Search This Blog

December 29, 2011

Resolutions

Lately I've been rolling bad, playing bad and losing my ass.  Riding this streak out is tough, and it is really tempting my resolve.  It is tempting to just say the hell with it and forget this stupid game.  If I do decide to continue, I suppose I need to persevere and redouble my improvement efforts.

To that end, here are some backgammon resolutions for 2012.

1. SLOW DOWN.  Take at least 10-15 seconds on all cube decisions.  And consider at least one alternative move before making a checker play.  Goal is to avoid snap judgments.  Consider the whole board, and hopefully game plan as well.

2. BE MORE SERIOUS WITH LIVE PLAY.  Look at these matches as an opportunity to learn, rather than just killing time before the chouette begins!  Goal is to take pictures of citical decisoins or record some matches for analysis later.  And to slow down as above.

3. STUDY STUDY STUDY!!  I feel like my game could use improvement in all areas, so it really doen't matter too much what I study so long as I study something!  Some ideas.  Racing, Match Equity, Opening Replies, Early Game checker play, Blitzing, Backgames, Holding Games, Pay Now/Pay Later, etc.  Also re-study the foundational materials such as Robertie, Magriel, Trice.  Take lessons as opportunities arise.

4. BE MORE SERIOUS WITH ONLINE PLAY.  Mostly the same points as 1 and 2 above.  Slow down and be serious.  Additionally, I want to use Stick's method of keeping a working folder of blunders.  Meaning, isolate blunders from matches and really look at them closely to try and eliminate the error in thinking from my game.  Continue to share errors in the blog.

December 28, 2011

Lamford on Racing

This is from Lamford's Starting Out in Backgammon and complements the EPC method for racing.

It covers endgames when each side is down to about 3 checkers or less.

So, my goal would be to use the following guidelines for racing...

1. Endgames (3 checkers or less with misses) -- Lamford. 

2. Pips versus Rolls and short races (~30 to 50 pips) --  EPC and Wastage (or "Stick's lecture)

3. Longer races (50 pips and up or short races with minimal wastage) -- Trice Rule 62 or Ballard 57 method


Endgame Race Cubes, Rules of Thumb:

These rules apply to positions with three checkers each (or less), and where each side has misses.

No distinction between Double and Redouble.  If note says Double, then Redouble also.

1.       One Checker against One Checker (or more):

a.       Always Double

b.      Always Pass (technically you are indifferent with checker on 6 point, since 9 rolls fail to get off)

2.       Two Checkers against. One Checker:

a.       Always Double if favorite to bear off in one roll

b.      Take if 9 or more misses (25%)

3.       Two Checkers against Two checkers:

a.       Double when race is Minus 2 pips or better

b.      Mimimum needed to take

                                                               i.      Less than 27 immediate winning rolls and

                                                             ii.      Pip Count at least equal

c.       *Exceptions: slightly better to pass in these mirror positions (two checkers each on three, four, or five point)

4.       Two Checkers against Three Checkers:

a.       Always Double

b.      Almost Always Pass

c.       *Exceptions: Two checkers on 6-point may be takeable with some three checker configurations

5.       Three Checkers against Two Checkers:

a.       Minimum Needed to Double

                                                               i.      Opponent not favored to get off next turn and

                                                             ii.      Race is Minus 2 pips or better and

                                                            iii.      Your pips are 12 or less

b.      Minimum Need to Take

                                                               i.      Pip count at least even and

                                                             ii.      Pip count 8 or less

6.       Three Checkers against Three Checkers:

a.       Player on roll has 10 or less pips

                                                               i.      Always Double

                                                             ii.      Always Pass

b.      Player on roll has 11-12 pips

                                                               i.      Always Double

                                                             ii.      Take if pip count even or better

c.       Player on roll has 13 pips or more

                                                               i.      Double if pip count even or better

                                                             ii.      Take if pip count minus one or better

December 25, 2011

Racing/EPC and Prop 57 Practice

Stick's racing lecture mostly covered Trice's concept of EPC. He went through countless examples of how to estimate wastage for various positions. Only at the end did he introduce the racing formulas to complement the EPC method.

In my opinion, this method is more work than it's worth for longer races (when no one has borne off yet) and when both sides have little wastage. The method seems most appropriate where you have Pips versus Rolls. Then you convert both sides to EPC and apply the racing rules from there.

