The position below is one that Petter posted recently:
| White | 124 |
|---|---|
| |
| Blue | 64 |
| GNUBg Id: Z3MGAFi29w4AAA:cAk1AAAAAAAA XGID=abBBDDC----------b-c-bb-c-:0:0:1:52:0:0:0:1:1 | |
5/off is forced and there are three plausible candidates for the 2: 6/4, 5/3, and 4/2. Petter wanted to know why 5/3 topped his rollout.
I don't have a full answer but after some analysis, I do have a proposed explanation for why either 6/4 or 5/3 is better than 4/2 in these sorts of positions (where by "these sorts of positions" I mean that Blue has this exact position and White has an ace-point anchor, with a third checker either on the anchor or on the bar, and White's other checkers are "reasonably" positioned).
The main point is that the purpose of 5/3 is to prepare to clear the 5pt. Though we're taught from childhood to clear from the rear against an ace-point game, clearing the 5pt is often almost as good or even slightly better than clearing the 6pt. In the position above, the long-term safety of clearing either the 6pt or the 5pt is sufficiently beneficial that it's (just barely) worth risking 5 immediate blotting rolls with either 6/4 or 5/3, rather than avoid any immediate blotters but fail to prepare to clear either the 6pt or the 5pt with 4/2. After 5/3, Blue will clear the 5pt with rolls such as 21, 32, etc. The three checkers on the 6pt look weird, but don't actually produce much blotting danger after the 5pt is cleared.
The choice between 6/4 and 5/3 is harder, and requires more analysis than I have (so far) performed. It seems to depend on the details of White's position.