territory $$ ------------------------------------------- $$ | X . X . . 4 3 . . . . . . . . . . O . O | $$ | . X X X X X O . . . . . . X O O O O O . | $$ | X X X O O O O . . . . . . X X X X O O O | $$ | . X O O . O . X . . . . O . X . X X O . | $$ | . X O . O O . X . . . . O . X X . X O . | $$ | 2 X O O O . X . . O X . . O . X X X O . | $$ | 1 O O . . X . . . O X . . . O . . X X . | $$ | . . . X X . . O . O X . X . . O O . . . | $$ -------------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$B Black's territory, sequence's prisoner difference 0 $$ ------------------------------------------- $$ | X C X C C X O . . . . . . . . . . O . O | $$ | C X X X X X O . . . . . . X O O O O O . | $$ | X X X O O O O . . . . . . X X X X O O O | $$ | C X O O . O . X . . . . O . X C X X O . | $$ | C X O . O O . X . . . . O . X X C X O . | $$ | X X O O O . X . . O X . . O . X X X O . | $$ | O O O . . X . . . O X . . . O . . X X . | $$ | . . . X X . . O . O X . X . . O O . . . | $$ -------------------------------------------[/go]

Black's territory including the sequence's prisoners is

8 + 0 = 8 points.

Click Here To Show Diagram Code: [go]$$B Sequence for determining White's territory $$ ------------------------------------------- $$ | X . X . . . . . . . . . . 3 4 . . O . O | $$ | . X X X X X O . . . . . . X O O O O O . | $$ | X X X O O O O . . . . . . X X X X O O O | $$ | . X O O . O . X . . . . O . X . X X O . | $$ | . X O . O O . X . . . . O . X X . X O . | $$ | . X O O O . X . . O X . . O . X X X O 2 | $$ | . O O . . X . . . O X . . . O . . X X 1 | $$ | . . . X X . . O . O X . X . . O O . . . | $$ -------------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$B White's territory, sequence's prisoner difference 0 $$ ------------------------------------------- $$ | X . X . . . . . . . . . . X O C C O C O | $$ | . X X X X X O . . . . . . X O O O O O C | $$ | X X X O O O O . . . . . . X X X X O O O | $$ | . X O O C O . X . . . . O . X . X X O C | $$ | . X O C O O . X . . . . O . X X . X O C | $$ | . X O O O . X . . O X . . O . X X X O O | $$ | . O O . . X . . . O X . . . O . . X X X | $$ | . . . X X . . O . O X . X . . O O . . . | $$ -------------------------------------------[/go]

Including the prisoners, White's territory is

8 - 0 = 8 points.

The territory count is Black's minus White's value and includes the initial prisoner difference:

8 - 8 + 100 = 0 + 100 = 100 points in Black's favour.

NAIVE PRINCIPLE

Count each prisoner only once.

BETTER PRINCIPLE

Count each prisoner only once, even if it occurs in different sequences that are not alternative variations.

REMARKS

Initial prisoners must be considered for each alternative variation. The sequences for determining either Black's or White's territory are not alternative to each other, but they coexist. We would be considering alternative variations if we continued the game from the initial position with different sequences and then make judgements of the black and white territories for the resulting game positions.

More principles for prisoners and positional judgement are needed but not discussed here at the moment.

Without knowing or thinking how and when to add or subtract prisoners during positional judgement, one can easily make mistakes and get wrong counts as a consequence of confusing addition and subtraction or counting prisoners several times that must be counted only once.

I have chosen the initial prisoner difference 100 to make it obvious why considering it twice is a mistake. If, however, the initial prisoner difference is small (a value such as 1) and more prisoners are made during one or both sequences, then it is easy to make wrong calculations because mistakes are not so obvious.

EDITS: corrections, early versions were wrong. More remarks.

