Positional Judgment
If you want to assess whether or not you will win a game, you may make a positional judgment by estimating territory, thickness, influence and many other positional aspects.
This article provides an overview of the various aspects of positional judgement and lists various techniques for assigning numerical values to some of them. It also lists some useful exercises, problems and books.
Table of contents |
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, a system developed by Rob van Zeijst, 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
- Constructint the current position by imagining different move orders
- Comparing unknown sequences or shapes to known josekis
Not all these methods are limited to assessing efficiency: some may be applied to distinguish better from worse moves, or correct ones from incorrect ones.
Territory efficiency
To make territory efficiently one should avoid overconcentration; the territorial efficiency of various sequences may be compared by calculating (estimated) territory gained per stone played.
Assessing various aspects
Local positional judgement
Local positional judgement involves assessing the value of influence or possibly territory on a local scale. Some tools for this are:
- Estimates of the miai value of early corner moves
- Estimates of the value of extensions
- Comparison of territory and influence
- 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 some formal definitions of these such as Robert Jasiek’s terms n-connection, n-alive and n-territory.
Global positional judgement
Global positional judgement combines other, more specific methods to assess current territory, 1-territory (i.e. that can be made with a single move), outside thickness and influence, mobility, strategic choices and options. Ideally, global positional judgement is applied to the whole board but a locally restricted application is also possible; this should not be confused with 'local positional judgement'.
Mobility and usefulness values
Mobility and usefulness can be assessed by these values:
- Mobility count and difference
- Usefulness count
- Usefulness difference
Unrest level
- An 'unrest level' may be 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'.
Literature
- Positional Judgement 1 / Territory by Robert Jasiek explains every aspect of territorial positional judgement.
- Positional Judgement 2 / Dynamics by Robert Jasiek explains every dynamic aspect of positional judgement, such as invasions, influence and fights.
- Positional Judgment High-Speed Game Analysis by Cho Chikun explains a few aspects of territorial positional judgement.
- Joseki Vol. 2 Strategy by Robert Jasiek explains a couple of other aspects of positional judgement.
- The Vital Points of Go by Takagawa Kaku explains a few tewari tools.
Problems
- PositionalJudgmentQuestions — A series of problems to develop one’s positional judgement, including these:
- Bill Spight's Elf Positional Judgment Exercises — A comparison of judgements by ELF and human players