Since the same seed wasn't used and also since the rollouts were only 5k I figured I'd go ahead and do this right and see what we have. 5k isn't enough in my opinion if you're looking to compare absolute equities. Even 15k still has a little give to it but it's a lot closer to the truth. I stopped the 24/20 play after 11k because it's the equity we're least concerned with and on top of that it lined up almost exactly with your GNU 5k.
Your 3 key equities from my rollouts are:
.080 - Equity if playing NO fourth deuce
.088 - Equity if playing 13/11 as fourth deuce
.063 - Equity if playing 22/20 as fourth deuce
This is a lot closer than your GNU results and leaves my questions dangling. Is the small difference because of the 2 pips or because 13/11 with the last deuce is a useful 2 to play?
                            |
| |||||||||||||||
| XGID=-B---bCBB-a-bE---cbe----A-:0:0:1:22:0:0:3:0:10 | ||||||||||||||||
 to play 22 | ||||||||||||||||
| 1. | Rollout1 | 24/22 13/11(3) | eq: +0.088 | |||
| ||||||
| 2. | Rollout2 | 24/20 13/11(2) | eq: +0.063 (-0.025) | |||
| ||||||
| 1 15552 Games rolled with Variance Reduction. Dice Seed: 1296 Moves and cube decisions: 3 ply 2 10996 Games rolled with Variance Reduction. Dice Seed: 1296 Moves and cube decisions: 3 ply | ||||||
is Player 2
score: 0
pip: 149 Money session
Jacoby Beaver pip: 141
score: 0
is Player 1 XGID=-B---bCBB-a-bD-A-cbe----A-:0:0:1:22:0:0:3:0:10 to play 22 1. Rollout1 24/22 15/11 13/11 eq: +0.080 Player:
Opponent: 52.98% (G:7.70% B:0.29%)
47.02% (G:10.47% B:0.38%)Conf: ± 0.005 (+0.075...+0.085)
Duration: 4 hours 02 minutes 1 15552 Games rolled with Variance Reduction.
Dice Seed: 1296
Moves and cube decisions: 3 ply