Match 1, Game 3, Move 2, factotum

[blitzxz]

2-1?

Score:
Black(forum) 2, White(factotum) 0, match to 5

Pips:
Black 162 White 161

GnuID:
4HPkASLg8+ABKA:MIGoAAAAEAAA

[diane]

Spoiler

I am really glad the game is going this way - this is a typical rubbish roll that I wouldn't know what is the best thing to do with.  I shall be glad to hear what the right thing to do is 

I would probably try to give myself more things that could be good in the future..hmm think that is what they call flexibility  by moving 13-11, 23-22, but hate it - cos it is a lot of blots 

Oh, and I might do a bit of bellyaching in shouts about rigged dice 

[close]

[playBunny]

Spoiler

this is a typical rubbish roll that I wouldn't know what is the best thing to do with. 

A common tactic is to hit loose, 6/4*, and then figure out what to do with the other dice. This would certainly be the case with the 5-point and the 4-point isn't too shabby for that treatment. Generally, three blots on the table should immediately remind the nostrils of the smell of gammon, even if there is plenty to be done to get it.

In this case the candidates for the 1 that I'd consider are 24/23 to make a deep anchor or 23/22 to advance the split backrunners. factotum doesn't really need an anchor at this stage, it reduces flexibility when, having the tempo, he isn't in that great a danger. Advancing the backrunner has the advantage here because our blot on 14 is vulnerable (that slight waft of gammon). If we miss factotum's slotted 4-point and he rolls 4-4 then he covers it and gets to hit our 14. That would be a nice joker for him but, on the pessimistic side, 23/22 also has advantages if we hit his slotted 4 because the chances of him making an decent anchor are better if he retains the split.

So it would be 6/4* 23/22 for me.

Oh, and I might do a bit of bellyaching in shouts about rigged dice 

[close]

[factotum]

Well neither side has made a point yet, so making your 3pt is too tame.   I should try for better.  Creeping up my back checker with 24-21 would leave you with the better attacking position and on roll.  I could try 23-22, 13-11, but having started my 4pt you are still in the better position and on roll.  I had the first roll, and to keep any kind of advantage now i must hit on my 4pt with the 2.   That leaves the best 1 as 23-22.     6-4*, 23-22.

[Zorba]

factotum and playBunny gave good explanations I think. 13/11 is not an option here, since we split to the 4pt. That means all sevens would hit, plus the usual 6-4. Sevens are six shots (the worst indirect), plus those two shots from the other backchecker makes eight shots total, that's quite a lot. Not worth taking such a risk for just getting one extra builder.

I'd guess the two real candidate plays are somewhat close, especially if the matchscore were even. But I still like knocking off an opponent of the fourpoint, if at all feasible. The 22pt anchor is decent, but gives away the initiative and lets us play freely. Surely, hitting loose on the fourpoint gives us a lot more shots than the 11pt builder would, but the (possible) gains are much higher:

• make it harder for the team to make the good 21(4)pt anchor
• take away half our roll, thereby taking away a lot of our point making rolls
• a favourite to make the good fourpoint when missed
• a small bonus overall, but still: we may roll 6-6, an otherwise very good roll

With the matchscore in mind, 6/4* becomes clearer, since it's quite obviously a more gammonish play.

For the one, 24/23 is way too conservative with opponent on the bar. Clearly better to play 23/22 and strive for something better while we're on the roof. Again the matchscore makes this even clearer, factotum shouldn't be worried much about losing a gammon here, and therefore needs an anchor less than usually.

[lewscannon]

Spoiler

6-4, 23-22 seems like the best thing to do with this mediocre roll, for all the reasons stated by the others. It's interesting to me to see how well thought out other player's strategies (with the obvious exception of diane  ) are, where I mostly just go with a first impression and instinct. It's also nice to know that a lot of the time I arrive at the right place as them.

[close]

[playBunny]

It's interesting to me to see how well thought out other player's strategies [...] are, where I mostly just go with a first impression and instinct.

I mostly play from intuition as well. Logic such as I've presented is after-the-fact justification and can sound just as impressive and convincing when I'm providing the rationale for a double-whopper!

What I find interesting, self-reflection-wise, is those occasions when I'm about to make a move that does have a readily available heuristic, such as "extend the back end of the prime", but I get a feeling that it's better to do such and such. I may or may not go with that feeling (if I don't it's usually cauition that makes me stick with the familiar) but often, when I ask GnuBg what I should have done, it confirms the feeling as having been correct, to my pleasant surprise. I then update my thinking with either a new rule or an exception to the old rule for that kind of position. But the fact that I had the let's do it differently here feeling in the first place means that my intuition knew what to do better than my logical mind. It shows that intuition is a growing front.

