(I've fixed the attachments)
In my experience, these kinds of positions (make the offensive or defensive point) are extremely sensitive to the score. The difference of only one point can create an abrupt transition between strategies. It took considerable rolling out to completely make sense of this position.
  White is Player 2
score: 0 pip: 145  7 point match  pip: 138 score: 0
Blue is Player 1  
XGID=bBBCCAcBccbAbA:0:0:1:31:0:0:0:7:10 
Blue to play 31 
1.  Rollout^{1}  8/5 6/5  eq: +0.431 
 Player: Opponent:  59.96% (G:23.35% B:1.72%) 40.04% (G:14.58% B:0.94%)  Conf.: ± 0.005 (+0.426...+0.436)  [100.0%] Duration: 43 minutes 06 seconds 

2.  Rollout^{1}  24/21 11/10  eq: +0.399 (0.032) 
 Player: Opponent:  58.43% (G:19.37% B:0.58%) 41.57% (G:8.57% B:0.41%)  Conf.: ± 0.003 (+0.395...+0.402)  [0.0%] Duration: 57 minutes 43 seconds 


^{1} 15552 Games rolled with Variance Reduction. Dice Seed: 58918962 Moves: 3ply, cube decisions: 4ply

eXtreme Gammon Version: 2.10, MET: Kazaross XG2
  White is Player 2
score: 0 pip: 145  7 point match  pip: 138 score: 1
Blue is Player 1  
XGID=bBBCCAcBccbAbA:0:0:1:31:1:0:0:7:10 
Blue to play 31 
1.  Rollout^{1}  8/5 6/5  eq: +0.4058 
 Player: Opponent:  59.97% (G:23.28% B:1.75%) 40.03% (G:14.43% B:0.87%)  Conf.: ±0.0034 (+0.4024...+0.4091)  [90.9%] Duration: 1 hour 55 minutes 

2.  Rollout^{1}  24/21 11/10  eq: +0.4029 (0.0029) 
 Player: Opponent:  58.44% (G:19.31% B:0.61%) 41.56% (G:8.55% B:0.42%)  Conf.: ±0.0025 (+0.4004...+0.4054)  [9.1%] Duration: 2 hours 25 minutes 


^{1} 31104 Games rolled with Variance Reduction. Dice Seed: 58918962 Moves: 3ply, cube decisions: 4ply

eXtreme Gammon Version: 2.10, MET: Kazaross XG2
  White is Player 2
score: 0 pip: 145  7 point match  pip: 138 score: 2
Blue is Player 1  
XGID=bBBCCAcBccbAbA:0:0:1:31:2:0:0:7:10 
Blue to play 31 
1.  Rollout^{1}  24/21 11/10  eq: +0.3175 
 Player: Opponent:  58.43% (G:19.29% B:0.64%) 41.57% (G:8.45% B:0.39%)  Conf.: ±0.0036 (+0.3140...+0.3211)  [100.0%] Duration: 1 hour 21 minutes 

2.  Rollout^{1}  8/5 6/5  eq: +0.2917 (0.0258) 
 Player: Opponent:  59.94% (G:23.09% B:1.87%) 40.06% (G:14.54% B:0.81%)  Conf.: ±0.0043 (+0.2875...+0.2960)  [0.0%] Duration: 1 hour 03 minutes 


^{1} 15552 Games rolled with Variance Reduction. Dice Seed: 58918962 Moves: 3ply, cube decisions: 4ply

eXtreme Gammon Version: 2.10, MET: Kazaross XG2
As an example in this position at the beginning of a 7pt match (7A,7A), making the 5 point Is clearly correct, with a one point lead (6A7A), the plays are equal, with a two point lead (5A,7A), the anchor is clearly correct.
As similar pattern emerges from a five or six point match. Two exceptions to this are 4A, 4A and unlimited games. There the anchor is with 2% of the 5 point. Other interesting scores are Crawford games with you opponent odd away. Then the two plays are equal with the 5 point slightly ahead.
I have an interesting visualization as a pdf, but I don't know how to attach it.