Fibsboard Forum Match 18 Game 4 Move 26 Herd to move

[stiefnu]

juanz board starts to crunch and now the Herd gets 54 to play. Put one on the roof or bear some in? The choice is yours.

5-pt match, juanz 0-4 Herd, Crawford
Pipcount: juanz (red) 105, Herd (black) 94

Steve

[ah_clem]

Spoiler

I don't think putting him on the roof makes sense here.  If gammons counted, sure, but we're just trying to win a single point.  10/1* means he could enter and we might be forced to send a fourth checker back - at that point we might be facing a 21 backgame instead of a deuce point game.

I think it's stronger to just let him crunch, and by the time we are forced to leave a shot (if we are)  he'll have no board.

That said, I doubt hitting loose on the ace is much of an error if at all.  If he enters with anything other than 21 he'll have to play from his six point (13 allows him to expose s second blot, but at the expense of tearing up the board).

I like 10/6 9/4, but 9/5 9/4 is also playable. 10/5 9/5 seems like a small error due to the stack on the 5.

10/6 9/4 is my choice, but now that I've seen the partial vote results I'll switch to 9/5 9/4 for strategic reasons.

[close]

[dorbel]

Spoiler

Whack him. Gammons aren't worth much, but they are worth something and backgammons are worth a lot! Nothing to lose, so put him on the roof!

[close]

[ah_clem]

Spoiler

Whack him. Gammons aren't worth much, but they are worth something and backgammons are worth a lot! Nothing to lose, so put him on the roof!

How is this possible when we are the leaders in the Crawford game?

Your comment makes sense if we were trailing 4a Crawford, but we are leading 4a Crawford.  Or am I missing something?

[close]

[dorbel]

Spoiler

How right you are, not the first time I've got the score arse about face. However, I think it's still the best way to win. All juanz's wins come when he gets a shot and hits it. Giving him that shot now is harmless, in fact if he hits it it does us a lot of good as he has to keep playing and crunches even more. If he enters on the 2 we are no worse off, if he dances we gain because we can now bear in while he is on the bar, always easier.

[close]

[NIHILIST]

Spoiler

My thinking is the reverse. I want him stranded on the 1 point. My strategy is to bring builders to bear and then, hoping he doesn't roll a 1 to move up to the 2 point, open up our prime and squeeze him off the 2 point when he rolls a 6.

By that time, if all goes according to plan, we'll have plenty of wood to bear on any blots in our board, he'll be even more crashed and much less of a threat to us.

9-5, 9-4 would be my play.

[close]

Bob

[stiefnu]

Spoiler

ah_clem's persuasive argument prevailed, though a BGBlitz rollout confirms his feelings that the hit on the ace point was not much of an error. In fact, all but one of the candidate moves produce match winning probabilities in excess of 97%, with only 7-3, 10-5 languishing at a mere 95.9%. Best of the bunch were 10-6, 9-4; 9-5, 10-5 and 9-5, 9-4 -- all at 98%.

[close]

Steve

[ah_clem]

rollout

Spoiler

XG likes having the spares on the 6 & 4 slightly more than the 5 & 4, but the herd's choice is only a .007 error.  Something James Bond might play.

Hitting loose is quite close at .011.

Stacking two spares on the 5 is next at .017

The trap play (8/4 8/3) is surprisingly playable, and only a .023 error.  Nobody mentioned it, but it should be on your radar in these sorts of positions.


XGID=-abBBBBBBBA--------a-bebb-:0:0:1:54:4:0:1:5:10

X:Player 1   O:Player 2
Score is X:4 O:0 5 pt.(s) match.
+13-14-15-16-17-18------19-20-21-22-23-24-+
|                  |   | O     O  O  O  O |
|                  |   |       O  O  O  O |
|                  |   |          O       |
|                  |   |          O       |
|                  |   |          O       |
|                  |BAR|                  |
|                  |   |                  |
|                  |   |                  |
|                  |   |                  |
|          X  X  X |   | X  X  X  X  O    |
|       X  X  X  X |   | X  X  X  X  O  O |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count  X: 94  O: 105 X-O: 4-0/5 Crawford
Cube: 1
X to play 54

   1. Rollout¹    10/6 9/4                     eq:+0.895
     Player:   94.75% (G:24.31% B:1.50%)
     Opponent: 5.25% (G:0.06% B:0.00%)
     Confidence: ±0.002 (+0.893..+0.897) - [100.0%]
     Duration: 3 minutes 09 seconds

