Life In 19x19

Doesn't AlphaGo answer that question? Train on KGS 1-9d game data (I'm still not clear if they used GoGoD or go4go with pro games as training data for Fan Hui or Lee Sedol version, there have been conflicting reports), play yourself and improve and then beat Lee Sedol. Deepmind have said they are ...

Fair enough, wow, I somehow managed to read right over that. That's awesome.

I'm really curious then about how strong the computer can become playing against itself and optimizing. It would seem to say something very interesting about people learning by playing against people their same strength.

I wonder if it's possible to train a policy network on kyu players in order to get bots that are useful for people learning.

There's a lot that gets in the way of comparison.

The old games were filled with genius, but the time controls were also significantly longer. I think if you took Lee Sedol or Lee ChangHo in their prime and gave them 5x time, you'd see a similar/stronger calibre.

Champions are statistical outliers ...

Greetings from not too far away from you, down in Curacao!

While I kind've like the special cases of "no result", positional superko just always seemed more elegant.

Maybe possible that the rest of the Japanese pro scene is catching up? Has Yamashita Keigo played in international competition recently?

Even if the processing power is similar, if the version you're using isn't parallelized over the cores, it might not matter, right?

But better than a particular coordinate system, I think, is to identify the rules that make a good coordinate system.

Golden Rule:
A good coordinate system should mirror the underlying symmetries in the topology.

1-19, 1-19 and A-T, A-T coordinates do not do this.

For what it's worth, I like it.

I think the full numbered board is effectively useless for the previously stated reasons. It's too big, and it's nonintuitive. A better perspective, in this case, from the all important corners, is ideal.

Just as I would not try to solve for spherical wave ...

Being based in Curacao now, it occurred to me that according to the AGA bylaws, I won't be in a position to compete for international representation for the foreseeable future. So, I started to think about go on the island. I wonder if Curacao is allowed to send a representative, and if I would ...

I feel like it has to be tengen. On a 9x9, tengen is brutal, splits all of white's plans, I feel like the same is true on a 19x19, but the method is considerably harder.

let me re simplify: Get all your pieces into the action. Just like in Go how you can't afford to make slow plays, in Chess you can't afford to have pieces doing nothing. Yes, of course, but this amounts to saying "play good moves".
( Notice, also, you don't really mean ALL your pieces in action ...

I like area scoring as the official method, if there are any problems it steamrolls right through (Examples: Did we accidentally lose one of our prisoners? was the stone replayed or not? is there a complicated endgame position?). But in general, I resolve games with no complications with territory ...

... And this long series of discussions is why I think we should just abandon territory scoring entirely. Why do we make this so hard on ourselves?

In any case, as Bill said, use area scoring, a point for each stone on the board, and a point for each empty space. If one side disagrees, resume play.