Game 3, move 7 roadkillbooks 4-1 NEXT, move 8 Forum 6-5

[Zorba]

The Forum chose B/24 9/7* last turn, roadkillbooks rolls a 4-1 from the bar and can't hit, which leaves him the obvious B/20 making the golden anchor.

Now the Forum rolls 6-5:

jGfwgCGM12DAIg:cIm6ACAAAAAA

[ah_clem]

Spoiler

I'm even more lost in this position than the last one.   I have no idea what the most important thing is here.  I'll try anyway...

The race is about even, although we have more men back. But this isn't a racing position so  I don't think a running play like lover's leap is the right idea. I think we need to keep exchanging hits and hope for a bad roll from rkb.  Anyway, none of the non-hiting plays do much for us position-wise. So, where to hit?

7/1* 6/1 points on his head and gives us a three point board.  But it kills two checkers in a position where there's still lots of contact and will be for some time.

7/1* 21/16  hits and activates a runner.  Not bad.

6/1* 13/7 hits and makes the bar point giving us a three point blockade.  But it abandons the midpoint, a valuable asset.  And 8-7-6 is not a very good blockade against an anchor on the 5.

7/1* 21/16 mostly because it seems least likely to undermine our position. 

[close]

[mrs.stog]

Spoiler

i like 7/1 6/1 we still have a mid point, a healthy reserve garrisoned in his board and you neverr know he may fan, we ro;ll a double and ...
if not - we have removed a possible  irritant in our home board. i like mr clem's choice too though

[close]

[diane]

Spoiler

I think 13-7 is a must to make that point and continue the game we have started, which seems to be a blocking game, holding RKB down in our homeboard.  Then I have a 5 to play, and I look around at what there is  - I can bring the other pip down from the 13, but that seems a bit risky , and loses us the advantage of that pip being there.  21-16 just gives RKB another target, with no hindrance to his next move at all.  The hit then on the 1, I like this, because it removs half of RKBs next roll and therefore stops him doing anything really dangerous. The worst case I can see from this position is that he hits us coming in, and I am happy for that, it will keep him behind the blockade we are building, and give us more ammo to come back at him with.

13-7, 6-1 for me.

[close]

[blitzxz]

Spoiler

The most important thing in nackgammonish games is (not in any particular order) a) getting back checkers moving b) race and c) blocking opponents back checkers.

In here these are conflicting objects. If we make the bar (trapping opponent) we will lose midpoint which is extremely important. We need that landing point because otherwice we are in major risk of getting trapped with back men and going into crashed position. The other choice is running but that way we leave bar point uncovered and risk losing more in race. So it looks like that we have to do very atypical move for this situation and burn checkers to 1-point with 7/1* 6/1. Hopefully rkb dances and we get some important race lead and time to organize our checkers.

To the future moves our general guide line should be: Race 'n' block (with exceptions like this move).

[close]

[lewscannon]

Spoiler

I would prefer to hit and make the 1 point here, though the other arguments make some sense. Good things can happen as a result of the hit and point, whether rkb dances or is forced into a bad situation as a result. One of the basic principles of bg is to give yourself as many chances to get lucky as possible, and this looks like an opportunity. I know one normally wouldn't want to make the 1 point this early, but I don't like abandoning the outfield pieces, and the lover's leap doesn't do much for me here.

[close]

[dorbel]

Spoiler

I would mark this as a very tough play. Making the bar is no great gain, particularly as it means giving up the midpoint in the middle of a positional struggle. At first I selected 24/13, but it doesn't really exert any pressure as all it does is threaten to make the not-very-useful bar point! I think we have to hit on the ace and when you do that, it's almost always correct to make it if you can. Some people mutter here that "being hit will improve our timing". Being hit loses us a minimum of 24 pips in the race is a bit more like it. 7/1*, 6/1 is my play, so good that lews names it twice.

[close]

[diane]

Oh well, I learned something here then    That is what we are here for!

[ah_clem]

Spoiler

At this point the vote is 9 to 1 to 1.  Seems like a good candidate for a "next", even though it's not my choice. 

For the record, though, allow me to state that I'm sure my move is correct at the zero-percent-confidence level.

[close]

[diane]

For the record, though, allow me to state that I'm sure my move is correct at the zero-percent-confidence level.

[Zorba]

NEXTed 

[ah_clem]

gnubg results:

Spoiler

Big error on my part.  Correct decision by the forum.




