Game 4, move 20 : Forum 1-1

[diane]

Ok - What a lot of choices, I have put a fair selection, but feel free to ask for more, if one you want isn't there.

Forum, blue, to move 1-1

mB+HIRDMBjA4Ng:QYmkACAACAAA

[ah_clem]

Spoiler

I like the hitting play, 3/2(2)* and then switching to the 14 point to bring it within a direct shot of his blot on the 8.  This gives us a chance to win "directly" instead of playing a backgame.

The other option is to eschew the hit and embrace the backgame - we have an excellent backgame structure and pretty good timing and  I think shifting from the 23 to the 22 improves our backgame structure.  But I look at the backgame as a fallback here; since we have another path to winning, we should continue to pursue that avenue while we have the chance.  I don't think the hit destroys our timing - we'll still have a viable backgame if this doesn't work out, so lets go for it.


15/14 (2), 3/2*(2)

[close]

[stog]

Spoiler

yes that was my thinking too -- and then looking at it further i began to dislike messing with our board - such as it is - and indeed pushing 2 stones further into the void.

by moving our 'pack in the rear' 24/22, 23/22 (2)sry ty dorb 24/22 (2) we do indeed ensure a better backgame and maybe, by not hitting , affect our opponent's timing - but of course we will still have to move ourselves, so i am really still unsure about this

[close]

[dorbel]

Spoiler

I think playing 3/2*(2) is revolting. How do we plan to win this game? We have to hit, possibly several times, then contain the hit checker(s) behind a prime, or at some point in the distant future hit while we have a closed board. Moving two checkers to the 2pt now does not fit in with either of those plans. It isn't only that moving two checkers to the 2pt is so weak, although it is, but moving two checkers to the 22pt is so strong! Then we will have the best two back game points and be able to bring our other men around to build the prime that we are going to need. Hitting now in the hope of winning going forward is dreamland. Even the illegal 24/22, 23/22 which has attracted somebody's vote(!) is better than any play that hits.

[close]

[diane]

Spoiler

Even the illegal 24/22, 23/22 which has attracted somebody's vote(!) is better than any play that hits.

[close]

The move option is 24/22, 23/22 (2) - which isn't illegal...as far as I can count...

[dorbel]

Yes that is what it is now. It lost its vote as well!

[NIHILIST]

Spoiler

It's bad enough we've made our 3 point. Trying to go forward here is, IMHO, a huge mistake. We have great timing, the 3rd checker on his 5 point certainly helps too.

We have a great opportunity to considerably upgrade our backgame by moving both checkers from the 24 to the 22 point. That's my play.

[close]

Bob

[diane]

Spoiler

I am confused with this move, a lot of you are advocating moving two checkers from 24 to 22, giving a 2,3 backgame.  A lot of popular folklore states that a 1,3 backgame is the strongest there is...is the folklore wrong? or is there something special about this position that makes 2,3 stronger?

[close]

[roadkillbooks]

I like making the midpoint here.  Its a great standard goto point and we have been weak in our coverage of our outfield.  It will serve us well.  I still not overly concerned about the placement of the back pieces cause this is just starting.

[PersianLord]

Spoiler

24/22(2)
PL

[close]

[NIHILIST]

Spoiler

The popular folklore in our old chouette was that the most powerful backgames, in rank order, are 2-3, 1-3 and 1-2.

Probably the best and shortest answer I can give you why 2-3 is best is that, during opponent's bearin, he almost always must play his full roll, especially 5s and 6s. OTOH, the 1-2 is the worst for the converse reason; opponent can slow things down by NOT having to play 5s and 6s. Indeed, the best defense against the 1-2 is to pile your spares on the 6 and 7 points ( assuming a 5 point prime) precisely to avoid playing 5s and 6s. The 1-3 is kinda in the middle. The reluctance of players to move to the 2-3 position vs the 1-3 is rooted in the age old notion of holding opponents 1 point til the bitter end.

Set up positions which use the 3 different backgames against an identical opposing structure and do some rollouts. It won't take long for you to observe for yourself why the 2-3 is best.

[close]

Bob

[ah_clem]

The popular folklore in our old chouette was that the most powerful backgames, in rank order, are 2-3, 1-3 and 1-2.

Here's a link to a comparison of the relative merits of each structure, but it doesn't give much detail as to how they arrived at the numbers, other than to say that they were calculated by Snowie.


The relevant chart:

Anchor Avg. pts lost  proper cube action   

2-1      1.03         Double       Pass
3-1      0.80         No double    Take
4-1      0.94         Double       Take
5-1      1.24         Double       Pass
3-2      0.58         No double    Take
4-2      0.71         No double    Take
4-3      0.56         No double    Take
5-2      0.80         No double    Take
5-3      0.86         Double       Pass
5-4      0.88         Double       Pass

According to this, 3-2 is much better than 2-1, and   4-3 is  best.

