Game 6, move 5 : Forum 5-4

[diane]

I think this is pretty nextish, but thought we would have a quick look at it...I will probably move it along very swiftly.

Forum, blue, to move 5-4

mGe3AATQ58EAMg:8ImyACAAIAAA

[ah_clem]

Spoiler

Hoo boy.  These are the positions where I make a dog's breakfast out of things.  I'm really not sure what to do, but I have to decide something so here goes:

I don't like running the backman with 22/13.  We'll be 25 pips behind after moving, and we need to maintain contact.  If we weren't so far behind in the race I'd probably do this.

We can hit with 13/4 or 8/4 8/3, but then we are basically playing a voluntary back game as he's almost certain to send a fourth checker back.  If we were farther behind in the race, I'd probably play 13/4, but right now our timing isn't right for a backgame (disclaimer - I suck at backgames).  Plus 13/4 abandons the midpoint - not a good idea here.

Making our 1-point just burries two checkers when we want to maintain contact. Nah.

Slotting his bar point with 22/18 just seems wrong - even if we make the bar point (doubtful), now that he's cleared his midpoint  it won't do us much good.

That leaves playing checkers behind his runner, hoping to build a board and get a shot.  I think that's the right approach here, although I hate playing behind the runner when maintaining contact is so essential.  It's not on the menu, but I'd like to order  8/3 6/2.

This unstacks the 6 and the 8 points, and maintains the slot on the 5. I think it's better than 8/3 5/1.

Ok.  That's my analysis.  Now I'm ready to hear why it's the wrong approach.... (c:

[close]

[stog]

looks a good move.. 

[diane]

Spoiler

It's not on the menu, but I'd like to order  8/3 6/2.

[close]

*cough*, there is a reason it isn't on the menu...however...I searched the freezer and found it for you...vote away.

Looks like this isn't anything like so cut and dried as I imagined, I will give this longer....

[dorbel]

Spoiler

I don't think hitting (13/4*) is bad, but with 10 men already in forward positions it's going to be hard to time a backgame. 8/3, 8/4* may be more feasible along the same lines. Both of these plays venture into the unknown as you might say. I would just play the simple 22/13, although if a partner wanted to go for one of the hit options I wouldn't argue too strongly against him. Sorry Clem, I think just burying checkers is awful! There's no other word for it. Running or hitting both have a reason and a game plan in mind, dumping doesn't.

[close]

[diane]

Spoiler

I really liked the 22/13 option, and surely thought it would be the one and only choice.  We retain the ace point if a back game is called for - but dont have too many back if we get the upper hand - and have more ammo to throw to try and get that upper hand. 
You dont see these kinds of arguments in any backgammon books, do ya   

[close]

[stog]

Spoiler

yep 22/13 for me too, strengthening our outer board - a chance to make our 5 point if rkb doesnt hit and if he does - its b'game for us boys -- or r we already there?

[close]

[stiefnu]

Spoiler

a chance to make our 5 point if rkb doesnt hit and if he does - its b'game for us boys -- or r we already there?

I'm inclined to think we're already headed that way, so I'm the lonely candidate for 13/4*.  If we get hit, we have a reasonable chance of dropping a second anchor and then have good timing.  If not hit, we'd hope to start building a board or outfield.  Even after this move we'll be 27 pips behind, so why run?

[close]

[dorbel]

Spoiler

Technically speaking, a backgame requires the player to hold two (or more) points in his opponent's home board. Holding a single point is a holding game. Note that some of the possible combinations of two points, notably 20-24, 20-23, 21-24 and perhaps 21-23, don't really offer any significan advantages over holding games, as the points held don't combine to create bad numbers for the opponent. "Let's get into a back game" often fails because the points obtained are useless or because one ends up with lots of checkers back but only a single point. 13/4* here probably is a reasonable play, as is the ultra comittal 8/3. 8/4* (!), but I didn't choose it because I haven't the faintest idea of how to put a value on all the possible outcomes. Either would be a fun play though!

[close]

[ah_clem]

gnubg rollout

Spoiler

Another close one.  The two hitting plays are a statistical tie with the running play. 

Fortunately, the forum chose one of the winners this time. Unfortunately, I didn't.  Playing behind the runner is a .07 error.




