On Mon, Jul 08, 2024 at 01:36:46PM +1200, Joseph Heled wrote: > I see (in the article) GNUbg 2-ply: 33.24 > Is that what we are comparing to? 16.74 is half. > Are you saying GNU got that much better?
Yes, error rates are approximately halved: 0ply chequer play cost -44.50 cube play cost -11.67 1ply chequer play cost -31.98 cube play cost -8.58 2ply wc chequer play cost -16.74 cube play cost -4.42 2ply supremo chequer play cost -16.11 cube play cost -4.42 3ply chequer play cost -12.66 cube play cost -3.80 4ply chequer play cost -9.29 cube play cost -3.51