OTB I make the ace pretty quick, as generally class all these types of holding games as don't double/build a board asap positions, but can see here that running both out to the 13 and 14 with the blot on the 21, weakish board and duplicate twos might be right somewhere deep in the neural net nuance where mere mortals fear to tread.

Black is going to break before white leaves a shot, so that plan won't work.

Waiting for doubles is still possible. Black is likely to get two more chances, maybe three. I'll give that about 25% chance. If it fails, white could still miss the shot black leaves. I'll give that another 25%, for 50% chance total.

Moving both back men treats the position as pay now or pay later. It's a pure double shot, with a return shot that won't be there later. That's probably about 60% total.

So I think running is best, which feels surprising.

Black is up 19 pips after the play and White has a blot which is hard to cover, so 18/13 18/14 looks clear. I'd even make this play with much less of a racing lead.

If White's 4-point were covered, say by moving the spare on the 3 back to the 4-point, then I'm not leaving the bar-point until I have to.

Hello,

Over the board I thought 18/13 18/14 would lose too many gammons and these 20 shots felt too risky.
I still can not believe that it`s a 0,121 blunder not to leave the anchor.


1. Rollout¹ 18/14 18/13 eq:+0,192
Player: 54,58% (G:7,80% B:0,13%)
Opponent: 45,42% (G:4,32% B:0,12%)
Confidence: ±0,006 (+0,186..+0,199) - [100,0%]
Duration: 4 minutes 09 seconds

2. Rollout¹ 6/1 5/1 eq:+0,072 (-0,121)
Player: 51,77% (G:3,44% B:0,06%)
Opponent: 48,23% (G:3,27% B:0,09%)
Confidence: ±0,008 (+0,064..+0,080) - [0,0%]
Duration: 6 minutes 03 seconds

3. 2-ply 6/2 6/1 eq:+0,018 (-0,174)
Player: 50,75% (G:1,56% B:0,00%)
Opponent: 49,25% (G:2,59% B:0,03%)

4. 2-ply 7/2 6/2 eq:+0,009 (-0,183)
Player: 50,20% (G:1,92% B:0,01%)
Opponent: 49,80% (G:2,40% B:0,02%)

5. 2-ply 7/3 6/1 eq:+0,003 (-0,189)
Player: 50,05% (G:2,04% B:0,00%)
Opponent: 49,95% (G:2,57% B:0,02%)


¹ 1296 Games rolled with Variance Reduction.
Moves: 3-ply, cube decisions: XG Roller


eXtreme Gammon Version: 2.10, MET: Kazaross XG2

I wouldn't worry too much about undoubled gammons when a play makes you a favorite to win the game.

If you think running is risky, you should see what happens when you stay. Your opponent's board is about to get stronger, and you're not going to be able to wait until he clears out all his outfield points before you're forced to run. Would you rather run now, when he has three home board points, or later when he has four?