Game 4, move 30 Blitzxz and 31 Forum : Forum 5-3

[diane]

Blitzxz rolled 5-5, forced.

Forum, blue, now has 5-3 to move.

H28HACA22yQCGA:QYmuACAACAAA

[ah_clem]

Spoiler

In our last episode, dorbel was explaining that we'll be willing to abandon the bar point in order to make the 4 point.  But is this really what we want to do?

Playing 9/4 7/4 exposes us to a direct shot, and if blitzy hits he'll have clear sailing to a win, probably a double pass leaving us at Crawford.  The chances of him hitting are about 1 in 3.

I still like the idea of liberating the runners.  With 23/18, white has some numbers to whomp us, but all except 4-2 and 1-1 vacate a valuable homeboard point or expose a juicy blot.  With that stack on the ace point, I'm not too worried about being hit.  Meanwhile, we contain the runner quite will with sixes blocked and the outfield covered.

I say vamoose with the runner. 23/18. Where to play the 3?

9/6 is too passive.  11/8 may be too aggressive, giving 2 more shots. 14/11 makes a point, but it's a fairly useless point and cedes outfield control.  So that leaves continuing on with 18/15.

23/15  With that play, he's got two numbers to escape and hit (3-5) and three numbers to point on our head (4-2, 1-1).  But even pointing on our head isn't decisive, since he'll still have a hard time escaping resulting in a board-crunch.  Double two's threes and fours play Ok.  That's 5 good rolls and three so-so rolls; the rest put him in a world of hurt.   Compare that to the 11/36 chance of him icing it with a direct hit/escape.

I don't think making the four is the way to go here.  Perhaps dorbel can explain why I've gotten it backwards again. (c:

[close]

[/spoiler]

[dorbel]

Spoiler

Making the 4pt is the way to go. Granted we won't be ecstatic if he does throw a 6, but he isn't going to be doubling us out if he hits! When he misses, he has 13 blot numbers in his own board and we can attack the blot on the ace or make a six prime. When he does hit, the five men already out of play and the gap on the 3pt are going to make it very hard for him to bear in safely and we will have a nice board ready for anything we can hit. I realise that I'm not going to be able to talk most people in to this, but please don't choose to leave the anchor! This is just so not the time to have five (!!) blots and a gap on our 4pt.
Backgammon is about trying to win, rather than trying to avoid losing. There is a large flashing sign pointing to our 4pt saying "Make me!" Let's do it. 23/16 is a humungous double whopper blunder if ever I saw one and any play other than making the 4pt is a large error. probably.

[close]

[diane]

Spoiler

I am afraid of this 'making the 4 pt' approach, even though I know it is right.  I am trying to summon the courage to change my vote, it would help if Clem can be strong too  

When I didn't vote for that, I was trying to duck the 'making the 4 now' dilemma, by sticking another builder in - playing 14-6

Meaning we keep the 23 anchor for the relatively small number of times he hits, and yet have another checker, so if he doesnt hit, and we find ourselves again needing to use a checker from the 7 to make the 4, we have one available.  There is also the tempting 'loads of rolls' where something bad happens for him at home, and we can get another checker  [note the mathematical expression of probability there..'loads', this is an internationally recognised term.]

Clem - did you add in an extra option there? I am sure I thought 23-15 was so horrible I wasn't even going to put the idea in their heads  

[close]

[ah_clem]

Spoiler

Diane, yes I added the 23/15 option.  I didn't even look at the options when making my original choice and only noticed it was missing when I went to vote.  Seeing that it wasn't there, I helped myself and added it.  Hope you don't mind - it seems easier to just do it rather than asking you to do it for me.

Re-evaluating the position, I see the value of keeping the anchor.  It really does make his life miserable by providing the opportunity to throw awkward rolls. If we could move both checkers out, I'd probably do so and make the runner walk the "path of fire", but leaving one checker behind offers too much of a target.  If the runner doesn't escape, he has a bunch of bad rolls, and if the runner does escape, we have the timing to wait for the nasty rolls to appear eventually.  The stack o' burried checkers means he'll likely leave a shot early in the bearoff if it gets to that.

