| |
BGonline.org Forums
Very nice, but just a tad conservative for raw pip counts
Posted By: Timothy Chow In Response To: Very nice, but just a tad conservative for raw pip counts (Nack Ballard)
Date: Thursday, 4 February 2010, at 10:13 p.m.
Nack wrote:
I think Tim is probably in agreement as well, though since he brought up the 24pt 23pt 22pt = 69 pipcount example, I am uncertain whether he is supporting 15 as the number of checkers used.
Actually, Bob Koca's surprising (to me, anyway) examples of positions A and B that have very different cube actions despite both sides having epc's of 50.00 make me think that there is a subtle but important issue here that we have been overlooking.
Prior to seeing Bob's examples, this is what I would have said: Why don't we sidestep the whole issue of minimum wastage versus typical wastage, or 15 checkers versus any number of checkers, by just using epc instead of raw pip count? Trice says in his book that his table applies to epc's as well as to raw pip counts of low-wastage positions. The concept of epc automatically takes wastage into account. So let's just develop a table for epc's, and worry separately about developing heuristics for estimating epc's.
To put it another way, when trying to decide if 67 vs. 75 is the point of last take, let's pit a position with an epc of 67 against a position with an epc of 75. We should get the same answer whether the 67-epc position or the 75-epc position happens to be a low-wastage position or not. So then we don't have to argue about which 67-epc or which 75-epc position to use.
I still think that this is basically the right approach. That is, if we're going to work on this problem, then the thing to do is to see if Trice's table is basically correct if we interpret it as a table of epc values.
Having said that, I now need to add a caveat in the light of Bob Koca's examples. These examples show that wastage isn't the only issue to consider. If it were, then epc's would solve our problems, since epc's take wastage into account automatically. Unfortunately, it appears that the variance in the number of rolls that it takes to bear off (and not just the average value, which is what epc measures) affects the correct cube action. This is a little annoying. It means that a truly "platinum standard" table must also take into account this variance. For now, I'm inclined to ignore this issue and just hope that most "normal" positions with a given epc have more-or-less the same variance. But if we want to be "scientific" about it, we should really investigate this question.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.