herd recube decision

[stog]

herd recube or wait?

[ah_clem]

Spoiler

elmicho's live takepoint is about 18% and the gammon value is about .2.

I doubt we are anywhere close to 72% GWC and we don't win very many gammons from here, and even if we do with the cube at 4 they are not worth much.  It's fairly volatile, so it looks like a fairly comfortable take.

The thing is, if we shake another six from the tree we probably lose our market. Or if elmicho rolls a 25, 26, 15, 16, 35, 45, 65 or any number of other ugly rolls. That's a lot of market losers.  Maybe the take is not so easy after all.

I say ship it and make it his problem.  At a minimum, if he takes and shakes  42 or 32 he'll have a tough decision to make.

[close]

[elmicho]

Spoiler

After the 2-cube I had a very comfortable position - most of my rolls played very well, and I had a lot of time to distribute them. 44, 66 and 66 put an end to that welth, and now the herd has a marginal cube. I think it is right there on the edge - to cube or not to cube.
Anyway - I have a trivial take.

[close]

_{spoiler tag added - Zorba}

[elmicho]

wealth - I meant to type  )

[Zorba]

Spoiler

Looks like a good redouble, we're good favourites now. We've managed to escape one checker, el_micho is a in a tough spot to escape his two checkers, he is crunching while our blockade can last for a few rolls probably. Rolling a six means losing our market, but even if we don't, el_micho could roll something bad.

[close]

[ah_clem]

rollout

Spoiler

gnu analysis calls this no double by about .03, but a full rollout says it's a cube by about .04.  More cycles or XG might reverse this, but I think this is  a good practical double regardless. For one thing, some may drop. For another, it's better to cube a little early occasionally than miss cube after cube by being more cautious.

The take is pretty clear.


Cube analysis
Rollout cubeless equity  +0.387 (Money:  +0.384)

Cubeful equities:
1. Double, take         +0.758
2. Double, pass         +1.000  ( +0.242)
3. No double            +0.717  ( -0.041)
Proper cube action: Redouble, take

Rollout details:
Player DeeDee owns 2-cube:
  0.671 0.125 0.004 - 0.329 0.087 0.001 CL  +0.387 CF  +0.717
[0.001 0.001 0.001 - 0.001 0.002 0.000 CL   0.004 CF   0.006]
Player gnubg owns 4-cube:
  0.684 0.124 0.009 - 0.316 0.100 0.003 CL  +0.934 CF  +0.758
[0.002 0.003 0.002 - 0.002 0.004 0.001 CL   0.010 CF   0.011]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 736285877 and quasi-random dice
Play: 0-ply cubeful [expert]
Cube: 0-ply cubeful [expert]

[close]

[Zorba]

GNUbg 2-ply rollout:

Spoiler

Difference a bit larger here. Good cube, easy take. GNUBG 2-ply and 4-ply evaluations miss this cube.

Cube analysis
Rollout cubeless equity  +0.3954 (Money:  +0.3933)

Cubeful equities:
1. Double, take         +0.7822
2. Double, pass         +1.0000  ( +0.2178)
3. No double            +0.7238  ( -0.0584)
Proper cube action: Redouble, take

Rollout details:
Player The Herd owns 2-cube:
  0.6745 0.1261 0.0047 - 0.3255 0.0844 0.0022 CL  +0.3954 CF  +0.7238
[0.0008 0.0011 0.0006 - 0.0008 0.0013 0.0004 CL   0.0024 CF   0.0034]
Player el_micho owns 4-cube:
  0.6901 0.1133 0.0104 - 0.3099 0.0955 0.0059 CL  +0.9492 CF  +0.7822
[0.0010 0.0017 0.0014 - 0.0010 0.0031 0.0011 CL   0.0053 CF   0.0057]
Full cubeful rollout with var.redn.
2592 games, Mersenne Twister dice gen. with seed 736195168 and quasi-random dice
Play: world class 2-ply cubeful prune [world class]
keep the first 0 0-ply moves and up to 8 more moves within equity 0.16
Skip pruning for 1-ply moves.
Cube: 2-ply cubeful prune [world class]

[close]