bogiesan wrote:Just noting the conversations I've been monitoring. Values I'm talking about are held by individual players, not the larger community. New go players, coming from gaming systems where running machine assistance is not seen as cheating so much as gaming the system and getting away with something, are going to bring their questionable behavior to go.
O tempora, o mores!
dfan wrote:Fox has a score estimator that is available to all players as a convenience. I don't see how using a feature openly provided by the site can be considered cheating. People who don't like playing against players who use it should play on a site that doesn't have that feature - there are plenty of them.
If I were playing against someone on a site without a score estimator feature and they made a copy of their game and ran a score estimator on it, I would indeed consider that cheating.
Me too.
Boidhre wrote:I don't think it is a cheat (ok, maybe in blitz games it's a bit much but longer games you are just saving a little effort really compared to counting regularly in your opponent's time).
If your score estimator gives the same results as human counting, it sucks the big one.
Pre-AI books on positional evaluation focused on counting secure territory. That can be useful information, because it indicates how much territory you need to make in less well defined areas of the board.
But it is not really score estimation. For real score estimation you want KataGo or the like.
To me, positional evaluation is an important skill to develop, and involves rather more than counting. In a casual game it should take a couple of minutes. IMO, following Znosko-Borovsky's chess advice, it should be done, as a preliminary to making a plan, only a few times in a game. That may not be so easy with common online time controls. {shrug} Regular counting during a game is a waste of time.
On the one hand, using an accurate score estimator is a crutch, which can hamper development, if you rely upon it rather than developing your own skills. And, OC, using a sucky score estimator sucks.
But thinking about it now, if I were to host an go server, I think that I would offer an accurate score estimator to the bottom half of users, maybe to the bottom two thirds. Instead of thinking of it as a crutch, I would think of it as a learning aid. We don't expect secondary school students to figure out everything for themselves. They get to look things up.
Here is how you might use an accurate score estimator as a learning device. At certain points in the game, estimate the score yourself, and then check your estimation against that of the score estimator. Then, following Znosko-Borovsky, make a plan. Later you can check to see how well you are doing, and maybe make a new plan. This way you get to practice both score estimation and planning, and you get real time feedback on how well you are doing.
