Game 4, move 11: Herd 3-1

diane · December 20, 2011, 08:01:29 PM

Tough one indeed...

ah_clem · December 20, 2011, 09:17:37 PM

This is a hard one. I'm glad I don't have to make this decision. It would be interesting to hear what various members of the Herd think about this play. I'll post something once the voting is over.

stog · December 21, 2011, 09:21:44 AM

it may be worth extending the time to vote for this move for some more comments; 3 different votes already

Spoiler

i voted to move up with the 3 because we need to use that 3 and hit with the 1 ; giving him 2 numbers to hit, but we could just hope for a hit later and go with the other plays; trying to maximise our response

[close]

stiefnu · December 21, 2011, 10:07:24 AM

Spoiler

A hard one indeed! Play the passive 10-6 or go for broke by spreading 'em out for recycling, with the hope of a back game? In which case we'd presumably have to hit with 3-2* and split at the back too. Gulp! Think I'm going to go passive and be thankful I'm not playing this in the final of a tourney, with the clock ticking down and big money at stake!

[close]

diane · December 21, 2011, 06:16:13 PM

Ok - voting period extended, I hope we can sort this one to a majority..it is a call I dont want to have to make alone

Moving that back 3 is sooo tempting, but it just looks soooo risky too!

Krazula · December 22, 2011, 03:04:27 AM

Spoiler

I like 24/21 3/2*. 24/21 puts a checker up to where we can escape or advance our anchor. ah_clem dances nearly half the time and when that happens we have a lot of good rolls that make a 5 point board, advance anchor, or escape with a checker. If we get hit it is bad, but even if that happens we still may enter, hopefully anchoring on 4 point at the same time. Because of the score and the 2cube we don't pay any extra for losing a gammon so while I think this play would be too risky at some other scores I like it here.

[close]

diane · December 22, 2011, 09:42:03 AM

oooooo..that is radical..but I like it...since this is a learning type match - this would be exactly the place to try such a strategy and see how it works out! Be brave, Herd!

ah_clem · December 22, 2011, 07:27:45 PM

Ok, the voting is closed and the vote is tied. Shall I cast the tiebreaker? (c:

diane · December 22, 2011, 07:48:10 PM

I have re-opened this voting, lets have more input - this is too tough for me to make the deciding call alone

diane · December 23, 2011, 08:19:51 PM

Ok - since it is still tied..I am going with the argument from the highest rated player..lets see what happens..this is all about learning, isn't it, and not about winning

diane · December 23, 2011, 08:22:55 PM

Although..that highest rated player did comment and argue without considering who is rolling the dice here...I am sooo sorry Herd..next time - make a case and don't leave it tied!!

ah_clem · December 23, 2011, 08:49:17 PM

I put thoughts in writing before looking at any of the Herd's opinions or bot results. Here's my take on this decision:

Quote
If the Herd stays on the ace they'll still be in trouble even if they hits the
runner. So that makes me want to get one of the runners moving. Plus,
I've never been a fan of sitting on the ace point and patiently waiting
to lose. Gammon losses are meaningless to the Herd, so bold play is more
justified. Plus, being behind in the race and behind in the match steers
me towards bold.

The problem is, moving up to the 21 gives clem lots of rolls that point on
a blot or pick and pass, and the rolls that don't allow clem to escape the
runner. To solve this little conundrum, make a tempo hit to take half
of clem's rolls away: 24/21 3/2*

Yes, this leaves 20 rolls that enter and hit, but the other 16 give X
quite a leg-up. That seems to be better odds than the ~20% chance that
the ace-point game will bear fruit. And if hit, X is 20/36 to enter &
anchor, so all is not lost if clem hits.

No, I'm not at all sure about this, but it's the best I can do.

So in my opinion the Herd chose the best move here. Of course, I didn't expect a devil-shake of 21, even with Diane rolling. (c: Actually, the double hit might not be all that bad - gammons don't count, and this gives the Herd great timing if they can manage to make an anchor or two.

Rollout in a sec...

ah_clem · December 23, 2011, 09:00:31 PM

While the Herd was agonizing over the decision I punched the position into gnu. Here's the rollout:

Spoiler

There are two very different ideas of how to play this. There's the bold play that the Herd eventually chose, and then several ways to creep up on the runner and try to get a shot as it comes around the board. Surprisingly, they're actually close in equity. The best move is to slot the bar point - this puts direct pressure on the runner in addition to threatening to make a five-prime. Nobody saw this move as desirable - I didn't even consider it.

Second best and not at all far behind is the Herd's play. The Herd's alternate play is third best, and pretty close to the leader. Nobody should feel bad about their choice - it was a tough call, even though it ultimately didn't matter much equity-wise which way it went.

1. Rollout 13/12 10/7 Eq.: -0.217
0.346 0.046 0.004 - 0.654 0.284 0.080 CL -0.217 CF -0.217
[0.002 0.001 0.001 - 0.002 0.007 0.005 CL 0.004 CF 0.004]

2. Rollout 24/21 3/2* Eq.: -0.234 ( -0.018)
0.332 0.051 0.003 - 0.668 0.466 0.097 CL -0.234 CF -0.234
[0.002 0.001 0.001 - 0.002 0.008 0.006 CL 0.004 CF 0.004]

3. Rollout 10/6 Eq.: -0.242 ( -0.026)
0.333 0.046 0.002 - 0.667 0.247 0.044 CL -0.242 CF -0.242
[0.002 0.001 0.000 - 0.002 0.008 0.004 CL 0.004 CF 0.004]

4. Rollout 13/12 13/10 Eq.: -0.258 ( -0.041)
0.331 0.040 0.004 - 0.669 0.258 0.047 CL -0.258 CF -0.258
[0.002 0.001 0.001 - 0.002 0.008 0.004 CL 0.004 CF 0.004]

5. Rollout 13/9 Eq.: -0.293 ( -0.076)
0.317 0.037 0.002 - 0.683 0.273 0.072 CL -0.293 CF -0.293
[0.002 0.001 0.000 - 0.002 0.007 0.005 CL 0.004 CF 0.004]

6. Rollout 24/21 10/9 Eq.: -0.318 ( -0.101)
0.298 0.043 0.003 - 0.702 0.409 0.036 CL -0.318 CF -0.318
[0.002 0.001 0.001 - 0.002 0.009 0.003 CL 0.004 CF 0.004]

7. Rollout 24/21 13/12 Eq.: -0.332 ( -0.116)
0.296 0.038 0.004 - 0.704 0.468 0.067 CL -0.332 CF -0.332
[0.002 0.001 0.001 - 0.002 0.008 0.005 CL 0.004 CF 0.004]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 935509925 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]

[close]

FIBS Board backgammon forum

News:

Game 4, move 11: Herd 3-1

diane

ah_clem

stog

stiefnu

diane

Krazula

diane

ah_clem

diane

diane

diane

ah_clem

ah_clem