> -----Original Message----- > From: Jonathan Kinsey [mailto:[EMAIL PROTECTED] > > >From: "Ian Shaw" <[EMAIL PROTECTED]> > > > >Have you gone for shading on relative equity, or on analysis > thresholds? > >It's not clear from your screenshots. The more I think > about, the more > >I'm convinced that the latter is better. > > The colours are fixed between 0 - VeryBad threshold. The > idea being that you'll associate certain colours with fixed > error rates.
Good plan. Ignoring the doubtful and bad thresholds makes sense, because that will gaurantee you a linear scale between 0 and the cut off point. I think it might be a bit better to extend the range so that all blunders aren't the same colour. With your implementation, blunders that we reasonably make in actual play will be the same colour as heinous 0.7 errors that no one in their right mind considers. Personally, I consider anything over 0.1 a blunder, but I still make 0.2 errors. You would probably get better information if your upper limit was VB + 0.1, or VB * 2 (arbitrary numbers). -- Ian _______________________________________________ Bug-gnubg mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-gnubg
