I agree. And it is easy to implement via code, just move all the board lines, lets say, “from A1 to Z0”. I will add this feature in my own repo too!
I knew that QR codes were inspired by Go, but I didn’t realise that it made use of this particular variant
See 0.45 and mainly from 10.10
Playing on the intersections works on any graph, even if not planar — I prefer that.
But if you analyse play on the complete graph Km (perhaps not a game of skill) it seems a lot easier with symmetric superko: you may not repeat a position equivalent by a symmetry of the empty board to a previously played position. In that case a position may simply be encoded by the numbers of Black and White stones on the board, it is definitely not a game of skill, and the game always stops by superko (as indeed it does on Km with any sort of superko), and ends in seki unless there is only one color on the board. The course of the game is entirely determined by m and may be represented as a line zigzagging diagonally through a right-angled triangle, dropping back orthogonally when it hits the hypotenuse, and finishing when it would re-reach a point. If I have not got muddled, White wins for m ∊ [0,5], but Black for m = 6 and White for m = 7. I see no obvious pattern, but perhaps drawing a few more cases would make it clearer, though it could be chaotic. However, it does seem that m = 2ⁿ is always a win for White.
But I have not thought about colour-symmetric superko, where colour-reversed repeats are also forbidden!