(If you're feeling deja vu, it's because I posted this about two weeks ago in another thread.)

[dorbel]

This chart has no value if we can't see any of the other checkers! What NIHo has to say about the relative merits of the deep backgames (24-23, 24-22 and 23-22) is pretty accurate and I too prefer 23 and 22 by a bit, because it is as he says, much harder to kill numbers for the defender (the defender being the side that is bearing in). The positions where the checkers are seperated by two or three points (24-20, 23-20 and 24-21) are very weak because the points dont combine to force bad numbers for the defender. 20-21 and 21-22 do quite well against primes that have to be rolled home but lose most of their value once there are only two or three points to be cleared. The last two 20-22 and 21-23, both quite rare for some reason, are a bit of an unknown quantity, probably ranked somewhere in the middle of the list all other things being equal, which of course they can't be. All this is of course anecdotal and is likely to remain so in the absence of any data that can truly be said to compare one backgame with another and I can't see how you would ever be able to produce that. This is particularly true because many commentators dont trust computers to play back games well anyway, although IMO they probably in general play them better than humans!

[ah_clem]

bot results

Spoiler

gnubg two ply analysis has 24/22(2) and 15/14(2) 3/2*(2)  exactly equal.   But the rollout shows the hitting play to be a blunder.  Interesting.



    1. Cubeful 2-ply    24/22(2)                     Eq.:  -0.126
       0.467 0.104 0.003 - 0.533 0.249 0.017
    2. Cubeful 2-ply    15/14(2) 3/2*(2)             Eq.:  -0.126 ( -0.000)
       0.448 0.131 0.008 - 0.552 0.249 0.023
    3. Cubeful 2-ply    15/13 3/2*(2)                Eq.:  -0.141 ( -0.014)
       0.446 0.128 0.008 - 0.554 0.261 0.027
        2-ply cubeful prune [world class]
   


    1. Rollout          24/22(2)                     Eq.:  -0.085
       0.492 0.103 0.014 - 0.508 0.254 0.049 CL  -0.011 CF  -0.085
      [0.003 0.003 0.002 - 0.003 0.004 0.003 CL   0.010 CF   0.010]
       
    2. Rollout          23/22(2) 15/14(2)            Eq.:  -0.126 ( -0.041)
       0.473 0.102 0.013 - 0.527 0.258 0.062 CL  -0.056 CF  -0.126
      [0.003 0.002 0.002 - 0.003 0.004 0.004 CL   0.010 CF   0.011]
       
    3. Rollout          15/13(2)                     Eq.:  -0.179 ( -0.094)
       0.454 0.097 0.016 - 0.546 0.263 0.076 CL  -0.107 CF  -0.179
      [0.003 0.002 0.002 - 0.003 0.004 0.005 CL   0.009 CF   0.010]
     
    4. Rollout          23/22(2) 3/2*(2)             Eq.:  -0.203 ( -0.119)
       0.443 0.096 0.011 - 0.557 0.263 0.053 CL  -0.111 CF  -0.203
      [0.003 0.002 0.002 - 0.003 0.005 0.004 CL   0.010 CF   0.010]
       
    5. Rollout          24/22 3/2*(2)                Eq.:  -0.208 ( -0.124)
       0.437 0.105 0.015 - 0.563 0.267 0.056 CL  -0.120 CF  -0.208
      [0.003 0.002 0.002 - 0.003 0.004 0.004 CL   0.010 CF   0.011]
       
    6. Rollout          15/14(2) 3/2*(2)             Eq.:  -0.223 ( -0.139)
       0.432 0.102 0.011 - 0.568 0.260 0.068 CL  -0.129 CF  -0.223
      [0.003 0.002 0.002 - 0.003 0.004 0.004 CL   0.010 CF   0.010]
       
    7. Rollout          24/23 15/14 3/2*(2)          Eq.:  -0.236 ( -0.152)
       0.430 0.099 0.011 - 0.570 0.270 0.072 CL  -0.145 CF  -0.236
      [0.003 0.003 0.002 - 0.003 0.004 0.005 CL   0.011 CF   0.012]
       
    8. Rollout          15/13 3/2*(2)                Eq.:  -0.247 ( -0.162)
       0.429 0.098 0.013 - 0.571 0.276 0.081 CL  -0.162 CF  -0.247
      [0.004 0.003 0.002 - 0.004 0.005 0.005 CL   0.011 CF   0.012]
       
    9. Rollout          23/22 15/14 3/2*(2)          Eq.:  -0.255 ( -0.170)
       0.429 0.098 0.014 - 0.571 0.281 0.090 CL  -0.166 CF  -0.255
      [0.003 0.002 0.002 - 0.003 0.004 0.005 CL   0.011 CF   0.011]
        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]