FIBS Board backgammon forum

So I think I've successfully shielded my eyes from the bot results long enough to weigh in on this one.  Actually, I didn't know bot-answer to the other recent cube decisions either (and still don't) , it's just that I'd seen enough to color my opinion so I was reluctant to post anything or vote; that said, had I posted an opinion anybody who followed it blindly would be a fool - a glimpse at equity numbers coupled with a mediocre understanding of how to apply them != actionable intelligence.  Probably just the opposite.  So what follows is my opinion not based on bot results, but if it has been subtly influenced, it's been filtered to the point where you shouldn't take my word for it.  I'm just saying...

Enough navel gazing, on to my analysis:

This looks like a ace point game to me. And a well timed one at that.  A well timed ace point game leads to about 20% wins and 20% gammon losses for the holder of the ace point (i.e. Zorba)  Th usual rule is that you have to win twice as many gammons as single-point losses to play on, and here it's 1-1.  So push the cube and take the single point.


Double.

This is marginally too good. Cubing here would not be a very big error.  Kudos to the Herd for correctly analyzing a close decision.



////////////////////////////////////////////////////////////
2-ply analysis:
Cube analysis
2-ply cubeless equity  +0.900 (Money:  +0.803)
 0.741 0.379 0.013 - 0.259 0.069 0.002
Cubeful equities:
1. No double            +1.045
2. Double, take         +1.913  ( +0.868)
3. Double, pass         +1.000  ( -0.045)
Proper cube action: Too good to double, pass (4.9%)

///////////////////////////////////////////////////////////////////////////

0-ply rollut

Rollout details:
Centered 1-cube:
 0.740 0.382 0.020 - 0.260 0.068 0.004 CL  +0.896 CF  +1.040
[0.001 0.002 0.002 - 0.001 0.001 0.001 CL   0.005 CF   0.010]
Player Zorba owns 2-cube:
 0.746 0.380 0.040 - 0.254 0.064 0.005 CL  +2.259 CF  +2.015
[0.002 0.003 0.003 - 0.002 0.001 0.001 CL   0.014 CF   0.017]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 903967096 and quasi-random

dice
Play: 0-ply cubeful prune [expert]
Cube: 0-ply cubeful prune [expert]



1. No double            +1.040
2. Double, take         +2.015  ( +0.975)
3. Double, pass         +1.000  ( -0.040)
Proper cube action: Too good to double, pass

See that lone voice dissenting up there..that is me that is, also feeling this was a bit too good - and that I would rather wait and play on..but I know I do that too often for my own good  ;)

Buuuut - can you explain your comment - you say it is marginally too good - then that the herd correctly decided to cube...

we stole us a game B)

Rollout details:
Centered 1-cube:
  0.787 0.133 0.008 - 0.213 0.026 0.001 CL  +0.723 CF  +0.963
[0.001 0.001 0.000 - 0.001 0.000 0.000 CL   0.002 CF   0.003]
Player Zorba owns 2-cube:
  0.805 0.133 0.019 - 0.195 0.022 0.001 CL  +1.799 CF  +1.607
[0.001 0.002 0.002 - 0.001 0.001 0.000 CL   0.010 CF   0.012]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 903967096 and quasi-random dice
Play: 0-ply cubeful prune [expert]
Cube: 0-ply cubeful prune [expert]




1.  Double, pass         +1.000
2.  Double, take         +1.607  ( +0607)
3.  No Double            +0.963 ( -0.037)
Proper cube action: Double, pass

So, double/pass is the correct cube action (according to 0-ply rollout anyway), the Herd made the right choice, but it was a fairly close decision. Sorry about the faulty numbers, I have no idea where those numbers came from.