Analysis:

    1. Cubeful 2-ply    7/1* 6/1                     MWC:  35.88%
       0.519 0.148 0.003 - 0.481 0.110 0.002
        2-ply cubeful prune [world class]
    2. Cubeful 2-ply    13/8 13/7                    MWC:  35.46% ( -0.43%)
       0.512 0.124 0.003 - 0.488 0.100 0.002
        2-ply cubeful prune [world class]
    3. Cubeful 2-ply    21/16 13/7                   MWC:  34.78% ( -1.10%)
       0.494 0.124 0.004 - 0.506 0.134 0.004
        2-ply cubeful prune [world class]
    4. Cubeful 2-ply    24/13                        MWC:  34.61% ( -1.27%)
       0.491 0.107 0.003 - 0.509 0.114 0.003
        2-ply cubeful prune [world class]
    5. Cubeful 2-ply    13/7 6/1*                    MWC:  34.24% ( -1.64%)
       0.474 0.121 0.004 - 0.526 0.133 0.004
        2-ply cubeful prune [world class]
    6. Cubeful 2-ply    21/16 7/1*                   MWC:  34.21% ( -1.67%)
       0.477 0.115 0.003 - 0.523 0.136 0.004
        2-ply cubeful prune [world class]
    7. Cubeful 2-ply    8/2 7/2                      MWC:  33.87% ( -2.01%)
       0.457 0.134 0.004 - 0.543 0.147 0.005
        2-ply cubeful prune [world class]
    8. Cubeful 2-ply    21/10                        MWC:  33.77% ( -2.12%)
       0.462 0.119 0.004 - 0.538 0.155 0.006
        2-ply cubeful prune [world class]
    9. Cubeful 2-ply    21/15 7/2                    MWC:  33.75% ( -2.14%)
       0.463 0.112 0.003 - 0.537 0.149 0.005
        2-ply cubeful prune [world class]
   10. Cubeful 2-ply    13/8 7/1*                    MWC:  33.44% ( -2.44%)
       0.449 0.116 0.003 - 0.551 0.145 0.005
        2-ply cubeful prune [world class]
   11. Cubeful 2-ply    21/15 6/1*                   MWC:  33.44% ( -2.45%)
       0.456 0.113 0.003 - 0.544 0.165 0.008
        2-ply cubeful prune [world class]

Rollout

    1. Rollout          7/1* 6/1                     MWC:  35.10%
       0.495 0.136 0.003 - 0.505 0.120 0.003 CL  34.58% CF  35.10%
      [0.001 0.001 0.000 - 0.001 0.001 0.000 CL   0.03% CF   0.04%]
        Truncated cubeful rollout (depth 11) with var.redn.
        826 games, Mersenne Twister dice gen. with seed 860737493 and quasi-random dice
        Play: 0-ply cubeful prune [expert]
        Cube: 0-ply cubeful prune [expert]
    2. Rollout          13/8 13/7                    MWC:  34.37% ( -0.73%)
       0.479 0.107 0.003 - 0.521 0.108 0.002 CL  34.14% CF  34.37%
      [0.001 0.001 0.000 - 0.001 0.001 0.000 CL   0.03% CF   0.03%]
        Truncated cubeful rollout (depth 11) with var.redn.
        826 games, Mersenne Twister dice gen. with seed 860737493 and quasi-random dice
        Play: 0-ply cubeful prune [expert]
        Cube: 0-ply cubeful prune [expert]
    3. Rollout          21/16 13/7                   MWC:  33.85% ( -1.24%)
       0.464 0.115 0.004 - 0.536 0.141 0.005 CL  33.63% CF  33.85%
      [0.001 0.001 0.000 - 0.001 0.001 0.000 CL   0.03% CF   0.04%]
        Truncated cubeful rollout (depth 11) with var.redn.
        826 games, Mersenne Twister dice gen. with seed 860737493 and quasi-random dice
        Play: 0-ply cubeful prune [expert]
        Cube: 0-ply cubeful prune [expert]
    4. Rollout          24/13                        MWC:  33.84% ( -1.26%)
       0.466 0.098 0.003 - 0.534 0.120 0.003 CL  33.66% CF  33.84%
      [0.001 0.001 0.000 - 0.001 0.001 0.000 CL   0.02% CF   0.03%]
        Truncated cubeful rollout (depth 11) with var.redn.
        825 games, Mersenne Twister dice gen. with seed 860737493 and quasi-random dice
        Play: 0-ply cubeful prune [expert]
        Cube: 0-ply cubeful prune [expert]

[close]

[diane]

OOh, big one for me too - eek, thanks for the steer folks 

[stog]

the forum is wise - it knows best- it is a collective consciousness of chocolate coffee and bg considered mindsssssssssssssssssssss

[Zorba]

Spoiler

7/1* 6/1 looks like the only move that actually gains us something; all other moves make our position worse than it was.

