This has nothing to do with Janowski.
The resason for the difference is that XGR(+(+)) are rollout and hence have some variance. Here the equity (7-ply) are -0.2611/-1.4121. Because the position is a beaver the variance on the D/B is inflated four fold.
     
 
      
 | | White is Player 2
score: 0
pip: 5 | Unlimited Game
Jacoby Beaver | pip: 18
score: 0
Blue is Player 1 |
|
| XGID=--BBB-------------------e-:0:0:1:00:0:0:3:0:10 |
| Blue on roll, cube action? |
| Analyzed in XG Roller++ | No double | Double/Take |
| Player Winning Chances: | 39.674% (G:0.000% B:0.000%) | 39.674% (G:0.000% B:0.000%) |
| Opponent Winning Chances: | 60.326% (G:0.000% B:0.000%) | 60.326% (G:0.000% B:0.000%) |
| Cubeless Equities | -0.2065 | -0.4130 |
| Cubeful Equities |
| No double: | -0.2605 | |
| Double/Beaver: | -1.3797 (-1.1192) | |
| Double/Pass: | +1.0000 (+1.2605) |
| |
| Best Cube action: No double / Beaver |
| Percentage of wrong pass needed to make the double decision right: 32.0% |
eXtreme Gammon Version: 2.01.175.pre-release
 
| | White is Player 2
score: 0
pip: 6 | Unlimited Game
Jacoby Beaver | pip: 18
score: 0
Blue is Player 1 |
|
| XGID=--BBB-------------------f-:0:0:1:00:0:0:3:0:10 |
| Blue on roll, cube action? |
| Analyzed in XG Roller++ | No double | Double/Take |
| Player Winning Chances: | 39.674% (G:0.000% B:0.000%) | 39.674% (G:0.000% B:0.000%) |
| Opponent Winning Chances: | 60.326% (G:0.000% B:0.000%) | 60.326% (G:0.000% B:0.000%) |
| Cubeless Equities | -0.2065 | -0.4130 |
| Cubeful Equities |
| No double: | -0.2646 | |
| Double/Beaver: | -1.3929 (-1.1284) | |
| Double/Pass: | +1.0000 (+1.2646) |
| |
| Best Cube action: No double / Beaver |
| Percentage of wrong pass needed to make the double decision right: 32.0% |
eXtreme Gammon Version: 2.01.175.pre-release