David J Wu

@OmnipotentEntity Thanks! I think `scoringRule` should be `AREA` instead since if you're tromp-taylor rules with a modification, then you're going to be area scoring as a baseline? @dergo853 - sorry...

Thanks for reporting this! What is going on is that "final_score" uses a smart evaluation that requires querying the neural net to determine dead stones and predict the final territories....

@monsterkodi - your problem is not the same problem as the one mentioned in this issue, and so far the behavior you've reported looks correct, it is only your configuration...

Can you provide an SGF?

@Chenvincentkevin - thanks for the analysis of that position. > It's a game against 60B played by 8 cards of 3090 for B and 4 for W(Not the complete game),...

> no matter how much reinforcement learning is done. I think I recall you expressing similar sentiments in the past, but if so, you might be surprised! :) In my...

> And here're some possible places for adjustment in a normal training game that I suppose if I understand what it means correctly (sorry if below I wrote wast your...

This is not a bug, see documentation for scoreMean at https://github.com/lightvector/KataGo/blob/master/docs/GTP_Extensions.md, this choice was made because there were many old tools coded to use scoreMean when they should have used...

Keep in mind that that part of the reason the flag for larger board sizes by default caps to 29 is that all of the neural networks don't generalize upward...

Except for the very fact of including more values to report, in what way does it not conform right now? Is it the fact that winrate is reported as a...