PR in one-point versus longer matches
Join Date: Aug 2019
Posts: 667
I've only played ~150 matches on Galaxy so far (so lol sample size), but I'm noticing a significantly lower error rate in my 1-point matches than in 3- and 5-point matches due to the lack of cube decisions (presumably I'm making bigger cube errors than checker errors). I'll have to export the match lengths to compute the exact PR for my 1-point matches, but I'm averaging 9.8 PR overall and ~7 PR in 1-point matches.
I'm curious whether others notice a similar trend. Do you have a significantly lower error rate with the cube out of play? Does the effect have a similar magnitude to what I described here?
Thanks!
Join Date: May 2004
Posts: 1,954
Yes, your error rate will be much lower in 1-point matches. You've subtracted out all the skill involved in cube decisions, and in decisions involving playing for the gammon versus playing to win, saving gammons, and so forth. You're just playing a very simplified version of real backgammon.
Join Date: Aug 2019
Posts: 667
It's even more dramatic of a difference than I thought -- I'm averaging 6.2 in one-point matches and 12.9 in 5-point matches (though still small sample size). Obviously I've got some work to do.