24/13 gives opponent sixes and twos to hit and set us back 18 pips for not so much gain. The double-direct shot for rkb means we're a big favourite to get hit after this play, so we're not really ahead in the race overall, making 24/13 kind of inconsistent.

All loose hits at the acepoint are terrible, the tactical gain is almost non-existent with a two-point board, and the returnhits are costly in pips, but also because they deplete the offense and make it 5 men back vs. 3. As others put it, when you hit loose on the acepoint, it's nearly always correct to make it too if that can be done somewhat conveniently. Here, it is very convenient, not leaving any blots after 7/1* 6/1 or any other bad things. The third homeboard point also means rkb's dancing numbers go from a measly 11% (4 rolls) to 25% (9 rolls) and also, that some of his entering rolls don't play that well (try 2-6 f.i.).

The most serious alternative to pointing on the acepoint is 13/7 13/8, also making a point. A clearly better point actually, but it doesn't hit, and its value is somewhat limited by rkb's 20pt anchor. But it removes the blot and creates an asset. The big downside is the loss of the midpoint. With 4 backmen, we need the midpoint to keep our army connected. Those four backcheckers have a lot of enemy territory to travel if we abandon the midpoint this soon. This is even more true considering rkb has spares on his midpoint, and is therefore unlikely to abandon it himself anytime soon.

Anyhow, while the acepoint itself is anti-thematical in this positional struggle, it has less deficiencies than the other plays. It leaves quite a good structure too. The threepoint board is valuable for future shots and putting a blot on the bar can reap benefits too.

Uhm oh, and did I mention the matchscore yet? Pointing on the acepoint must be clearly the most gammonish play here: an extra homeboard point and a man on the bar.

[close]

[Zorba]

0/2ply rollout:

Spoiler

It's not close... 7/1* 6/1 wins most games to start with, and really gains in the gammon wins department. At this score, that puts it way ahead of the other plays.


    1. Rollout          7/1* 6/1                     Eq.:  +0,1238
        50,79  13,03   0,27 -  49,21  12,19   1,03 CL  +0,0593 CF  +0,1238
      [  0,13   0,11   0,03 -   0,13   0,15   0,11 CL   0,0038 CF   0,0079]
        Full cubeful rollout with var.redn.
        2592 games, Mersenne Twister dice gen. with seed 860448159 and quasi-random dice
        Play: 0-ply cubeful prune [expert]
        Cube: 2-ply cubeful prune [world class]
    2. Rollout          13/8 13/7                    Eq.:  +0,0303 ( -0,0936)
        49,51  10,36   0,30 -  50,49  10,92   0,73 CL  +0,0091 CF  +0,0303
      [  0,13   0,11   0,02 -   0,13   0,15   0,10 CL   0,0037 CF   0,0077]
    3. Rollout          21/16 13/7                   Eq.:  -0,0030 ( -0,1268)
        48,21  10,65   0,45 -  51,79  13,84   1,09 CL  -0,0323 CF  -0,0030
      [  0,14   0,11   0,05 -   0,14   0,17   0,10 CL   0,0042 CF   0,0084]
    4. Rollout          24/13                        Eq.:  -0,0239 ( -0,1477)
        48,26   9,52   0,30 -  51,74  11,97   0,83 CL  -0,0323 CF  -0,0239
      [  0,13   0,11   0,03 -   0,13   0,16   0,10 CL   0,0041 CF   0,0080]
    5. Rollout          21/16 7/1*                   Eq.:  -0,0679 ( -0,1917)
        46,81  10,37   0,35 -  53,19  14,07   1,17 CL  -0,0760 CF  -0,0679
      [  0,14   0,11   0,05 -   0,14   0,17   0,11 CL   0,0040 CF   0,0080]

    6. Rollout          21/10                        Eq.:  -0,1057 ( -0,2295)
        45,69  10,35   0,31 -  54,31  16,15   1,94 CL  -0,1240 CF  -0,1057
      [  0,44   0,33   0,10 -   0,44   0,48   0,51 CL   0,0135 CF   0,0255]
        216 games
    7. Rollout          8/2 7/2                      Eq.:  -0,1084 ( -0,2322)
        45,27  11,60   0,37 -  54,73  17,11   1,93 CL  -0,1051 CF  -0,1084
      [  0,43   0,43   0,06 -   0,43   0,75   0,61 CL   0,0139 CF   0,0269]
    8. Rollout          13/7 6/1*                    Eq.:  -0,1352 ( -0,2590)
        44,85  10,71   0,30 -  55,15  14,83   1,32 CL  -0,1032 CF  -0,1352
      [  0,47   0,42   0,10 -   0,47   0,77   0,53 CL   0,0143 CF   0,0328]

[close]