So, now that I've convinced myself that the anchor is too valuable an asset to give up, I'll consider other moves.  The logic of making the four point is that if he doesn't hit, we're in great shape.  And if he does hit we get extra timing to wait for a shot.  The five point homeboard vs the candlestick on the ace point is a strategic assymetry that won't change soon.

What about other options?  Slotting the four point looks attractive because we get to keep bar point as a blocking point, but it doesn't provide the strategic long term advantage of the solid five point board.  If he hits, he'll probably escape and we're back to playing for a shot, but without as solid a board. So, I don't think it's as good.

Shuffling the outer checkers works as long as his runner stays put. But if he escapes, again, we don't have the big asset of a solid board.

So, reluctantly, I'm going to go with making the four.  I see the reasoning (I didn't before) but I'm not convinced that it's best, and I doubt I'd have the cojones to play it over the board.

[close]

[diane]

Spoiler

Diane, yes I added the 23/15 option.  Hope you don't mind - it seems easier to just do it rather than asking you to do it for me.

[close]

Spoiler

No problem at all on adding extra options, unless they happen to be that one 

Yes, it is easier than asking, when you have the power.

So - you were brave, it is up to me now...I might just see what everyone else does first - i shall make my final decision in the morning when i come to update it 

[close]

[stog]

Spoiler

tactical voting here! i have switched to 14/6 from my original choice 14/9 11/8 -i don't want to slot yet! nor break up our "reserves" - indeed i think that if he doesn't '2' or hit next, then there is a good chance his board will crumble roll after next, and i want to be there for the spoils - we have if we don't throw huge dubbs - good timing, and if we huge dubb then we will be good to catch his straggler - but all this is in the gods, and why risk a possible match loosing dance on the bar? although i do see mr dorbel's good thinking on making the 4, or even having a good chance to b'game if he was to pick up our other stones

[close]

[ah_clem]

Spoiler

Speaking of tactics, I think the big dilemma here is strategy vs tactics.

Keeping the bar point is the tactical approach - this play seems much better if you look at what's likely to happen in the next roll.  Making the 4 point is the strategic approach - by creating a nearly airtight board we solidify our position which will remain an asset for the rest of the game.  We have to take a considerable risk to attain this permanent strategic advantage - is it worth it?

IOW, which trumps which here?  Hard to say.  I'm basically deferring to dorbel here.  It'll be an interesting rollout, probably worth doing something more than 0-ply.

[close]

[ah_clem]

And the winner is.................

rollout

Spoiler

It's not even close.  The strategic play of making the four point is much better than the outfield-checker-shuffling plays.  And abandoning the anchor is a triple-whopper.

A very eye-opening position. 



    1. Rollout          9/4 7/4                      Eq.:  +0.753
       0.750 0.194 0.022 - 0.250 0.120 0.018 CL  +0.782 CF  +0.753
      [0.002 0.003 0.003 - 0.002 0.002 0.002 CL   0.007 CF   0.008]

    2. Rollout          14/9 11/8                    Eq.:  +0.673 ( -0.080)
       0.718 0.178 0.017 - 0.282 0.093 0.008 CL  +0.722 CF  +0.673
      [0.002 0.003 0.002 - 0.002 0.003 0.001 CL   0.007 CF   0.007]

    3. Rollout          11/8 9/4                     Eq.:  +0.658 ( -0.094)
       0.725 0.181 0.019 - 0.275 0.143 0.026 CL  +0.689 CF  +0.658
      [0.002 0.003 0.002 - 0.002 0.002 0.003 CL   0.007 CF   0.008]

    4. Rollout          14/6                         Eq.:  +0.648 ( -0.105)
       0.707 0.179 0.019 - 0.293 0.098 0.007 CL  +0.702 CF  +0.648
      [0.002 0.003 0.002 - 0.002 0.003 0.001 CL   0.007 CF   0.008]
        Full cubeful rollout with var.redn.
        1296 games, Mersenne Twister dice gen. with seed 845045210 and quasi-random dice
        Play: 0-ply cubeful prune [expert]
        Cube: 0-ply cubeful prune [expert]

[close]

[dorbel]

That's a relief! I was stricken with doubts after being so dogmatic.