Game 1, move 6: Herd cube decision

[diane]

2GbJATBm2xYCIA:cAmgAAAAAAAA

[ah_clem]

Spoiler

I see this going two ways - we hit one of his outfield blots now (or escape with 6-5) and win a gammon, or we don't and he contains our runner while our board crunches.  If the former, we've lost our market, but there's only a 1/6 chance of that.  5/6th of the time we're headed for trouble and likely will face a recube.

So, despite our formidable 5-prime, we need to sit on the cube.  No double.

Can we apply Woolsey's law here? Most likely this is headed in the direction of the "well timed ace point game", which indicates a take.  So, Woolsey's law doesn't apply.

[close]

[diane]

Ok r_monk...seems we would like to cube you...will you take or pass?

[r_monk]

Take.
Situation is dangerous, but my board is strong; and even if you roll a 6-2, 6-3 or 6-5 I think I can get away without a gammon.

[Zorba]

Spoiler

Looks like take, but not a comfortable one. r_monk's backcheckers are very trapped behind a 5-prime and not at the edge of it.

Even a perfectly timed acepoint game is a drop BTW. Although it gets a shot 90% of the time if you stay right until the end and probably hits a little over 30% of the time, not all hits will be winners since several checkers might be borne off already.

What's worse, an acepoint game where you stay until the end will lose lots of gammons and even backgammons.

Here, r_monk has more than an acepoint game, since he could trap or attack our backchecker. But it's still a difficult gameplan to complete succesfully, since extracting the backcheckers will be hard to do and in the process, any acepoint game equity might get lost.

Big double since we're already very strong here and some good rolls make it a big drop, whereas bad rolls still leave us in a strong position.

[close]

[ah_clem]

 Rollout:

Spoiler

gnu 0-ply thinks cubing is a whopper.  And it's a massive take.



Rollout details:

Cubeful equities:
1. No double            +0.276
2. Double, pass         +1.000  ( +0.724)
3. Double, take         +0.182  ( -0.093)
Proper cube action: No double, take (11.4%)


Centered 1-cube:
  0.555 0.212 0.023 - 0.445 0.075 0.003 CL  +0.278 CF  +0.276
[0.002 0.002 0.001 - 0.002 0.001 0.001 CL   0.007 CF   0.015]
Player r_monk owns 2-cube:
  0.554 0.218 0.033 - 0.446 0.073 0.006 CL  +0.637 CF  +0.182
[0.003 0.003 0.002 - 0.003 0.002 0.001 CL   0.013 CF   0.021]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 896490702 and quasi-random dice
Play: 0-ply cubeful prune [expert]
Cube: 0-ply cubeful prune [expert]

[close]

[diane]

I was waiting for someone to do that..   And wondered even more after Zorbas post...clearly there is something we are all not seeing...

[ah_clem]

...clearly there is something we are all not seeing...

It seems clear to me.  We'll have trouble escaping our runner, he's poised to strengthen his blockade, and it's likely our board will crunch unless we can get it out of there.  Luckily for us he wasn't able to make his bar point - if he had we'd be hurting here.

Our blockade looks formidable, and it is - he's not going anywhere for awhile.  But he's got good timing and as long as we've got a checker stuck on the other side of the board all we can do is watch helplessly as our prime deteriorates while his strengthens. We're still ahead here, but there are many likely scenarios where we'll be eating a recube.

[diane]

I was the other 'NO' vote here...but there were a lot of 'yes', and Zorba describes this a borderline take for monk...not a 'massive' take...

[Zorba]

I didn't mean borderline take, but say something like a 0.9 take. Way off in any case!

Apparently I overrated our own prime/blockade somewhat, and more importantly, seriously underestimated monk's chances to block/attack our backchecker which eventually might cause trouble with our prime upfront, forcing us to crack it and perhaps in a bad way even. After such a scenario, escaping two backcheckers will be that much easier for monk to do.

The loss of our own midpoint, combined with monk's 3 checkers on his midpoint and some other outfield checkers means monk will have outfield control for several turns, making it hard to escape the backman safely, much harder than I'd thought earlier.

So, it's a gigantic take. Even so it's less than a 0.1 error to cube here, probably because an escape and a miss by monk is a big marketloser, so volatility is high.

Practically speaking, this might not be such a bad cube against many players, since I can see a significant percentage of drops here, and then you gain enormously.

[diane]

Practically speaking, this might not be such a bad cube against many players, since I can see a significant percentage of drops here, and then you gain enormously.

Totally agree there...

[ah_clem]

Even a perfectly timed acepoint game is a drop BTW.

Agreed, but a well timed ace point game is almost a take (for money), and in this position there are several other winning scenarios for white, most importantly containing blue's runner.  What I was trying to say was his well timed ace point game plus his other assets implies a take since he's almost there on the ace-point game alone.

[Zorba]

Well it depends on what you mean with "acepoint game". I would think of opponent having filled in all his homeboard points and starting to bear in from the barpoint, or to bear off, against the acepoint game player having a full prime ready to contain a hit. The position of the prime determines how well it's timed.

Using XG to get some quick numbers on this scenario:

A very well timed acepoint game (80 pips behind) may win upto ~23% of games, but then it loses ~40% gammons: a huge drop at -1.8.

A typical well timed acepoint game (trailing 60 pips) wins only 19% of games, and loses 21% gammons for a huge drop at -1.6.

Less well timed acepoint game (trailing 40 pips) wins only 15% and loses 15% gammons too, again for a huge drop at -1.6.

Without the acepoint game player having a full prime in place, things get worse of course. The only way an acepoint game alone can be almost a take, is if opponent has some very ugly gaps in his homeboard and stacked points so the gaps may not fill and the bear-off could be very inflexible.

Normally speaking though, I'd say it's best to think of acepoint games as very huge drops, twopoint games are normally big drops too (unless the race is somewhat close, much easier to run from the 2pt...), threepoint games are almost takes: decent racing chances can makes them takes, or succesful early hitting chances because of opponent's ugliness.

[stiefnu]

Normally speaking though, I'd say it's best to think of acepoint games as very huge drops, twopoint games are normally big drops too (unless the race is somewhat close, much easier to run from the 2pt...), threepoint games are almost takes: decent racing chances can makes them takes, or succesful early hitting chances because of opponent's ugliness.

A good 'rule of thumb', thanks!