BGonline.org Forums

0-0/19 cube decision - RO

Posted By: Daniel Murphy
Date: Monday, 7 February 2011, at 4:57 p.m.

In Response To: 0-0/19 cube decision (Daniel Murphy)

The score is: Daniel_Murphy 0, Asbjoern_Arntzen 0 (match to 19 points)

Move number 8: Daniel_Murphy on roll, cube decision?

Daniel_Murphy	147
Asbjoern_Arntzen	145
Position ID: 4LsjARjQ24MBJA Match ID: MAFgAgAAAAAA

White doubled. Bad move. Probably at least a 0.120 error, I'd guess larger, if this short rollout were extended.

What I recall thinking is first, that White has an obvious advantage. (Thinking now), many numbers make a 4 prime, after which Blue is poorly anchored on a blocked 23 point. Blue's lost his midpoint, aces hit ... except that aces make the 5 point. White is split, has 6s out ... but 6-2 and 6-4 are duplicated. Second, it can be worthwhile to give a somewhat speculative cube early in a long match or session to test one's opponent's mettle.

In retrospect, duplication means White has few multiple threats, and few hit and cover numbers; some 1/3 of White's rolls don't even make the 5 point; if Blue's left alone, at least half his rolls will make the one or two priming points; and most important, this position is so obviously not a pass that if Blue did pass, all I would have learned is that Blue is very timid. Since I didn't expect that to be true -- I had no reason to think my opponent was weaker than I -- there's really no justification for a "testing" double here. No doubt I was riding a 9-game winning streak high. The corn was as high as an elephant's eye, oh what a beautiful day, with everything going my way.

On the other hand, this particular double gave up less than 0.5% match winning chances. There are certainly times to try to push a position and your opponent with less than optimal decisions. This wasn't one of them.

Alert: wrong double ( -0.1208)! [very bad]

Cube decision
Rollout cubeless equity		+0.3044	(Money: +0.3022)
Cubeful equities:
1.	No double	+0.4892
2.	Double, pass	+1.0000	+0.5108
3.	Double, take	+0.3684	-0.1208
Proper cube action:		No double, take (19.1%)
Rollout details
Win W g W bg Lose L g L bg Cubeless Cubeful Centered 1-cube 0.6249 0.1645 0.0068 - 0.3751 0.1115 0.0073 +0.3044 +0.4892 Standard error 0.0037 0.0032 0.0008 - 0.0037 0.0050 0.0034 0.0144 0.0345 Player Asbjoern_Arntzen owns 2-cube 0.6288 0.1679 0.0070 - 0.3712 0.1094 0.0057 +0.6517 +0.3684 Standard error 0.0033 0.0044 0.0011 - 0.0033 0.0041 0.0012 0.0210 0.0358
Full cubeful rollout with var.redn.
339 games, Mersenne Twister dice gen. with seed 879699691 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]

Playing around with the position to give White some improvements reinforces how far White is from a double in the original position.

(1) Move the 4th checker from the 8 to the 7 point, giving all 2s to cover the 5 point in addition to 1s, 3s, 6-2, 4-4 and 2-2. Still far from a double.

(2) Advance both back checkers to the 21 point. Still no double.

(3) Advance a checker from the 22 to the 20 point. Now 6s threaten one of Blue's builders. 6-2 is duplicated, but White has more hit and cover numbers. This is nearly a marginal double.

White	145
Blue	145
Position ID: 4LsjARjQ24MBIQ Match ID: MAFgAgAAAAAA

(4) Give White a made 4 point. Well, that's a big improvement, and White now as a good double with EMG equity approaching 0.900. But it's still an easy take.

White	146
Blue	145
Position ID: 4LsjARiw24MBJA Match ID: MAFgAgAAAAAA

Messages In This Thread

• 0-0/19 cube decision
  Daniel Murphy -- Saturday, 5 February 2011, at 6:39 p.m.
  • 0-0/19 cube decision
    Timothy Chow -- Saturday, 5 February 2011, at 6:52 p.m.
  • 0-0/19 cube decision
    Casper van der Tak -- Saturday, 5 February 2011, at 6:57 p.m.
  • 0-0/19 cube decision
    Golden -- Sunday, 6 February 2011, at 12:37 a.m.
  • 0-0/19 cube decision
    neilkaz -- Sunday, 6 February 2011, at 1:38 a.m.
  • 0-0/19 cube decision - RO
    Daniel Murphy -- Monday, 7 February 2011, at 4:57 p.m.