Ok, I badly mis-understood the implication. 3/2* is a genuinely awful play that Magriel would not have made. That said, I've seen plenty of genuinely awful plays advocated by respected authors from the pre-bot days so nothing would surprise me.
As to the question at hand, lifting wins more games and more gammons, so 4/3 is not wrong because it ignores gammon losses. It's just plain wrong period. The additional gammon losses (and there are a lot) just make it more wrong.
I just hope nobody dredges up anything I wrote in 1979 for analysis. Last week's stuff is bad enough.
        
 
         
 | | White is Player 2
score: 0
pip: 85 | Unlimited Game
Jacoby | pip: 127
score: 0
Blue is Player 1 |
|
| XGID=--aBCBBBA--------c-cBbbbbA:1:-1:1:51:0:0:1:0:10 |
| Blue to play 51 |
| 1. | XG Roller++ | Bar/20 8/7 | eq: -0.020 |
| Player:
Opponent: | 57.35% (G:5.57% B:0.13%)
42.65% (G:6.93% B:0.04%) | |
|
| 2. | XG Roller++ | Bar/20 4/3 | eq: -0.125 (-0.106) |
| Player:
Opponent: | 54.14% (G:5.31% B:0.12%)
45.86% (G:16.61% B:0.09%) | |
|
| 3. | XG Roller++ | Bar/20 3/2* | eq: -0.864 (-0.844) |
| Player:
Opponent: | 39.34% (G:3.54% B:0.08%)
60.66% (G:38.26% B:0.87%) | |
|
eXtreme Gammon Version: 2.10