Thank you for your reasoning! Indeed, there is a huge QF due to my question and due to the simultaneous presentation. I overlooked that.
Position 2 is an original postion which occured when I played a match (unlimited session) with XG2. The other positions are variations of position 2. They were created "artificially".
Rollouts:
Position 1:
                         |
| ||||
| XGID=--aB--D-CC-AbB---b-ccb-b--:0:0:1:41:0:0:3:0:10 | |||||
| Blue to play 41 | |||||
| 1. | Rollout1 | 9/5 6/5 | eq: +0,459 | |||
| ||||||
| 2. | Rollout1 | 11/7 8/7 | eq: +0,338 (-0,121) | |||
| ||||||
| 1 2592 Games rolled with Variance Reduction. Dice Seed: 7134413 Moves: 3-ply, cube decisions: XG Roller | ||||||
eXtreme Gammon Version: 2.10
To me, this position is the easiest to solve. Making the golden 5-point uses the checkers very efficiently. The valuable 9-point doesn't have to be cleared in this context.
Position 2:
  |
| ||||
| XGID=-a-B--C-CB-AbD---b-ccb-b--:0:0:1:41:0:0:3:0:10 | |||||
| Blue to play 41 | |||||
| 1. | Rollout1 | 11/7 8/7 | eq: +0,108 | |||
| ||||||
| 2. | Rollout1 | 9/5 6/5 | eq: +0,022 (-0,087) | |||
| ||||||
| 1 2592 Games rolled with Variance Reduction. Dice Seed: 24754213 Moves: 3-ply, cube decisions: XG Roller | ||||||
eXtreme Gammon Version: 2.10
This is the original position which occured in the match. I made the barpoint, and flagged this move using CRTL+F in order to analyze it afterwards, since I wasn't completely sure.
Yet, making the barpoint leaves no blots, maintains the 9-point, and results in a small 4-point prime which provides a probably good blocking structure.
Position 3:
|
| ||||
| XGID=--aB--C-CB-AbD---b-ccb-b--:0:0:1:41:0:0:3:0:10 | |||||
| Blue to play 41 | |||||
| 1. | Rollout1 | 9/5 6/5 | eq: +0,151 | |||
| ||||||
| 2. | Rollout1 | 11/7 8/7 | eq: +0,120 (-0,032) | |||
| ||||||
| 1 2592 Games rolled with Variance Reduction. Dice Seed: 24754213 Moves: 3-ply, cube decisions: XG Roller | ||||||
eXtreme Gammon Version: 2.10
This is a variation of position 2 insofar as White's backchecker stepped up one pip from the 24-point to the 23-point. Over the board, I would definitely make the barpoint for the same reasons as given in position 2. Yet, this is wrong, as the rollout indicates that making the 5-point is better. To be honest, I still do not understand it.
One reason for the 5-point might have something to do with the fact that Blue will have a stripped position which makes it more difficult to make the 5-point subsequently (same situation like in position 2). Moreover, to make the 5-point may be justified, since White has only a limited number of hits (52 and 54). I admit that this explanation is not very suitable.