[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

Bear-Off cube decision

Posted By: Hans Marius Eikseth
Date: Friday, 27 April 2018, at 6:45 a.m.

In Response To: Bear-Off cube decision (Mochy)

This is a nice problem, as it poses a challenge to the usual array of adjusted pipcount methods.

OTB: My bread-and-butter method would be iSight, which estimates the counts L: 37 pips + 1 for open 4 + 1 for open 6 (as opponent has them)=39 T: 44 pips

win% estimated as 80-L/3+2*(L-T)= 77%

and we know from other postings here that this overshoots - Gnu gives L 72% here. iSight is between 2 and 3 correction pips wrong here.

SO, how to recognise the problems of iSight here? Gap severity (six checkers on 5 with open 4), and gap persistence (stacks on odd numbers with gap on even number never self-correct) are key, and also the 6es wastage that Mochy points towards may be considered.

Then any adjustment must take the relative severities into account - as trailer her has a gap on 4 in the making (with six checkers on higher points).

So, with 9 checkers and around 40 pips, a 5 or 6 rolls bear-off is to be expected, and two rolls with six to consider punishing L for. However, iSight already punishes L with 1, so at most 1 correction pip more can be given here.

Can we formulate other corrections generally? Like, still 1 pip for the gaps, but then add 1 pip for each third checker more than opponent on a high point, in the spirit of the punishment for the one, two and three-points? 2 pips if the gap is extra wasteful/persistent?

If this can be applied to the 5-point stack as well, it would yield a correction of another pip (6 vs 2 checkers on a high point) or two.

Thoughts on these kinds of adjustments?

Messages In This Thread

 

Post Response

Your Name:
Your E-Mail Address:
Subject:
Message:

If necessary, enter your password below:

Password:

 

 

[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums is maintained by Stick with WebBBS 5.12.