EPC is widely known and can be found in Trice's excellent book Backgammon Boot Camp. The online article at bkgm.com (http://www.bkgm.com/articles/EffectivePipCount/) also covers most of the concepts.

Wastage Basics

For a Pips position: EPC = Pip Count + estimated Wastage

1. Nice positions waste 7 pips  (the 7-5-3 position).  Note pretty much any configuration where all checkers are on 4-5-6 points are 7 pips wastage.
1a. Nice with 7-8 checkers remaining is 6.5 pips wastage
1b. Nice with 5-6 checkers remaining is 6 pips wastage
2. Flat positions waste 10 pips (2 each on 1pt thru 6 pt)
3. "1/2 flat" wastes 6 pips (1 each on 1pt thru 6 pt)
4. Single checker on 4 pt wastes 4 pips
5. Single checker on 10 pt wastes 5 pips

For a Rolls position: EPC = 7N + 1

1. Additions to pip count are necessary when misses are possible.  (ie when not all checkers are on ace and deuce point).  Use table below to estimate additional wastage to 7N + 1.  

Even and Odd columns represent the average pip count per checker when there are Even or Odd number of checkers remaining to bear-off.

So if you had even number of checkers left and average pip was 2.6, then you would add 3 pips to the 7N + 1 calculation.
 Add       Even   Odd  
1 2.0 2.6
2 2.3 2.9
3 2.6 3.2
4 2.8 3.4
5 3.0 3.6

2. Stack & Straggler (1 straggler).  EPC = 3.5c + p

3. Stack & Straggler (2 stragglers).  EPC = 3.5c + p1 + p2 - 4

where ...

c is number of checkers remaining
p is pip count of the straggler


Application / EPC Racing Formula

Point of Last Take = (# of Rolls to go) - 3

Note: # of Rolls to go can be approximated as EPC / 7 (round down)

Double within 2 pips of last take
Redouble within 1 pip of last take


*** This breaks down when you get to 3 rolls or less ***
*** So you need to use other methods when EPC in low 20's or about 4 checkers per side ***

Practice / Prop 57

Stick recommended using Trice's Prop 57 as practice estimating wastage and using the EPC racing formula.

Here are a few sample positions taken from practice versus XG .

is XG Roller+

score: 0
pip: 19
Unlimited Game
Jacoby Beaver
pip: 38
score: 0

is Player 1
XGID=-----DC---------------bde-:0:0:-1:00:0:0:3:0:10
on roll, cube action?

Analyzed in XG Roller+ No double Double/Take
  Player Winning Chances: 72.27% (G:0.00% B:0.00%) 72.17% (G:0.00% B:0.00%)
  Opponent Winning Chances: 27.73% (G:0.00% B:0.00%) 27.83% (G:0.00% B:0.00%)
  Cubeless Equities +0.445 +0.887
Cubeful Equities
     No double:+0.714 (-0.072)
 Double/Take:+0.786
     Double/Pass:+1.000 (+0.214)
Best Cube action: Double / Take
eXtreme Gammon Version: 2.01


So XG's EPC is 6 roll position or 43.
Last take is 43/6 -3 or 3 pips.

So I need to have EPC of 46 or less to take.

I have 38 + wastage. Wastage for nice position with 7-8 checkers left is 6.5. I have a gap on the 4 pt which must be a bit worse than "nice" but I am at 44.5 so far so must be easy take. Turns out the penalty is only .2 so the take is "super-easy" but how easy would it be to estimate this without the practice?

is XG Roller+

score: 0
pip: 26
Unlimited Game
Jacoby Beaver
pip: 41
score: 0

is Player 1
XGID=--A-CCB---------------ded-:0:0:1:00:0:0:3:0:10
on roll, cube action?

Analyzed in XG Roller+ No double Double/Take
  Player Winning Chances: 72.63% (G:0.00% B:0.00%) 72.64% (G:0.00% B:0.00%)
  Opponent Winning Chances: 27.37% (G:0.00% B:0.00%) 27.36% (G:0.00% B:0.00%)
  Cubeless Equities +0.453 +0.906
Cubeful Equities
     No double:+0.682 (-0.083)
 Double/Take:+0.765
     Double/Pass:+1.000 (+0.235)
Best Cube action: Double / Take
eXtreme Gammon Version: 2.01

I missed the cube here. 

