BGonline.org Forums

41S-11 question

Posted By: David Rockwell
Date: Saturday, 30 January 2010, at 3:34 p.m.

I did a rollout to try to understand the value lost in 41S-11F (Float 24-23, 6-5(3)) due to the floating checker versus the value lost by not splitting all of the way to the 22 point with e (each 24-22, 6-5(2)), the best play. I did this by playing only three aces instead of the required four. I don't know how to Nactate this illegal play, if indeed there is a way. I will use ? for this one time usage. I have:

41S-11 [e ?16 F33 E61] - 5k

I was quite surprised to see that the float was as costly as the failure to split all the way. I expected the float to be relatively insigificant. Of course, if I was sure about that, I wouldn't have done the rollout. I am wondering if I need to extend these rollouts or not. Does anyone have any knowledge of this position and whether or not the rollouts look reasonable?

As a side note, I have done several rollouts of the (legal) position using MWC settings. The absolulte equity of these vary by a surprising amount from the Supremo rollouts. The differences between plays are somewhat consistent between the two settings.

The score (after 0 games) is: GNU_bg 0, user 0

Move number 3: user to play 11

GNU_bg	162
user	168
Position ID: 4HPhASjgc/ABUA Match ID: cIkEAAAAAAAA

• user moves bar/23 6/5(2)

#		Ply	Move	Equity
•	1	R	bar/23 6/5(2)	+0.1385
			0.5348 0.1462 0.0064 - 0.4652 0.1317 0.0054 +0.0851 +0.1385 0.0008 0.0009 0.0002 - 0.0008 0.0008 0.0002 0.0020 0.0057
			Full cubeful rollout with var.redn.
			5184 games, Mersenne Twister dice gen. with seed 857636419 and quasi-random dice
			Play: supremo 2-ply cubeful prune [world class]
			keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
			Skip pruning for 1-ply moves.
			Cube: 2-ply cubeful prune [world class]
	2	2	bar/24 6/5(3)	+0.0751 ( -0.0635)
			0.5249 0.1442 0.0051 - 0.4751 0.1408 0.0054
	3	2	bar/24 8/7 6/5(2)	+0.0133 ( -0.1252)
			0.5063 0.1450 0.0059 - 0.4937 0.1481 0.0081
	4	0	bar/21	-0.1486 ( -0.2871)
			0.4638 0.1114 0.0039 - 0.5362 0.1588 0.0042
	5	0	bar/22 24/23	-0.1636 ( -0.3021)
			0.4600 0.1102 0.0038 - 0.5400 0.1618 0.0043

The score (after 0 games) is: GNU_bg 0, user 0

Move number 2: user to play 11

GNU_bg	162
user	167
Position ID: 4HPhASjgc/ABMA Match ID: cIkEAAAAAAAA

• user moves 24/22 6/5(2)

#		Ply	Move	Equity
•	1	R	24/22 6/5(2)	+0.1546
			0.5403 0.1408 0.0059 - 0.4597 0.1288 0.0050 +0.0936 +0.1546 0.0008 0.0009 0.0002 - 0.0008 0.0008 0.0002 0.0020 0.0053
			Full cubeful rollout with var.redn.
			5184 games, Mersenne Twister dice gen. with seed 857636419 and quasi-random dice
			Play: supremo 2-ply cubeful prune [world class]
			keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
			Skip pruning for 1-ply moves.
			Cube: 2-ply cubeful prune [world class]
	2	R	24/23 6/5(3)	+0.1221 ( -0.0326)
			0.5307 0.1473 0.0066 - 0.4693 0.1372 0.0056 +0.0725 +0.1221 0.0008 0.0009 0.0002 - 0.0008 0.0008 0.0002 0.0020 0.0055
			Full cubeful rollout with var.redn.
			5184 games, Mersenne Twister dice gen. with seed 857636419 and quasi-random dice
			Play: supremo 2-ply cubeful prune [world class]
			keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
			Skip pruning for 1-ply moves.
			Cube: 2-ply cubeful prune [world class]
	3	R	24/23(2) 6/5(2)	+0.0938 ( -0.0608)
			0.5227 0.1456 0.0065 - 0.4773 0.1296 0.0056 +0.0622 +0.0938 0.0008 0.0009 0.0002 - 0.0008 0.0008 0.0002 0.0021 0.0057
			Full cubeful rollout with var.redn.
			5184 games, Mersenne Twister dice gen. with seed 857636419 and quasi-random dice
			Play: supremo 2-ply cubeful prune [world class]
			keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
			Skip pruning for 1-ply moves.
			Cube: 2-ply cubeful prune [world class]
	4	2	8/7(2) 6/5(2)	+0.0689 ( -0.0858)
			0.5207 0.1489 0.0058 - 0.4793 0.1387 0.0063
	5	2	8/5 6/5	+0.0677 ( -0.0869)
			0.5202 0.1525 0.0059 - 0.4798 0.1423 0.0059

Messages In This Thread

• 41S-11 question
  David Rockwell -- Saturday, 30 January 2010, at 3:34 p.m.
  • 41S-11 question
    Paul Weaver -- Saturday, 30 January 2010, at 6:09 p.m.
  • 41S-11 question
    Nack Ballard -- Saturday, 30 January 2010, at 9:36 p.m.
    • 41S-11 question
      David Rockwell -- Saturday, 30 January 2010, at 10:21 p.m.