Search This Blog

August 19, 2012

PaulofAtlanta

Below are selected errors from an online match against PaulofAtlanta.

First game of 7 point match.  64 to play.
is Paulofatlanta

score: 0
pip: 160
7 point match
pip: 167
score: 0

is Chutzpah
XGID=-a--a-E-C---dE--ac-e----B-:0:0:1:64:0:0:0:7:10
to play 64

1.XG Roller++24/14eq: -0.223
Player:
Opponent:
45.46% (G:9.25% B:0.35%)
54.54% (G:15.37% B:0.55%)
2.XG Roller++8/2 6/2eq: -0.230 (-0.006)
Player:
Opponent:
44.60% (G:11.98% B:0.42%)
55.40% (G:16.03% B:0.80%)
3.XG Roller++24/18 8/4* eq: -0.258 (-0.034)
Player:
Opponent:
44.85% (G:11.02% B:0.49%)
55.15% (G:17.96% B:0.90%)
eXtreme Gammon Version: 2.02, MET: Kazaross XG2

All this work on reply rolls and I still make mistakes!! I have a tough time remembering this one. I played down and hit loose (13/7, 8/4*).  Apparently this is an exception to the hit loose when the opponent splits to the 5 point or 4 point plays. 

Looking at my earlier posts (http://bigmoneybill.blogspot.com/2012/06/replies-to-43z.html) I even made the comment "Alert!  I play this wrong currently."  Looks like I still do. 

I think I fiinally figured out why this play is wrong.  Let's look at the non-double rolls: 56, 46, 36, 26, 16, 54, 53, 52, 51, 43, 42, 41, 32 and 31 all hit back from the bar.  Only 21 misses!!  So this play is basically impossible to work. 

The running play 24/14 gets hit a lot too but might work, and if he does get hit then at least the opponent is distracted from making an inside point or anchoring in my home board.  Also, I lose less in the race getting hit there.

And while making the 2 point (second best) isn't the most desirable point, at least this play does something positive (for sure!!).

***********

Later in the same game.  7 away  7 away.  Cube action?

is Paulofatlanta

score: 0
pip: 155
7 point match
pip: 160
score: 0

is Chutzpah
XGID=-b--BaDBB---bB---cabbb--C-:0:0:-1:00:0:0:0:7:10
on roll, cube action?

Analyzed in XG Roller++ No double Double/Take
  Player Winning Chances: 70.19% (G:27.60% B:2.50%) 70.26% (G:27.38% B:2.48%)
  Opponent Winning Chances: 29.81% (G:8.66% B:0.58%) 29.74% (G:8.61% B:0.59%)
  Cubeless Equities +0.619 +1.269
Cubeful Equities
     No double:+0.808 (-0.192)
     Double/Take:+1.039 (+0.039)
xg Double/Pass:+1.000
Best Cube action: Double / Pass
eXtreme Gammon Version: 2.02, MET: Kazaross XG2


PaulofAtlanta doubled and I took.
The position looks to be a small drop. Let's use Roberties old "PRaT" system of evaluation - Position, Race, and Threats.

Position: advantage Paul. He has the better board, better prime, and I have three ugly guys on the ace point.

Race: small advantage Paul. The race is close but Paul is ahead and on roll.
Threats: Paul's rolls are well diversified. He is threatening to make the bar and/or simply make an advanced anchor on my 5 point. Based on the above, it looks bad like a pass.

And apparently it is per the computer. Three on the ace spells danger - but I do have counterplay and I do have an anchor. as the computer indicates I do win a fair amount of the time (~30%) but the gammons are too high. XG gives Paul a 20% net gammon advanatge. Too much vig to eke out a take.

************

I trail 6 away 5 away and have a 3rd roll 41 to play. Your move?
is Paulofatlanta

score: 2
pip: 163
7 point match
pip: 161
score: 1

is Chutzpah
XGID=-b----E-D---eD---b-db--AA-:0:0:1:41:1:2:0:7:10
to play 41

1.XG Roller++23/22 13/9eq: -0.247
Player:
Opponent:
42.68% (G:11.43% B:0.52%)
57.32% (G:17.53% B:0.83%)
2.XG Roller++23/18eq: -0.273 (-0.025)
Player:
Opponent:
42.88% (G:9.97% B:0.45%)
57.12% (G:18.32% B:0.71%)
3.XG Roller++13/8eq: -0.278 (-0.031)
Player:
Opponent:
42.25% (G:10.67% B:0.45%)
57.75% (G:17.71% B:0.67%)
4.XG Roller++24/23 13/9eq: -0.294 (-0.047)
Player:
Opponent:
40.96% (G:11.04% B:0.38%)
59.04% (G:16.15% B:0.72%)
5.XG Roller++23/22 8/4eq: -0.350 (-0.102)
Player:
Opponent:
40.78% (G:10.01% B:0.45%)
59.22% (G:19.23% B:0.92%)
6.XG Roller++13/9 6/5eq: -0.358 (-0.110)
Player:
Opponent:
41.46% (G:10.69% B:0.49%)
58.54% (G:21.39% B:1.49%)
eXtreme Gammon Version: 2.02, MET: Kazaross XG2

My play ( 13/9, 6/5) is way way down on the best list - 6th best!!

