Interested in measuring your Go Strength?

There are times where there are worse moves than passing

1 Like

Sure, I also thought about not bounding the score (-10000,10000), when the move is worse than pass/better than ai. In this case, we can’t define the possible range of score, since how bad a move can be is not easy to determine.

Choosing “pass” as “worst” is what I can think of so far, since searching through all point on the board to find the worst is not feasible in this moment.

Of course measuring between “worst” and “best” is just my preliminary thought. I’m open to any idea to improve / redesign the formula.

Why not score it based on katago score estimate, and rank users by some average of points lost per move?