Petition to add stone scoring/group tax

General Go Discussion
1

The developers have indicated that currently, it’s not in their immediate plans. However, if there’s significant interest, they may consider starting work on it.

I’m 12k and I’ve been playing this cool style of go where I always go for the center, but without a group tax scoring system, it makes it less fun :slight_smile:

I would also suggest a “self fill in” button that let’s GNU Go fill in all but two of the intersections on every group. This “button” would only be available after move 200 or so.

But yeah, I think professional go championships should be played with the stone scoring/group tax method, but what do I know :slight_smile:

Cheers!

7 Likes
looking for sage advice teaching in a game how to play
2

I’d also like to point out that the code that defines the rulesets/scoring lives in an open source library:

So “significant interest” could also be manifested as code contributions, which would significantly reduce the barrier to making the feature a reality.

FWIW, I think it would be fun ruleset to try :slight_smile:

6 Likes
3

INB4 someone plays this ruleset then uses the 3 pass force score button :rofl:

3 Likes
4

So, this rule slightly punishes big number of groups.

Rule with more unusual result:
Player who has more than 1 group loses.
It may result in very peaceful game where both opponents are afraid that part of their group is separated. Separated group is possible to convert to 1-eye group later, but its big loss of points.
Or someone may attempt to cut board in half like in Hex. Then opponent would be able to take territory from 1 side of board only and it most likely would not be enough.

1 Like
5

I think stone scoring also allows to explore important aspects of the game instead of being told about them (eyes, life, territory) and it would be a good scoring method for beginners. So hereby I’m signing this petition.

4 Likes
6

As a compromise, what about player with the fewest groups wins, ties are broken with area scoring (+komi)? seems more elegant than just a special case for 1 group

That said, I’d see that as more the purview of variants, whereas stone scoring is not so much a variant per se, but just a different Scoring system which is not so popular today, but has some interesting (if subtle) strategic implications

1 Like
7

Could I win by sacrificing groups until I had zero?

6 Likes
8

Stone scoring is like area scoring with group tax, but it accomplished by essentially making area worth zero points, and thus leaving eyes is penalty (for having many groups).

Area scoring: 1 point per living stone + 1 point per territory surrounded
Stone scoring: 1 point per living stone + 0 points per territory surrounded

So, another way to fine tune the penalty of group tax is to adopt something like this:

Hypothetical scoring: N point(s) per living stone + M point(s) per territory surrounded

Of course, it’s really only the relative size of M and N that matter. For example, any N = M > 0 is essentially equivalent to N = M = 1 (provided that one similarly scales komi).

To make the group tax effect greater, one could use N > 1, or even M < 0.

1 Like
9

thanks for all the engagement gang!

I think one of the coolest reason for stone scoring/group tax is because it codifies the objective of the game very eloquently.

The object of the game is to get the most stones on the board.

Now if you play this game for 500 years you end up seeing why Japanese rules are cool.

But yeah, get the most stones on the board, this is how the game should be explained to beginners imho.

And I also think these rules seem more primal and organic :slight_smile:

4 Likes
10

Note that if you do implement it, you’ll already have AI analysis for it, since Kata is trained on it already. That’s not true of most other rules variants though, just group tax rules.

4 Likes
11

I guess it needs a minimum of 1 :smiley:

1 Like
12

Topic: counting group tax on a real board.

The “group tax” seems natural and elegant, but I have read it was abandoned because it was harder to count that way. I know from experience that this can be true when you try to count the “Chinese way”, because there is an added complication. The total eye space for example reduces the total stones on the board down from the predictable 361. Many people prefer “Japanese” counting, which they say is simpler/faster, but isn’t the following method just as simple as “Japanese counting”, even with group tax, while you are effectively counting the stones difference and hence doing away with all the Japanese complications. (From what I read, this is close to or perhaps the same as what happened in the Chinese Tang dynasty: History of Go rules, PDF.)

If you make sure both sides have played the same amount of stones to the board (like AGA rules: passing gives the opponent a prisoner, all prisoners are returned to the board in the end as usual), then you end with the same amount of stones for both players on the board. You can then ignore all those stones in the counting, because you want the difference between the opponents. This difference now only lies in the remaining open but enclosed spaces (territory).