[diane]

It's interesting to me to see how well thought out other player's strategies (with the obvious exception of diane  )

I will have you know I thought about that for ages before coming to the wrong conclusion 

I didn't like the hit, because that blot is soooo going back to the beginning, and with my danceability, that is usually game over for me  Especially when I am playing Lews!!      The match score was relevant too, at 0-2, I would really want to avoid being gammoned, and three checkers back seems a good way to try my best to make that happen....
I try to think 'recycling', but the hit still didn't seem to gain much to me - but, I shall see how it pans out. And reading the other opinions certainly helped 

[socksey]

I woulda made the "too lame" 3 pt!     But, what do I know?!   

socksey



"To know what is right and not to do it is the worst cowardice." â€" Confucius

[Zorba]

A quick'n'dirty 0-ply rollout:

Spoiler

    1. Rollout          23/22 6/4*                   Eq.:  +0,0573
        49,27  13,24   0,57 -  50,73  15,46   1,96 CL  +0,0303 CF  +0,0573
      [  0,19   0,14   0,03 -   0,19   0,25   0,19 CL   0,0057 CF   0,0127]
    2. Rollout          24/23 6/4*                   Eq.:  +0,0266 ( -0,0306)
        48,25  12,96   0,55 -  51,75  15,74   1,75 CL  +0,0021 CF  +0,0266
      [  0,17   0,14   0,08 -   0,17   0,24   0,18 CL   0,0057 CF   0,0123]
    3. Rollout          24/22 23/22                  Eq.:  +0,0072 ( -0,0501)
        48,62  10,54   0,47 -  51,38  12,39   1,07 CL  -0,0039 CF  +0,0072
      [  0,20   0,15   0,07 -   0,20   0,23   0,16 CL   0,0060 CF   0,0126]
    4. Rollout          6/4*/3                       Eq.:  -0,0044 ( -0,0616)
        48,15  12,71   0,61 -  51,85  15,97   2,24 CL  -0,0136 CF  -0,0044
      [  0,22   0,18   0,07 -   0,22   0,29   0,25 CL   0,0067 CF   0,0144]
    5. Rollout          23/22 13/11                  Eq.:  -0,0430 ( -0,1002)
        46,85  11,69   0,59 -  53,15  15,68   1,33 CL  -0,0557 CF  -0,0430
      [  0,28   0,26   0,14 -   0,28   0,35   0,23 CL   0,0095 CF   0,0192]
    6. Rollout          24/21                        Eq.:  -0,0465 ( -0,1037)
        47,17  10,29   0,44 -  52,83  14,37   1,77 CL  -0,0546 CF  -0,0465
      [  0,27   0,22   0,05 -   0,27   0,31   0,33 CL   0,0084 CF   0,0178]

        Full cubeful rollout with var.redn.
        648-1944 games, Mersenne Twister dice gen. with seed 819749348 and quasi-random dice
        Play: 0-ply cubeful prune [expert]
        Cube: 0-ply cubeful prune [expert]

[close]

[dorbel]

Don't be too sad socksey. Making the 22pt anchor probably isn't the best play, but it isn't very bad either, only a small mistake at worst. Taking a reasonable anchor preserves you from being knocked out by a bad joker sequence and exerts some pressure on the opponent's blots. The hit play here is to take away half the next roll, a commonplace tactic in the early game, but anchoring gives you a solid base from which you can attack next turn if the opponent can't tidy up his blots. Denying your opponent any target to attack is a form of pressure in itself.

[NIHILIST]

It's known as THE FENCEPOST PLAY.

Bob

[socksey]

   I know the devil made you do that, Niho!   

Seriously, thank you, dorbel, for that extensive comment.   

socksey



"Marriage changes passion. Suddenly you're in bed with a relative." - Anonymous

[playBunny]

 
GnuBg Rollout 1296 trials, 2-ply for 12 plies, 0-ply thereafter, 2-ply cube throughout

Spoiler

Hitting is the order of the day. The top two moves are both very acceptable as shown by the difference between them (0.009) compared with the standard error (0.011).

# .. Move ......... Equity .. (Diff) .... Win ... WinG ... WinBg -- Lose ... LoseG .. LoseBg
1 .. 23/22 6/4* .... +0.045 ............ 48.7% .. 12.8% ... 0.7% -- 51.3% .. 15.0% ... 1.4%
2 .. 6/4*/3 ........ +0.036 (-0.009) ... 48.5% .. 13.1% ... 0.6% -- 51.5% .. 15.7% ... 1.8%
3 .. 24/23 6/4* .... +0.015 (-0.031) ... 47.7% .. 12.8% ... 0.6% -- 52.3% .. 15.2% ... 1.9%
4 .. 24/22 23/22 ... -0.003 (-0.048) ... 48.0% .. 11.0% ... 0.3% -- 52.0% .. 12.1% ... 1.0%
5 .. 23/22 13/11 ... -0.058 (-0.103) ... 46.7% .. 11.5% ... 0.5% -- 53.3% .. 15.5% ... 1.3%
Std Err ............. 0.011 ............. 0.2% ... 0.2% ... 0.1% --- 0.2% ... 0.3% ... 0.2%

        Full cubeful rollout with var.redn.
        1296 games, Mersenne Twister dice gen. with seed 858403456 and quasi-random dice
        Play: supremo 2-ply cubeful prune [world class]
        keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
        Skip pruning for 1-ply moves.
        Cube: 2-ply cubeful prune [world class]
        Different evaluations after 12 plies:
        Play: 0-ply cubeful prune [expert]
        Cube: 2-ply cubeful prune [world class]

[close]

[dorbel]

There is a point to this and it is that there isn't, in many cases an absolute best play. There is, as there is here, a theoretical best play, but note that this is the best play for a bot playing near perfect backgammon. For a player of more moderate skills, it is quite possible that a simpler game plan, such as making the 22pt anchor here, is better, because it will lead to a game that is easier to play. For a strong checker player like factotum for example, the hit is clear and we can assume that he will use the cube well if he can engineer a suitably gammonish position. For socksey, it may well be that the simplifying play is better, because she feels comfortable with an anchor and because she will find the cube easier to handle.
Just as strong players seek complication and risk as a matter of course, it seems probable to me that intermediates are often better served doing what they know, which is keep it simple.