This is the "splot" play - or slotting while split.  This is usually wrong, and a whopper in this case when the opponent already has the 5 point made.  "Spreading the legs" like this simply allows the opponent to have his way with me next turn, never a good idea in backgammon!

I saw the stacks and wanted to take a chance now before it was too late, but I guess my position isn't desperate enough to justify this risky manuever.  The computer likes anchoring on the 2 point - a beginner's play but it is an improvement and less risk.  My 9 point builder is the best at making a point naturally next turn and the risk is much smaller.
************

Later in the same game.  How to play this 51?

is Paulofatlanta

score: 2
pip: 146
7 point match
pip: 161
score: 1

is Chutzpah
XGID=----BbCBBA--cB-a-c-cc--BA-:0:0:1:51:1:2:0:7:10
to play 51

1.XG Roller++23/18 9/8eq: -0.638
Player:
Opponent:
35.20% (G:6.46% B:0.25%)
64.80% (G:22.17% B:1.10%)
2.XG Roller++24/18eq: -0.712 (-0.074)
Player:
Opponent:
32.41% (G:6.25% B:0.28%)
67.59% (G:20.49% B:1.05%)
3.XG Roller++23/22 9/4eq: -0.731 (-0.094)
Player:
Opponent:
32.18% (G:5.97% B:0.21%)
67.82% (G:20.61% B:0.99%)
eXtreme Gammon Version: 2.02, MET: Kazaross XG2

I ran the opponent's bar point with 24/18. Right idea, but wrong execution. The opponent has an anchor and I don't. This means I desperately need one too before it's too late. So going to the opponent's bar point is necessary. However, I shouldn't leave that extra blot on my 9 point. My play has the virtue of keeping the anchor but that's all I have going for it. note the diversifcation of Paul's numbers - 4's hit me on the outside while 6's, 1's and 3's all hit on the bar point. Not too many bad rolls there - stupid play really.

The best play is the triple split 23/18, 9/8. This cleans up a blot while making the needed attempt at an advanced anchor. Classic distraction play where many of his hits are on the outside instead of making an inside point. Additionally, on some hits I get return shots from the bar or at least the chance to make an anchor from the air.

************

I trail 6 away 4 away, have already doubled and have boxes to play.

is Paulofatlanta

score: 3
pip: 98
7 point match
pip: 153
score: 1

is Chutzpah
XGID=--a-aDC-B-A--B-A--bbabccB-:1:-1:1:66:1:3:0:7:10
to play 66

1.XG Roller+15/9 10/4* 8/2*(2) eq: +0.386
Player:
Opponent:
62.71% (G:27.41% B:1.70%)
37.29% (G:15.36% B:1.89%)
2.XG Roller+15/9 13/7(2) 10/4* eq: +0.290 (-0.096)
Player:
Opponent:
62.66% (G:19.63% B:1.32%)
37.34% (G:17.01% B:2.10%)
3.4-ply13/7 10/4* 8/2*(2) eq: +0.323 (-0.063)
Player:
Opponent:
61.17% (G:26.34% B:1.49%)
38.83% (G:16.42% B:2.57%)
eXtreme Gammon Version: 2.02, MET: Kazaross XG2

This one really should have been easy but I managed to screw it up.   I moved 15/9, 13/7(2), 10/4*.  This play goes for structure by making a nice prime and putting him on the bar.  Not a terrible play but way wrong.

Why?  becuase of the gammons you dummy.

Best by far is 15/9, 10/4*, 8/2*(2).  This play doesn't look as pretty but puts two checkers on the bar with enormous pressure on Paulofatlanta to perfrom.  Clearly this play wins more gammons.  It turns out it wins the same number of plain games too.

4 comments:

  1. Good blog, Bill...I should visit here more often.

    In your third position, I think you misread the rollout, because the computer doesn't want to anchor. By the way, splotting is rarely correct when you have no new home-board points and the opponent has made her 5pt, unless your position is so desperate that counter-priming is your only realistic game plan.

    ReplyDelete
  2. Thanks Tim. Yep, looks like I misread the third one. Even so, the "splot" is a bad idea.

    ReplyDelete
  3. In the first position by running you are duplicating opponent's 2's. Every 2 except 32 plays well and makes a point so duplication is why the run is correct. More so if it was 54D-64 as now 32 makes the bar point. comment by Grant Hoffman

    ReplyDelete
  4. In the above comment I meant 43D-64 not 54D-64 for 32 becoming good.

    ReplyDelete