I have 41 + wastage of about 7 for 48 EPC.  48/7 - 3 is 3 so last take is 51 for XG.

I show XG with 7 roll position or EPC of 50.

I got confused on interpreting the cube action.  Since last take is 51 I should double if XG has 49 or more... ie, within 2 pips of last take.  XG gas 50 so proper action is Double/Take as the analysis shows.

Clearly I need more practice, especially since I was doing this as an "open notes" exercise.



is XG Roller+

score: 0
pip: 40
Unlimited Game
Jacoby Beaver
pip: 25
score: 0

is Player 1
XGID=-EDD---------------bdb----:0:0:-1:00:0:0:3:0:10
on roll, cube action?

Analyzed in XG Roller+ No double Double/Take
  Player Winning Chances: 76.57% (G:0.00% B:0.00%) 76.58% (G:0.00% B:0.00%)
  Opponent Winning Chances: 23.43% (G:0.00% B:0.00%) 23.42% (G:0.00% B:0.00%)
  Cubeless Equities +0.531 +1.063
Cubeful Equities
     No double:+0.782 (-0.165)
 Double/Take:+0.947
     Double/Pass:+1.000 (+0.053)
Best Cube action: Double / Take
eXtreme Gammon Version: 2.01

This is basically the same position but now I am on the Take/Drop side.

XG has 40 + 7 wastage or EPC of 47.
Last Take is 47 + (47/7) - 3 or 50.

I have 7 roll position or EPC of 50.

So proper cube action is Double/Take.

is XG Roller+

score: 0
pip: 30
Unlimited Game
Jacoby Beaver
pip: 14
score: 0

is Player 1
XGID=-DE----------------bbb----:0:0:-1:00:0:0:3:0:10
on roll, cube action?

Analyzed in XG Roller+ No double Double/Take
  Player Winning Chances: 67.21% (G:0.00% B:0.00%) 67.21% (G:0.00% B:0.00%)
  Opponent Winning Chances: 32.79% (G:0.00% B:0.00%) 32.79% (G:0.00% B:0.00%)
  Cubeless Equities +0.344 +0.688
Cubeful Equities
     No double:+0.523 (-0.014)
 Double/Take:+0.537
     Double/Pass:+1.000 (+0.463)
Best Cube action: Double / Take
eXtreme Gammon Version: 2.01

Here is another example from the Rolls side.

XG is just barely on the front edge of the doubling window.

XG has 30 + 6 wastage.
Last Take is 36/7 -3 or 2.
So Trailer last take is 38 pips. 

I have 5 roll position or 36 so the take is very comfortable.
The Double is also correct since the window opens at 38- 2 pips or 36.

December 17, 2011

Chutzpah-SixPack

OK.  Time to get the ball rolling by looking at mistakes from my matches.

These are from a match against SixPack.

First game of 7 point match.  Cube Action?

is Chutzpah

score: 0
pip: 146
7 point match
pip: 121
score: 0

is SixPack
XGID=--a--BDaC---bE-a-a-cb-bAb-:0:0:-1:00:0:0:0:7:10
on roll, cube action?

Analyzed in Rollout No double Double/Take
  Player Winning Chances: 62.88% (G:31.43% B:0.19%) 63.07% (G:32.49% B:0.19%)
  Opponent Winning Chances: 37.12% (G:8.69% B:0.51%) 36.93% (G:8.78% B:0.59%)
  Cubeless Equities +0.485 +1.032
Cubeful Equities
     No double:+0.619 (-0.122)±0.011 (+0.608..+0.630)
 Double/Take:+0.741±0.016 (+0.725..+0.758)
     Double/Pass:+1.000 (+0.259)
Best Cube action: Double / Take
Rollout details
1296 Games rolled with Variance Reduction.
Moves: 3-ply, cube decisions: XG Roller
Double Decision confidence:100.0%
Take Decision confidence:100.0%
Duration: 6 minutes 50 seconds
eXtreme Gammon Version: 2.01, MET: Kazaross XG2
He had been Too Good the past couple rolls.  I rolled 2-3 from the bar, coming in and covering the 5-point.  My mindset was that a joker roll would merely save the gammon.  I dropped in a flash.

Still, looking at the position now I don't really understand why it is a take. I could see myself making this error again.