RobertJasiek · **#32**

POSITIONAL JUDGEMENT III - OTHER MISTAKES

Click Here To Show Diagram Code: [go]$$W Sequence for determining Black's territory $$ ------------------------------------------- $$ | X . X . . 4 3 . . . . . . . . . . O . O | $$ | . X X X X X O . . . . . . X O O O O O . | $$ | X X X O O O O . . . . . . X X X X O O O | $$ | . X O O . O . X . . . . O . X . X X O . | $$ | . X O . O O . X . . . . O . X X . X O . | $$ | 2 X O O O . X . . O X . . O . X X X O . | $$ | 1 O O . . X . . . O X . . . O . . X X . | $$ | . . . X X . . O . O X . X . . O O . . . | $$ -------------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$B Black's territory, current prisoner difference 100 $$ ------------------------------------------- $$ | X C X C C X O . . . . . . . . . . O . O | $$ | C X X X X X O . . . . . . X O O O O O . | $$ | X X X O O O O . . . . . . X X X X O O O | $$ | C X O O . O . X . . . . O . X C X X O . | $$ | C X O . O O . X . . . . O . X X C X O . | $$ | X X O O O . X . . O X . . O . X X X O . | $$ | O O O . . X . . . O X . . . O . . X X . | $$ | . . . X X . . O . O X . X . . O O . . . | $$ -------------------------------------------[/go]

Black's territory including the prisoners is

8 + 100 = 108 points.

Click Here To Show Diagram Code: [go]$$B Sequence for determining White's territory $$ ------------------------------------------- $$ | X . X . . . . . . . . . . 3 4 . . O . O | $$ | . X X X X X O . . . . . . X O O O O O . | $$ | X X X O O O O . . . . . . X X X X O O O | $$ | . X O O . O . X . . . . O . X . X X O . | $$ | . X O . O O . X . . . . O . X X . X O . | $$ | . X O O O . X . . O X . . O . X X X O 2 | $$ | . O O . . X . . . O X . . . O . . X X 1 | $$ | . . . X X . . O . O X . X . . O O . . . | $$ -------------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$B White's territory, current prisoner difference 100 $$ ------------------------------------------- $$ | X . X . . . . . . . . . . X O C C O C O | $$ | . X X X X X O . . . . . . X O O O O O C | $$ | X X X O O O O . . . . . . X X X X O O O | $$ | . X O O C O . X . . . . O . X . X X O C | $$ | . X O C O O . X . . . . O . X X . X O C | $$ | . X O O O . X . . O X . . O . X X X O O | $$ | . O O . . X . . . O X . . . O . . X X X | $$ | . . . X X . . O . O X . X . . O O . . . | $$ -------------------------------------------[/go]

Including the prisoners, White's territory is

8 - 100 = -92 points. // Here I alter the mistakes.

The territory count is Black's minus White's value:

108 - (-92) = 108 + 92 = 200 points. // In effect, here we have counted the initial prisoners twice. In Positional Jugdement I, we did not count them at all:)

RobertJasiek · **#33**

It is particularly tempting to make the mistakes in Positional Jugdement I or III when one imagines prisoners as stones collected in the lids. Such "lid collections" do not distinguish between initial prisoners and prisoners made during imagined sequences! However, analysis such as positional judgement can require a more sophisticated calculation of prisoners!

schawipp · **#34**

otenki wrote:

...How did you learn to count well, and did this have impact on your game?

I still feel quite unsure in counting, and improvement in that aspect would for sure help a lot... What I sometimes try is to visualize the partly imaginary boundary lines (including boundary stones, if possible...) of black and white groups and see if they mutually cancel out (e. g. this W top left corner is about the same size as the two B groups at the right...). By this way I get at least a gut feeling about who is leading. Recently I finished a game where at the end (=after resigning) both players were still hallucinating about the L&D state of a large ~70pt group; thus L&D practice is also always helpful... :mrgreen:

Magicwand · **#35**

Professionals can not count 100% there are things you can corn and things you have to estimate. Count what u can and you decision is making estimate on what is left. First learn how to count what you can. Next is your feeling on what you can not count. Professionals call them alpha.
Edit. If any one is claiming that they have a formula for counting they are moran.

topazg · **#36**

Magicwand wrote:

Professionals can not count 100% there are things you can corn and things you have to estimate. Count what u can and you decision is making estimate on what is left. First learn how to count what you can. Next is your feeling on what you can not count. Professionals call them alpha.
Edit. If any one is claiming that they have a formula for counting they are moran.

