Game 5, move 5 : Forum 3-3

[diane]

Now we get a 33    Forum, blue, to move 3-3

[I think this is pretty obvious too, so if it is all decided by the time I get home, I will move it along]


zLaDASSYZ/AxAA:cImtACAAGAAA

[ah_clem]

Spoiler

Agree that 8/5(2), 6/3*(2) is an obvious choice. 

But, I'm not completely certain that it's the best play.  Sometimes the "obvious" choice is not... especially when playing doublets.

[close]

[diane]

Spoiler

That was my choice too...but I did want to see if anyone wanted to consider the double hit in particular - what with the score and all.  Someone voted for it - but they have done it anonymously, and not said why 

[close]

[ah_clem]

Spoiler

That was my choice too...but I did want to see if anyone wanted to consider the double hit in particular - what with the score and all.  Someone voted for it - but they have done it anonymously, and not said why 

Well, any time you can put two checkers on the roof you should take a serious look at it.  In this case, though, owning 4 points which also happen to be the best four points seems to outweigh the double hit.  If there were more builders nearby I'd think harder about the double hit followed by a blitz, but the reserves are just too far away...

BTW, my first instinct was to clear the 16 point, but that was before I saw the other moves.

[close]

[diane]

ok, this seems like a consensus, I will move it along before I go to bed...

[ah_clem]

rollout

Spoiler

First the cube, as expected a huge blunder to double:



Cube analysis
Rollout cubeless equity  +0.012 (Money:  +0.046)

Cubeful equities:
1. No double            -0.063
2. Double, pass         +1.000  ( +1.063)
3. Double, take         -0.776  ( -0.714)
Proper cube action: No double, take (40.2%)
Rollout details:
Centered 1-cube:
  0.526 0.121 0.005 - 0.474 0.125 0.006 CL  +0.012 CF  -0.063
[0.002 0.004 0.001 - 0.002 0.002 0.001 CL   0.005 CF   0.011]
Player rkb owns 2-cube:
  0.528 0.146 0.010 - 0.472 0.157 0.015 CL  +0.423 CF  -0.776
[0.002 0.005 0.002 - 0.002 0.005 0.002 CL   0.011 CF   0.016]
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]



And now the checker play.  We made the right call by a wide margin.


    1. Rollout          8/5(2) 6/3*(2)               Eq.:  +0.803
       0.714 0.289 0.026 - 0.286 0.077 0.008 CL  +0.487 CF  +0.803
      [0.002 0.007 0.003 - 0.002 0.003 0.002 CL   0.006 CF   0.016]
       
    2. Rollout          13/10(2) 6/3*(2)             Eq.:  +0.439 ( -0.364)
       0.643 0.207 0.016 - 0.357 0.093 0.007 CL  +0.358 CF  +0.439
      [0.002 0.005 0.002 - 0.002 0.002 0.001 CL   0.006 CF   0.012]
       
    3. Rollout          13/7(2)                      Eq.:  +0.326 ( -0.478)
       0.611 0.140 0.008 - 0.389 0.079 0.003 CL  +0.263 CF  +0.326
      [0.002 0.004 0.002 - 0.002 0.002 0.001 CL   0.006 CF   0.012]
       
    4. Rollout          13/10(2) 8/5(2)              Eq.:  +0.261 ( -0.543)
       0.594 0.139 0.007 - 0.406 0.080 0.004 CL  +0.226 CF  +0.261
      [0.002 0.004 0.001 - 0.002 0.002 0.002 CL   0.006 CF   0.012]
        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]

[stiefnu]

and quasi-random dice

Why only quasi-random, I wonder, and what does it actually mean?

[spielberg]

This is well off the thread but I'll respond here as it's where the question has been asked. Quasi-random just means that they're not truly random dice. The dice generator at FIBS  was state of the art when marvin wrote it. Which ever algorithm he chose it's far better than Von Neumann's  and that was easily good enough to generate dice - it's just slow in processing time. marvin used a far more efficient algorithm for the fibs dice. If you ask Patti she might know which one he picked.