I just uploaded a 9x9 game i played the other day to have it reviewed by KataGo. Since it happened on a real board, i had to replay it from memory in an SGF-editor (I used Sabaki).
The analysis offered seems quite off however and i was wondering, why this is. Is it because its a 9x9 game or because of some other reason, for example that it’s not a “real” game file, but rather a replay?
Here is the SGF-file I uploaded. @jlt Hmmm … If thats actually the reason, I find it a bit weird. You might as well make 9x9 analysis availible for higher tiers only, if it shows nonsensical moves like this one
Games are randomized in board size, with 37.5% of games on 19x19 and increasing in
KataGo’s main run to 50% of games after two days of training. The remaining games are
triangularly distributed from 9x9 to 18x18, with frequency proportional to 1, 2, . . . , 10.
And 7x7 and 8x8, in the years following that paper.
Which is why we can now “solve” 7x7 and 8x8 better than any prior attempts at determining the best play on those board sizes (https://katagobooks.org/).