Game 3, move 9: rkb 3-3 DANCE, move 10: Forum double?

[Zorba]

The Forum played B/21*/15 last turn, roadkillbooks (white) dances with a 3-3, cube decision for the Forum (blue):

jDN4wEIzzmCCIQ:cAmgACAAAAAA

[ah_clem]

Spoiler

We're ahead in the race by 30 pips.  We have a slight lead in the threat department since we're on roll and he has a blot exposed, and we have a slight lead in the posititon category due to our three point board.

Is this enough?  Applying Woolsey's law, I'm pretty sure it's a take.  4-2 and big doubles are market losers, but othere that those, we'll probably have another decision next turn.  I say wait.  IF this is actually a double, I doubk no double is that big of an error.

No double,

[close]

[dorbel]

Spoiler

Again, as earlier in this game, a double is not a big mistake because the cube has little value for White. However, the chances of a wrong pass are just about zero and personally I hate to give the leader a trivially easy take. Hard to see many market losing sequences, so no double.
It's worth looking more closely at the features that lead me to this decision. This is a mutual holding game, where both sides hold a high anchor. The doubling opportunities in this type of game usually come when one side successfully clears their anchor with a large double, or when one side is forced to leave their anchor with one checker thus coming under attack. Blue has a race lead and some attacking chances, but still has three checkers back and worse still, two checkers out of play on the 1pt. This means that Blue must play with 13 checkers, a serious handicap that will restrict choice on every roll from now.

[close]

[lewscannon]

Spoiler

This is a spot where I am prone to making over-aggressive mistakes. I see an opportunity like this and instinctively reach for the cube, smelling a potential gammon, especially now when we're down 2-0 and can turn it around quickly. I have to admit that I read the other comments before voting here, which I normally never do, so that tempered my reaction. So, I'll keep the lion in the cage for another roll and vote 'no', though I think if we were down 3-0, this would be a cube for sure.

[close]

[ah_clem]

Another Next?

[Zorba]

Spoiler

No Double for me either, with rkb's 20pt anchor and our three-and-a-half backmen, there just aren't enough marketlosers: with good rolls we can still double next turn and get a take (or a tiny pass at worst), with bad sequences we will be happy that the cube is still in the center.

[close]

[Zorba]

NEXT!

[Zorba]

GnuBG 2-ply rollout:

Spoiler

It's not a double, but a pretty small error to actually cube here.

If you suspect there's any real chance of your opponent passing, then it's a good cube. But personally I wouldn't guess many opponents pass this.


No double: +0.5383
Double, take: +0.4985 (-0.0398)

Correct cube action: NO DOUBLE (TAKE)

Rollout details:
Centered 1-cube:
   57,49  22,61   0,52 -  42,51  10,51   0,93 CL  +0,3346 CF  +0,5383
[  0,12   0,13   0,03 -   0,12   0,20   0,13 CL   0,0037 CF   0,0074]
Player roadkillbooks owns 2-cube:
   57,59  22,97   0,29 -  42,41  11,28   1,87 CL  +0,6591 CF  +0,4985
[  0,15   0,17   0,03 -   0,15   0,25   0,20 CL   0,0083 CF   0,0096]

Full cubeful rollout with var.redn.
2592 games, Mersenne Twister dice gen. with seed 860861592 and quasi-random dice
Play:  2-ply cubeful prune [world class]
keep the first 0 0-ply moves and up to 8 more moves within equity 0,13
Skip pruning for 1-ply moves.
Cube: 2-ply cubeful prune [world class]

[close]