DotsGo is the game that evolves from the special setup shown below, which resembles Dots and Boxes. Even though common Go rules apply, the game is a little strange. It makes heavy use of miais and hence is interesting for practicing the concept of miai.
Something might be interesting to note:
Play on an a would result in 2 liberties for both Black and White.
A black play on a b would result in 3 liberties, whereas a white one would result in 4. On c correspondingly, a white play would result in 3 liberties, whereas a black one would result in 4.
Either player can create a string with 5 liberties at d.
A black stone played on e would have the full 6 liberties, whereas a white one would only have 5 there. At f points it's vice versa.
I don't know... how would you analyze such a game? (extracted from a comment by aib)
Like any game of Go, it is important to understand how to get life. Various shapes found by playing this game include:
With , White must allow Black to play at "2", otherwise Black can run this one eyed group along the left edge to death.
If at a, at b and Black kills the White stones in gote. If at b, Black can potentially play "c" next threatening to play "d" if White doesn't play "d" or "a". It's sente, but Black must be happy with losing the chance to play at "a" for the extra outside push.
White at allows White have the option of attacking under Black's stones (a move played later in the game), while being able to connect out to one of the two circled White stones.
The area to the right side of the Black three-in-a-row is under a slight Black influence. Black can play at "a" and be confident that a second eye awaits (or White will find trouble).
There are two miai shapes which are useful when it comes to thinking about connecting:
As you can see, a three in a row facing another three in a row makes a connection (unless a lack of liberites mean that connecting is a bad idea). In this example, a third three-in-a-row angled so that it runs along the ends of the other two also makes a connection.
A three-in-a-row joined to another three-in-a-row enables the stone equidistant from the end points to be virtually connected. As shown in this diagram
To ensure that games between players of different DotsGo strength are exciting to both, White can give handicap to Black by passing for the first few moves (after the setup). However be aware, DotsGo ranks don't necessarily correspond with usual ones.
Dots Go is the balanced version of a position initially suggested at Global Seki.
blubb: I've played some dozens games of Dots Go at kgs. A 13x13 setup grid can be mutually created by the players within a minute. The actual game turnes out to be quite challenging, since hardly any of the Go knowledge I have gained so far is really useful here, regardless of the rules being exactly the same.
Anyway, I couldn't find a clear evidence yet if the first player tends to win or to lose, nor which komi may be appropriate.
Bill: blubb, I am curious about your impressions about the game. It seems to me that there are no kos, and superkos would be rare, as would sekis. It also seems to me that the tactics are simpler than regular go. For instance, there are no snapbacks or eye-stealing tesuji, and it is practically impossible to form an empty triangle. So strategy and whole board thinking are even more important in dots go than in regular go. Comments?
- blubb: Sorry, Bill, about the huge delay. At about the time when our game ended I got thoroughly distracted from Go alltogether, both time- and attentionwise. The game was very interesting and also special in that various decent players contributed and discussed their ideas and views live. Thanks for your participation.
I have never encountered a ko or superko throughout the 40 to 45 games I have played so far, whilst sekis seem to occur rather frequently (actually, even more often than in regular Go). The typical seki shape is a single "five in a row" eye with three foreign stones in the middle. Not much variation there. Local fighting tactics look different and indeed simpler, however there are some parallels, e. g. the bamboo joint discussed in the game.
Concerning complexity, I believe that the branching entropy (modulo symmetries) of a high level Dots Go game collection would be lower than the one of a similarly high level regular Go game collection using the same board size (referring to the empty points count). Anyway, I don't have a substantiated estimate of by what factor, yet. In my view, distant effects accross the Dots Go board tend to be more straightforward. I am not sure whether global thinking matters more, but due to stronger distant interactions and the apparently not so rich local tactical spectrum, I tend to think so, too.
Although in a game of Dots Go between skilled players, some regular Go stuff is quite unlikely to happen, hardly any is impossible. If Black passes while White captures all black stones and then fills the own eyes, Black obviously can enter an almost regular Go game tree by capturing the white dumpling (the tree is subject to a few additional superko restrictions). Therefore I suppose that the tree of Dots Go is quite a bit larger than the regular Go tree on a rectangular board with the same count of initially empty points.
Anonymous: This game has the same structure as the game "Bridg-it" or "Gale", described by Martin Gardner. In this game the objective is to connect two opposite sides, and the solution to this is trivial.
hk: Yup, but when the goal is to make territory, it doesn't seem nearly as trivial!
A few observation (perhaps obvious): Cutting is impossible.
To surround a group you use a single group (actually the minimum number of groups required to surround one group = the number of places the surrounded group connects with the sides). When capturing anything you always get at least one eye. The dotgo bamboo joint is not a normal bamboo joint, sometimes when connecting in sente due to shared liberties with an outside group you can get an extra 2 point eye.
See also DotsGoOngoingGame.