I think there's a huge amount in what magicwand has said in this thread. The amount of professional games where, even whilst playing, the commentator will say something like "and White's 1.5-2.5 points ahead so doesn't need to play this ko". I've always been very impressed about the length and depth and accuracy of professional counting, and I don't believe for a second it's by some comprehensive and deep formula.

A good, precise endgame visualising what's going to happen to the end and counting the points. The more I've watched professional games and observed professional game commentaries, the more I see them doing exactly what magicwand says, and they never seem to be off the mark except where one player, knowing they're behind, complicates things for the sake of catching up.

RobertJasiek · **#37**

Magicwand wrote:

Professionals can not count 100% there are things you can corn and things you have to estimate. [...] If any one is claiming that they have a formula for counting

Nobody can count 100%, but the point of positional judgement is the construction of move sequences so that estimates and guesses can be replaced by good approximations of regions that can be counted.

If there is a player's region that can be counted, then (basically) the "formula" is very easy: each empty intersection is 1 point, each dead opposing stone intersection is 2 points, each opposing prisoner removed from that region is 1 point.

There are more complicated cases: e.g., teire is 0 points, a gote endgame option is 1/2 of its points etc.

Quote:

they are moran

What is moran?:) Poor Magicwand does not realise that positional judgement "formula" are very simple, see above, in most practical cases. The difficulty lies elsewhere: in constructing suitable sequences to transform fighting positions into easily countable positions.

RobertJasiek · **#38**

topazg wrote:

the more I see them doing exactly what magicwand says

What exactly does he say? He remains nebulous or contradicts himself when saying both "good endgame counting is necessary" and "they just guess instead of using a formula".

What exactly do you see the pros doing? Usually, pros I have seen would, if at all, only state some resulting count number but say nothing about how they got it.

topazg · **#39**

RobertJasiek wrote:

topazg wrote:

the more I see them doing exactly what magicwand says

What exactly does he say? He remains nebulous or contradicts himself when saying both "good endgame counting is necessary" and "they just guess instead of using a formula".

What exactly do you see the pros doing? Usually, pros I have seen would, if at all, only state some resulting count number but say nothing about how they got it.

It's all in the thread so I don't see any point in rehashing it, but you're pretty badly misinterpreting (and rephrasing) what he's saying.

RobertJasiek · **#40**

topazg wrote:

you're pretty badly misinterpreting (and rephrasing) what he's saying.

Ok, let me cite of what he says:

"you must factor thickness, weakness, aji, etc."

This is nebulous because he does not say at all HOW to do so.

"they look at each position by deciding who has more rights to the position"

This is nebulous because he does not say at all HOW to determine who has more rights.

***

"i would like to compare it with other good endgame books i have read. [...] I suggest everyone to study endgame before counting. you must have basic knowledge on endgame in order to apply counting."

Here he suggests that good endgame knowledge / study is needed.

"If any one is claiming that they have a formula for counting they are moran."

Here he suggests that using formula for counting aka positional judgement is bad. This contradicts his previous statements referring to good endgame knowledge / study (everybody knows that endgame knowledge relies on formula etc. such as forming averages for Black versus White first). So let us assume that he does not mean this contradiction. The other way he offers is guessing:

"things you have to estimate"

However, this also contradicts his endgame knowledge statement because guessing does not need any endgame knowledge!

How to learn to count during midgame

Who is online