Then use additional stones to fill in all the minimum necessary eyes (two stones per group). Then count the remaining open spaces (territory) for both players, and compare the result to find the winner.

Example:

example-game-abs.stone.count
example-game-abs.stone.count849×848 235 KB

Black: 24 total stones + territory (“area”). 1 groups, 2 eyes, 22 total stones minus eyes, 8 surrounded empty points (“territory”), 6 surrounded empty points minus the eyes.
White: 25 total stones + territory (“area”). 2 groups, 4 eyes, 21 total stones minus eyes, 9 surrounded empty points (“territory”), 5 surrounded points minus the eyes.

In both methods, Black wins by +1 stone. The benefit is that it is easier to count to 5 and 6 and compute the 1 difference, than it may be to count to 24, subtract that from the 49 (board size), then subtract 2 eyes from 24 = 22 (Black), and subtract 49 - 24 - 4 = 21 (White), and subtract 22 - 21 = 1. The underlying reason may be that there are more stones than empty spaces, and/or that it is easier to rearrange the empty spaces.

In currently popular rules, White wins by 1 (eyes are counted as points).

Do dame need to be played to the end ? In the historical PDF linked above, The History of Go Rules by Chen Zuyuan 陈祖源 © 2011, page 10:

As a result White would lose by 9. But the result is still 8, as in present Japanese rules. How can that be? A reasonable assumption is: If Black makes the last move, in order to have equal stones for each side, Black will remove his last stone. Once territory counting is adopted, dame are naturally not played.

This I have now tested twice, and (I am sorry to say) it does not seem to work out this way, but rather it works as the AGA rules seem to prescribe. If black takes the last dame point, then white will play a pass stone, and then both the absolute stone count (counting all stones and territory), and only the territory (because the stones are equal) give the same result (with or without the group tax, which has no impact on this matter). From the perspective of the absolute stone count, this makes sense: the last dame is a point, and there is no more benefit for white in its last move. To not play that dame point for black, means to loose that point. When white plays the pass stone, that difference becomes accurately codified into the remaining territory for both. If black omits merely as a counting short cut, white’s territory is one larger than it should have been, even though that black stone itself later gets ignored in the counting.

This leaves the question of why they didn’t play the dame points. Perhaps the one who started the game first, then withdrawing his last useful move was a form of komi, since komi points back then where not awarded. It restores a balance, so that each side plays an equal amount of useful moves. It seems courteous for the one who played first, to in the end not force the other to have to give up a pass stone (or play inside their own territory), when there is no longer a useful move. If the goal is to allow the player who played second in the beginning the benefit of getting the last useful dame playing move if he gets it, then you could omit playing not only the last dame point but more or all of them, since the score doesn’t change when both players fill in a dame point. If the scoring is done by putting back in the pass stones (prisoners) and count the territory (enclosed empty space), to prevent confusion these empty dame points could be randomly filled in.

(Corrections welcomed.)

2 Likes
13

You are doing the “long way around” of the group tax, instead of adding stones in every group, just put aside two stones, for each extra group one side has, and give them to the other side.

Or using the original Ming-Qing ancient rules with half-counting, so only need one extra stone for each extra group, and compared to half of the board (hence only need to count one side and know the result, without even needing to backfill, or ensure the same amount of stones being played). There’s a reason why the ancient “road scoring” evolved into group tax half counting of area.

1 Like
14

As far as I understood things, a major reason to stop using pass stones as a way to ignore the stones and reach the total absolute stone count with a faster method, is to avoid cheating. Someone from China even said this in person at our club recently.

The way here argued of adding the group tax is rather a short way around, rather than a long way around, isn’t it ? It is a trivial matter to fill in a few eyes with stones, and you will hardly go wrong with it (unlike when you try to count to 180, or wrap your head around what to do with the excess eyes or seki points in Chinese counting if you want to add the group tax; or get your opponent to agree to it ;-).

