Positional Judgment

Path: <= Skills =>
Keywords: Strategy

Chinese: 形势判断 (xing2 shi4 pan4 duan4)
Japanese: 形勢 判断 (keisei handan)
Korean:

If you want to assess whether you will win this game or not, you make a positional judgment using analysis methods to assess territory, thickness, influence and many other positional aspects.

Assessing territory

Counting territory

The current territories in enclosed regions are counted and the score is estimated. Prior quiescence can help. The count of a quiet position is estimated by imagining local peaceful sente endgame reduction sequences to determine the 'current territory'. Optionally and additionally 1-territory, which requires the defender to add 1 more local move and counts as 50% territory, can be determined.

Quiescence

The idea here is to count quiet positions. Imagine typical sequences that make all the important groups stable and settled. Then count current territories. The idea of quiescence comes from expert system computer go programs but can be applied by human players as well. See the other wiki: http://en.wikipedia.org/wiki/Quiescence_search

Unsettled group average

For an unsettled group, miai value (also called average local move value) and count (also called average locale value) can be determined after imagining an attacking sequence and a defending sequence.

QARTS

QARTS reduces the count by -20 for every group without eye and by 0-10 for every weakish group with space to build one eye.

Assessing efficiency

Tewari

Different tewari tools are known:

• difference of numbers of played stones
• removal of plays
• forming stone pairs
• assessing efficiency
• move reversion
• construct the current position by imagining different move orders
• comparing unknown sequences or shapes to known josekis

Not necessarily all of these methods assess efficiency. Some might instead compare between worse and better, incorrect and correct.

Territory efficiency

Territory efficiency avoids overconcentration and calculates points per played stone to compare variations.

Assessing various aspects

Local positional judgement

Local positional judgement is a collection of tools for assessing values of influence or both territory and influence on a local scale. Possible tools are

• estimate of the miai value of an early corner move
• comparing territory with influence
• estimate the value of an extension
• the method 'excess influence stones', which imagines additional territory constructing stones by one player and then estimates the 'territory efficiency'
• the method 'attacking an unsettled group', which imagines an attacking sequence and a defending sequence until quiescence followed by counting territory
• assessing influence and thickness by means of their formal definitions relying on n-connection, n-alive, n-territory.

Global positional judgement

Global positional judgement combines other, more specific judgement methods to assess current territory, 1-territory, outside thickness and influence, mobility, strategic choices and options. Ideally global positional judgement is applied for the whole board but a locally restricted application is also possible (this is not to be confused with 'local positional judgement' though).

Mobility and usefulness values

Mobility and usefulness can be assessed by these values:

Unrest level

• An 'unrest level' is calculated for every local region to assess whether local groups are stable and playing elsewhere is possible.

Discussion

Charles Matthews Very often you actually want to make a positional judgement as a guide of what to do (that is, in relation to some current decision). It is also quite common that it is too hard to assign a definite value to some potential territory, or other factor. Therefore the conclusion is quite likely to be 'I need/don't need a deep invasion' or 'I'm winning as long as my opponent only makes 15 points in the centre' or 'I must now make at least 10 points attacking before the endgame starts, so as to avoid being clearly behind'.

During ko fights, too, one has to judge hypothetical positions after large exchanges as won or lost.

In all these cases there may be at least one unknown factor - and of course one hopes no more than that, though it is possible to operate plans like 'make more on the upper side than the opponent makes on the right side'.

Other SL pages

Path: <= Skills =>
Positional Judgment last edited by RobertJasiek on December 21, 2015 - 13:31