         
 
         
 | | White is Player 2
score: 0
pip: 146 | Unlimited Game
Jacoby Beaver | pip: 160
score: 0
Blue is Player 1 |
|
| XGID=------E-CA-afD---c-e-A--A-:0:0:1:63:0:0:3:0:10 |
| Original position |
| 1. | XG Roller++ | 9/3 6/3 | eq: -0.263 |
| Player:
Opponent: | 41.60% (G:8.69% B:0.20%)
58.40% (G:8.43% B:0.27%) | |
|
| 2. | XG Roller++ | 24/18 21/18 | eq: -0.292 (-0.028) |
| Player:
Opponent: | 40.37% (G:1.71% B:0.05%)
59.63% (G:1.83% B:0.05%) | |
|
| 3. | XG Roller++ | 24/21 13/7 | eq: -0.300 (-0.036) |
| Player:
Opponent: | 40.18% (G:3.69% B:0.12%)
59.82% (G:3.61% B:0.11%) | |
|
The above diagram shows the XGR++ evaluation of Stick's original position. His rollout (not shown here) of [P @19 Z52], which is 2592 trials for P and @ and only 1296 trials for Z, is sufficient for determining best play but not for nailing down margins. I don't want to feel uncertainty about the variance or do a bunch of rollouts in order to make a point that can be made by less time-consuming means. Rather, to achieve an apples-to-apples comparison, we'll start with the above XGR++ eval [P @28 Z36].
The point of this post: Whether to anchor or not largely depends on the strength of the near-side-only alternative.
P (Point, 9/3 6/3) makes an inside point while unstacking the 6pt, attaining a solid blotless offensive formation that will shine if Blue is able to hit a shot on the next roll.
Suppose we lift Blue's 9pt builder to his 8pt. Now the XGR++ evaluation [@ Z12 &30], diagrammed below, swings the near-side play from being best by .028 to it being third best and a mistake of .030.
| | White is Player 2
score: 0
pip: 146 | Unlimited Game
Jacoby Beaver | pip: 159
score: 0
Blue is Player 1 |
|
| XGID=------E-D--afD---c-e-A--A-:0:0:1:63:0:0:3:0:10 |
| 9/8 variant |
| 1. | XG Roller++ | 24/18 21/18 | eq: -0.269 |
| Player:
Opponent: | 41.06% (G:1.72% B:0.05%)
58.94% (G:1.66% B:0.04%) | |
|
| 2. | XG Roller++ | 24/21 13/7 | eq: -0.281 (-0.012) |
| Player:
Opponent: | 40.75% (G:3.54% B:0.10%)
59.25% (G:3.50% B:0.09%) | |
|
| 3. | XG Roller++ | 13/7 8/5 | eq: -0.298 (-0.030) |
| Player:
Opponent: | 40.98% (G:8.58% B:0.21%)
59.02% (G:9.52% B:0.39%) | |
|
Finally, I decided to test 9/8 plus 6/7 for Blue. This evens out the +/- pip adjustment for more race precision, but it loses the forest for the trees, because it makes Blue's strongest near-side move even better than in the original position! Being able to cover the 7pt instead of having a blot there makes a big difference in the event Blue hits a shot. The combination of owning the 7pt plus 5pt-slot is better than owning the 3pt. See the [$ D40 Z45 @55] evaluation below.
Summary: In the main position, the strongest near-side play is best by .028. In the 9/8 variant, the strongest near-side play is a mistake of -.030 (swing of .058). Add 8/7 to 9/8, though, and the strongest near-side play switches back from -.030 to being best by a larger margin than in the original position: the 21pt is -.045 and the least-shot-provoking 18pt play is -.055. (Rollouts will alter the margins but not radically, and they will not change the best play conclusions.)
In short, the decision to anchor or not is relative to the strength of the near-side alternative.
| | White is Player 2
score: 0
pip: 146 | Unlimited Game
Jacoby Beaver | pip: 160
score: 0
Blue is Player 1 |
|
| XGID=------DAD--afD---c-e-A--A-:0:0:1:63:0:0:3:0:10 |
| 9/8 + 7/6 variant |
| 1. | XG Roller++ | 13/7 8/5 | eq: -0.241 |
| Player:
Opponent: | 42.49% (G:9.02% B:0.22%)
57.51% (G:9.06% B:0.31%) | |
|
| 2. | XG Roller++ | 13/10 13/7 | eq: -0.281 (-0.040) |
| Player:
Opponent: | 41.62% (G:8.27% B:0.23%)
58.38% (G:9.29% B:0.34%) | |
|
| 3. | XG Roller++ | 24/21 13/7 | eq: -0.287 (-0.045) |
| Player:
Opponent: | 40.53% (G:3.18% B:0.10%)
59.47% (G:2.96% B:0.06%) | |
|
| 4. | XG Roller++ | 24/18 21/18 | eq: -0.296 (-0.055) |
| Player:
Opponent: | 40.24% (G:1.63% B:0.06%)
59.76% (G:1.75% B:0.04%) | |
|
Supplement: In the original position (top), or for that matter in either of the variants, if Blue anchors on the 18pt, White's roll of 66 leaves a blot on her 8pt. This gives the 18pt play a specific tactical advantage that is not accounted for in a comparison of the general value of the 21pt vs 18pt anchor in this type of position. Giving either 8/13 or 13/8 to both Blue and White (and doing evals or rollout comparisons) could be one approach for trying to quantify that abnormal advantage.
Nack