   2. Rollout¹    9/5 9/4                      eq:+0.888 (-0.007)
     Player:   94.39% (G:22.62% B:1.46%)
     Opponent: 5.61% (G:0.08% B:0.00%)
     Confidence: ±0.002 (+0.886..+0.890) - [0.0%]
     Duration: 2 minutes 21 seconds

   3. Rollout¹    10/5 9/5                     eq:+0.886 (-0.009)
     Player:   94.32% (G:22.30% B:1.50%)
     Opponent: 5.68% (G:0.10% B:0.00%)
     Confidence: ±0.002 (+0.885..+0.888) - [0.0%]
     Duration: 1 minute 49 seconds

   4. Rollout¹    10/1*                        eq:+0.884 (-0.011)
     Player:   94.20% (G:23.04% B:1.53%)
     Opponent: 5.80% (G:0.05% B:0.00%)
     Confidence: ±0.002 (+0.882..+0.886) - [0.0%]
     Duration: 3 minutes 37 seconds

   5. Rollout¹    8/4 8/3                      eq:+0.872 (-0.023)
     Player:   93.60% (G:22.92% B:1.27%)
     Opponent: 6.40% (G:0.13% B:0.00%)
     Confidence: ±0.002 (+0.870..+0.874) - [0.0%]
     Duration: 2 minutes 46 seconds


¹  1296 Games rolled with Variance Reduction.
  Moves: 3-ply, cube decisions: XG Roller


eXtreme Gammon Version: 2.10, MET: Kazaross XG2

[close]

[ah_clem]

rollout variant

Spoiler

Is the match score relevant?  You betcha.  At 5a Crawford gammons don't really matter for either side. But what if they did?

If we were at a score where gammons counted I would have hit loose.  And this would have been incorrect.

For money play without Jacoby (this is a good proxy for "normal' match scores) the trap play of 8/4 8/3  comes out on top.  With the 8 point open, it means O is forced to break anchor with any six other than 61.  This gives X a good chance to close out two checkers.  Trap plays are easy to overlook.  Tim is doing a nice series on trap plays over at rgb this week.




XGID=-abBBBBBBBA--------a-bebb-:0:0:1:54:0:0:0:0:10

X:Player 1   O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
|                  |   | O     O  O  O  O |
|                  |   |       O  O  O  O |
|                  |   |          O       |
|                  |   |          O       |
|                  |   |          O       |
|                  |BAR|                  |
|                  |   |                  |
|                  |   |                  |
|                  |   |                  |
|          X  X  X |   | X  X  X  X  O    |
|       X  X  X  X |   | X  X  X  X  O  O |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count  X: 94  O: 105 X-O: 0-0
Cube: 1
X to play 54

    1. Rollout¹    8/4 8/3                      eq:+1.217
      Player:   92.51% (G:26.52% B:1.27%)
      Opponent: 7.49% (G:0.38% B:0.01%)
      Confidence: ±0.003 (+1.214..+1.220) - [100.0%]
      Duration: 16 minutes 23 seconds

    2. Rollout¹    10/6 9/4                     eq:+1.205 (-0.012)
      Player:   93.94% (G:24.35% B:0.78%)
      Opponent: 6.06% (G:0.15% B:0.00%)
      Confidence: ±0.003 (+1.202..+1.207) - [0.0%]
      Duration: 16 minutes 50 seconds

    3. Rollout¹    10/5 8/4                     eq:+1.200 (-0.018)
      Player:   91.61% (G:26.05% B:1.05%)
      Opponent: 8.39% (G:0.40% B:0.00%)
      Confidence: ±0.003 (+1.197..+1.202) - [0.0%]
      Duration: 19 minutes 04 seconds

    4. Rollout¹    9/5 9/4                      eq:+1.195 (-0.022)
      Player:   93.65% (G:23.80% B:0.78%)
      Opponent: 6.35% (G:0.17% B:0.00%)
      Confidence: ±0.002 (+1.192..+1.197) - [0.0%]
      Duration: 14 minutes 00 second

    5. Rollout¹    10/1*                        eq:+1.183 (-0.034)
      Player:   93.26% (G:23.29% B:0.74%)
      Opponent: 6.74% (G:0.13% B:0.00%)
      Confidence: ±0.003 (+1.181..+1.186) - [0.0%]
      Duration: 26 minutes 14 seconds


¹  2592 Games rolled with Variance Reduction.
   Moves: 3-ply, cube decisions: XG Roller


eXtreme Gammon Version: 2.10

[close]