I have been scoring quite a few of my games now according to the Chinese rules: counting half the stones then comparing to 180,5. Making easily counted squares first, and then also with the group tax in the way you described (subtracting half of the excess eyes to the other, which is indeed how it works out). I find it quite easy to do, but some people seem to disagree.

You could argue: there is a reason the Tang Dynasty rules of efficient counting was developed, just like you say there is a reason the Ming - Qing rules where developed, since the most basic rules of the game may have been there earlier. Complexity developing from simplicity ? The airplane comes after the ox cart ? A simple origin may have involved filling up the entire board except the eyes, and then counting all the stones. This was likely on smaller boards (the 17x17 board certainly was smaller) where this was not so tedious, and/or the play was tight anyway so the board may have been quite full. Play at distance seems to have been discovered later as a winning method, or so I have read. I wonder if they just re-arranged the stones after a full play to see who reaches over half, because they could not formally count yet (which was normal in the past). This is speculation I suppose, since the most ancient rules seem to be increasingly unknown.

Playing the board completely full besides eyes and then counting all stones, seems to be a good first hypotheses about original Go rules and counting. In my experience it mattered a lot how much time it took to explain someone the rules of a game, to how willing they still where to try it. The shorter the better. Keeping pass stones around to make the stones on the board equal and then counting only the spaces, seems to be a complexity and understanding which develops after the game has been played for longer ? It is less intuitive, but once you discover it gives the same result and is easier to do, it may have caught on ?

Then apparently the cheating started to become a serious problem, and the rules where changed again to the Ming - Qing variant (one color counting with group tax). Some counting efficiency was sacrificed, as the example I gave seems to illustrate, which seems to be in line with what most people think about this (area counting being more work than territory counting). Then under Japanese influence apparently, after the Qing Dynasty in modern times, the group tax (the underlying absolute stone counting) was lost.

The method I try to argue is not about Chinese versus Japanese counting, but rather Japanese counting versus Tang Dynasty / AGA (…) like counting, so that the counting efficiency which they both try to achieve (apparently) can be paired with solid and simple rules while also respecting the group tax, the absolute stone counting principle of the game.

It seems to me that the Tang Dynasty like method, with some kind of simple group tax method perhaps such as proposed, may be the better answer (if you prefer simple & elegant rules). If cheating is the bigger problem, then the best way to Go is perhaps one color counting and dealing with the group tax in the way you described; which seems to be relatively complex, a long way around if you will, but does result in a solid and elegant set of rules (unlike Japanese rules). Any corrections from anyone appreciated, thank you, because I’m also not too sure about any of this ;-).

1 Like
15

My comment of “long way around” was a satirical remark about the length of your reply, which can be summed up in one sentence - backfill extra stones representing eye space (beside other historical customs stuff).

And your new reply is so convoluted and full of conjectures, if I reply to all of them, it would be prohibited long. If you really want to talk about methods that can be applied as substitutes and give the same result as stone scoring (whether historical or simply hypothetical), we can start a new thread for it. And if you do, please separate all your major points in order, or it would be exceeding difficult for anyone to follow.

1 Like
16

You seem to be advocating a long way around of the group tax, by giving those eye filler stones first to the opponent, who is then to place them into the opponent’s groups, together with the prisoners / pass stones. They can be placed directly unto the board, coming out of the remaining stones in the bowl, which is clearer and faster.

The reason to also discuss an equal amount of stones to be played, together with the group tax and its methods, is because the group tax implements the simple principle of only the living stones played to the maximum are the points. The resulting rules seem to be simple (unlike Japanese rules), the counting method can be efficient (unlike Chinese counting). If you don’t like this principle, you might also dispense with the group tax (as in modern rules).

Your claims seem to not be argued and to be incorrect. Your text I could not comprehend due to your use satire, sorry. I didn’t know you are sarcastic, but assumed you referred to the topic honestly. You made an historical claim about the rules being changed “for a reason”, prompting me to start debating the issue in historical terms.

It is fine with me if the topic is moved to its own header. I propose we let the matter rest between us. I am sorry I wrote long, but I prefer clarity and correctness over hard to follow brevity with errors. Have a good day.

1 Like
17

the object of the game is to get the most stones on the board…