    1. Rollout          13/4*                        Eq.:  -0.509
       0.295 0.081 0.013 - 0.705 0.409 0.035 CL  -0.509 CF  -0.509
      [0.003 0.003 0.002 - 0.003 0.004 0.003 CL   0.009 CF   0.009]

    2. Rollout          8/4* 8/3                     Eq.:  -0.514 ( -0.005)
       0.297 0.065 0.011 - 0.703 0.461 0.039 CL  -0.514 CF  -0.514
      [0.003 0.003 0.002 - 0.003 0.004 0.003 CL   0.010 CF   0.010]

    3. Rollout          22/13                        Eq.:  -0.515 ( -0.006)
       0.260 0.053 0.007 - 0.740 0.241 0.010 CL  -0.515 CF  -0.515
      [0.002 0.003 0.002 - 0.002 0.003 0.001 CL   0.005 CF   0.005]

    4. Rollout          22/18 8/3                    Eq.:  -0.556 ( -0.047)
       0.253 0.058 0.007 - 0.747 0.361 0.019 CL  -0.556 CF  -0.556
      [0.002 0.003 0.003 - 0.002 0.003 0.002 CL   0.006 CF   0.006]

    5. Rollout          8/3 6/2                      Eq.:  -0.578 ( -0.069)
       0.244 0.061 0.008 - 0.756 0.367 0.021 CL  -0.578 CF  -0.578
      [0.003 0.003 0.002 - 0.003 0.003 0.002 CL   0.007 CF   0.007]

    6. Rollout          22/18 13/8                   Eq.:  -0.578 ( -0.070)
       0.243 0.054 0.005 - 0.757 0.375 0.020 CL  -0.578 CF  -0.578
      [0.002 0.003 0.001 - 0.002 0.003 0.002 CL   0.006 CF   0.006]

    7. Rollout          6/1 5/1                      Eq.:  -0.581 ( -0.072)
       0.229 0.060 0.005 - 0.771 0.276 0.011 CL  -0.581 CF  -0.581
      [0.002 0.003 0.001 - 0.002 0.003 0.001 CL   0.004 CF   0.004]

    8. Rollout          13/9 13/8                    Eq.:  -0.589 ( -0.080)
       0.241 0.049 0.006 - 0.759 0.383 0.023 CL  -0.589 CF  -0.589
      [0.003 0.003 0.001 - 0.003 0.003 0.002 CL   0.006 CF   0.006]

    9. Rollout          22/18 6/1                    Eq.:  -0.604 ( -0.095)
       0.228 0.048 0.006 - 0.772 0.375 0.018 CL  -0.604 CF  -0.604
      [0.002 0.003 0.002 - 0.002 0.003 0.002 CL   0.006 CF   0.006]

   10. Rollout          8/3 5/1                      Eq.:  -0.618 ( -0.110)
       0.218 0.055 0.005 - 0.782 0.321 0.016 CL  -0.618 CF  -0.618
      [0.002 0.004 0.001 - 0.002 0.003 0.002 CL   0.006 CF   0.006]
        Full cubeful rollout with var.redn.
        1296 games, Mersenne Twister dice gen. with seed 872235963 and quasi-random dice
        Play: 0-ply cubeful prune [expert]
        Cube: 0-ply cubeful prune [expert]

[close]

[adrian]

I will keep following the onlinematches, but seems I will never vote in one again. The pace is too fast for my agenda.  Don't take this as blame Diane (and all the others who were working on this GREAT topic.) 24 hours for voting is a must for a poll imho. 

[stog]

i think you will find the pace of these matches varies a lot -- even with the same facilitator on the job - as life intervenes

this match has been running fasy slow fast seeeeellllllooowwwww and has been looked after by a few of us now. Diane is doing well and her preferred pace i think is brisk..... but i think zorba may take a little more time -- and then -- it depends on the other player too -- rkb has been vey available and up for a brisk match -- so all has been well and we are still getting good voting numbers even on shorter plays!

i trust you will vote when you can - it is fun to vote then read y what u voted for was useless! and so often true i assure you

thx to all of you for contributing to this very educational board.

btw zorba r u up for the next as top player or fibsboarder?

and dorbedl you really must play too and nihooooo!

maybe even vic!

[diane]

24 hours for voting is a must for a poll imho.  

Poll was posted «  on: October 22, 2009, 08:02:27 PM » [strange fibsboard time  ]

It was closed earlier this evening, approx  October 24, 2009, 06:50:56 PM  [still strange fibsboard time  ]

That means this was open for about 46 hours    

I have been asked repeatedly to move it along more quickly, by a lot of the regulars - so sorry Adrian, you will have to visit a bit more frequently  

[ah_clem]

24 hours for voting is a must for a poll imho. 

I'd agree with that - 24 hours is about the right amount of time for a move.  The thing is, this move had been open for 46 hours and diane sensibly moved it along after 24 hours of inactivity.

So, my $.02 is to leave the poll open for 24 hours, but move it along earlier if the vote is really lopsided.  I don't think we've been getting very many responses in the second 24 hours, so the 48 hour window doesn't add much in the way of responses.  With a standard 24 hour vote period, participants would know to check in once a day, and most of the time there'd be a move to consider.

Of course, the moderators and the box have lives, so if a move is a day late that's no big deal.