Game 5, move 7 : Forum Cube?

[diane]

Well, rkb danced with 44, so....any sentient life forms out there want to do this?  

zLaDAVBsG8UxAA:cAmgACAAGAAA

[spielberg]

Spoiler

Far too easy a take and recube for the match. The gammon chances are better on our side too and given the score our gammon chances are worth more than his

[close]

[diane]

Spoiler

Far too easy a take and recube for the match. The gammon chances are better on our side too and given the score our gammon chances are worth more than his

[close]

Spoiler

Never mind easy take, I don't want him dropping this!! 

[close]

[dorbel]

Spoiler

Easy Take spiel? I don't think so. This is a big pass, but if you knew that you could get a take, it's probably right to double! RKB needs 25% winning chances (cubeless) to make a take correct. I don't think that he has that here, or even close. Is a double correct then? Over the board I would keep playing on for a closeout and a gammon. If rkb manages to anchor, we could think again. Of course if he manages to enter and hit a fly shot we probably won't have a cube at all, so we don't need to play too big going for the gammon.
This is the type of position that occurs frequently, particularly with all the three pointers that we play, so learning how to handle the cube and the checkers at this score pays big dividends.

[close]

[ah_clem]

Spoiler

As I said before, at this match score, the window between the DP and too good is narrower than usual. At even match score, I'd double here, but gammons are more valuable than usual and recubes are more costly. 

So basically, I'm uncomfortable with a drop (what if we played on and won gammon?) and I'm uncomfortable with a take (zOMFG if we lose we lose  the match) so I'm afraid to turn the cube.

Rational? no.  I should break out the MET and try to evaluate it cooly.  BUt where I am right now is "hold the cube cause I'm not sure what to do."

no double.


Now I'm going to read what others have to say and se if I change my mind.

[close]

[diane]

This is a lot less clear cut than earlier decisions, so I will leave this running longer - maybe ah_clem can convince a few more before he changes his mind 

[ah_clem]

gnubg rollout.  I've been holding back on looking at bot results because I didn't want to influence my cube decisions.  Now I think I can look without feeling like I'm cheating.  If you feel differently, don't look. (c:

Spoiler

Too good.  Cubing would be a whopper.  Did we miss a correct double somewhere along the way, or did we go right from no double to too good?  I'll try to rollout some of the past couple of decisions to see.

Cube analysis
Rollout cubeless equity  +1.037 (Money:  +0.858)

Cubeful equities:
1. No double            +1.137
2. Double, take         +1.390  ( +0.253)
3. Double, pass         +1.000  ( -0.137)
Proper cube action: Too good to double, pass (35.1%)
Rollout details:
Centered 1-cube:
 0.765 0.362 0.022 - 0.235 0.050 0.005 CL  +1.037 CF  +1.137
[0.001 0.004 0.002 - 0.001 0.002 0.001 CL   0.007 CF   0.010]
Player rkb owns 2-cube:
 0.796 0.382 0.035 - 0.204 0.045 0.004 CL  +1.907 CF  +1.390
[0.002 0.008 0.003 - 0.002 0.003 0.001 CL   0.011 CF   0.015]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 871181490 and quasi-random dice
Play: 0-ply cubeful prune [expert]
Cube: 0-ply cubeful prune [expert]

[close]

[dorbel]

Cubing is a blunder, but it only costs about a tenth of a point. In the days before bots we couldn't draw lines this accurately and this decision would nave been considered marginal. A conservative player here might reasonably pay that tenth as an insurance and cash his point, with the added bonus that he might even get a wrong take. Note the 35% figure in brackets. That is the number of wrong takes that would make a double correct. Against a weaker player then rkb, you might get that and you will certainly get some, reducing the cost of your insurance still further. There is one other hidden factor that makes doubling reasonable. Turn it now and you can't make a cube error, perhaps even a succession of cube errors later in the game!
Given that the decision to play on for a gammon is so clear, why does doubling only cost 0.137? It's because the most likely result here (40%) is still a plain win for the team, when we cease to be too good and cash later.
In answer to clem's question, no we didn't miss a double. The 3-3/dance sequence was a big joker.

[ah_clem]

In answer to clem's question, no we didn't miss a double. The 3-3/dance sequence was a big joker.

Hmmm...  agree that move 5 (33 / dance) was a very lucky sequence for us, but immediately after it we rolled move 6 (4/3) without stopping to consider a double.  That was a double according to the rollout, so we did miss a double.  Not the end of the world by any means as gnubg would have flagged the decision as "doubtful".

And I would have voted 'no double" had we put it to a vote.