How does one count points, territory, etc. How do you know what is your territory and what is not, and how can you add this up in your mind during a game to get a rough estimate on who is winning?

Knowing what is territory and what isn't through the middle game can be difficult until you get a fair amount of experience under your belt. Generally speaking, if you have a tight framework with stones on the 3rd line on both sides and no wide open gaps, you can consider it territory. Similarly, in the absence of strong enemy stones nearby, corner enclosures can generally be counted. If there was a joseki where you took the territory side and your opponent got influence, that is often territory as well.

For a very rough middle game estimate, you can just drop imaginary lines to the end of the board from your 3rd line stones and count the spaces inside that or multiply width*height. If you don't have 3rd lines stones, the position is much more likely to be undermined or reduced, and should be considered as potential. Repeat this for all the relatively solid territory and add prisoners, and you have a rough estimate of cash in hand. If I am estimating this way, I usually don't look to be much more accurate than counting in 5s, because it's pretty imprecise. Once this is done, take a look at areas of potential for both sides, like frameworks, to get an idea of who can make the most points out of the rest of the board. You can count that in rough rectangles too. If one side has a weak group that the opponent can beat up on, consider that additional points or possibilities for invasions may open up as a result. If one side is ahead in both territory and potential, they're ahead in the game. If it's mixed, you need to figure out a plan to reduce the potential or maximize the territory made from the potential to get yourself ahead in the end.

You can count much more accurately than this as the game goes farther along, borders get sealed up, and the reductions available to the opponent become much more clear, in which case you can take them in to account. Additionally, as you get stronger, your judgements of territory vs potential will become more clear and accurate earlier in the game, and you should be able to count more accurately as a result, taking into account sente reductions the opponent will likely be able to play.

I have heard that it's a good idea to keep the counts for each group or region separate, so that you only need to recount one group or section when something changes instead of the whole board, but I am not very good at this.

Hopefully this is enough to get you started, at least.

If you search for "counting" or "positional judgement" or "analysis", you can find related threads during the previous ca. 10 months.



First, consider your purpose of determining the territory intersections. Second, choose some method suitable for the purpose:

- territorial positional judgement
- a player's number of successive initial plays so that then the opponent cannot prevent territory on a particular intersection or set (n-territory)
- pragmatic 'the opponent cannot invade and live there' for fighting considerations
- model go rules scoring definition (such as in the Japanese 2003 Rules)

This is true for two space jumps, but at three spaces or more, you should assume that an invasion is possible, unless the three space extension has extra support.

_________________
Occupy Babel!

Yes, I probably should have mentioned that

Really, judging when something is invadable and when it isn't is something that requires experience. A two space extension could also potentially be worth nothing if your opponent has it surrounded, after all.