Public
The graphs were dramatic. It's not often you can watch software quality shoot up, as measured by a key objective metric, after a community shift towards open source code and norms.
It would be interesting to see a similar analysis for Go programs.Jan 6, 2012
I wonder if that's possible. The graphs feature a sort of exogenous shock: the release of powerful open-source chess engines into a previously mostly proprietary closed community. In Go, I don't know of any sudden increases, whether the Go computer/AI community was closed, or whether the dan rankings are fine-grained like ELO is.Jan 6, 2012
Given that Go is a 2-player game with win/lose/draw outcomes, shouldn't Elo translate naturally to Go?Jan 6, 2012- Go differs from chess in some respects; ELO was invented for chess, so it may not generalize. For example, the entire system of komi and handicaps might defeat a simple application of ELO. Maybe someone has done it - I'm not an expert on either. (20 kyu or so at Go and I stopped playing chess before I got any ELO.)Jan 6, 2012