I am outboarded, down in the race, and may be attacked next roll.  I suppose these are the strengths of my position:  I have the 5-point, always big.  No other loose checkers.  My back-guy is relatively deep on the 2-point.  Probably mostimportant are the defects in Sixpack's position.  He does have four loose blots after all.  One hit could turn the game around, and he can't do everything in one roll.  Also the guys on the ace-point aren't the best.

*************
In the next position, I lead 4 away post-crawford.  I have 44 to play.

is SixPack

score: 3
pip: 122
7 point match
pip: 154
score: 6

is Chutzpah
XGID=--AB-BBbB-a-bC-----bccbC--:1:1:1:44:6:3:0:7:10
to play 44

1.XG Roller++13/5 8/4(2)eq: -0.785
Player:
Opponent:
30.78% (G:7.59% B:0.22%)
69.22% (G:19.10% B:0.80%)
2.XG Roller++13/9(2) 13/5eq: -0.839 (-0.054)
Player:
Opponent:
23.88% (G:4.02% B:0.12%)
76.12% (G:15.10% B:0.55%)
3.XG Roller++13/9(3) 6/2eq: -0.864 (-0.079)
Player:
Opponent:
23.34% (G:3.88% B:0.13%)
76.66% (G:15.74% B:0.59%)
eXtreme Gammon Version: 2.01, MET: Kazaross XG2
My mindset was "gammon save."  Therefore, I figured I should get as close to home as possible and went for 13/9(2), 13/5.

And while it is true the gammon losses are fewer, the problem with this play is I basically give up on trying to win.  Making the 4-point is obviously a big improvement.  I establish a 4.5 point board which will be very strong should I hit a shot later.  True enough, I will have to clear the mid-point soon and could get another blot sent back.  But then again, this may not happen.  And even if it does, I still have the 2-point anchor - which while not the greatest still provides some gammon insurance.

In short, I let gammonphobia clould my judgment here.

*********

Later in the same game, I made the same mistake.  51 to play.

is SixPack

score: 3
pip: 70
7 point match
pip: 90
score: 6

is Chutzpah
XGID=-BCBBBB--------b---bbebBb-:1:1:1:51:6:3:0:7:10
to play 51

1.XG Roller+23/17eq: -0.772
Player:
Opponent:
20.33% (G:0.75% B:0.01%)
79.67% (G:8.52% B:0.18%)
2.XG Roller+6/1 2/1eq: -0.865 (-0.093)
Player:
Opponent:
12.90% (G:0.11% B:0.00%)
87.10% (G:5.86% B:0.14%)
3.3-ply6/5 6/1eq: -0.853 (-0.081)
Player:
Opponent:
13.64% (G:0.20% B:0.00%)
86.36% (G:6.02% B:0.12%)
eXtreme Gammon Version: 2.01, MET: Kazaross XG2
Without even thinking, I lifted the six point.  Gammonphobia again.  His awkward builder distro and the possibility of duplicating 2's gives me a golden opportunity to make a break for it. 

I do get gammoned a bit more often but give myself more than enough chances to win to compensate. 

******

Finally, a careless error at DMP.  21 to play. 

is SixPack

score: 6
pip: 123
                         
7 point match
                          pip: 61
score: 6

is Chutzpah
XGID=bBBBBCBB----b----b-b-bbba-:0:0:1:21:6:6:0:7:10
to play 21

1.XG Roller+6/5 6/4eq: +0.965
Player:
Opponent:
98.27% (G:63.91% B:0.31%)
1.73% (G:0.00% B:0.00%)
2.XG Roller+7/6 7/5eq: +0.926 (-0.039)
Player:
Opponent:
96.32% (G:78.81% B:0.77%)
3.68% (G:0.00% B:0.00%)
3.4-ply7/6 5/3eq: +0.954 (-0.011)
Player:
Opponent:
97.70% (G:79.80% B:0.65%)
2.30% (G:0.00% B:0.00%)

eXtreme Gammon Version: 2.01, MET: Kazaross XG2


The only explanation for this is that I was playing on autopilot.  I just made the first play I saw without even thinking.

I know better than this, especially since I just reviewed DMP strategy in recent posts. This is the one-time when pre-clearing is right. All you care about is getting the mofo out of your hair. So open up right away and let the coming in begin.