Study Group 41-50

I participated in Julius High's backgammon study group this weekend.  Julius picks 10 problems, sends them ahead of time so we can review (without the computer of course) and then he leads a live discussion of the positions.  Anyone who wants to offer their play and reasoning is free to do so, then Julius proceeds to give the correct answer and analysis.

While Julius is being very generous with his time I do enjoy this, I would hesitate to call this a study group.  It's more like a quiz or a free lesson led by Julius.

What follows are the problems discussed along with answers provided by Julius and my own further comments.

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

#41 A-C: 61P-22

#41A  - 61P-22 at 6 away 6 away
#41B  - 61P-22 at 6 away 2 away
#41C -  61P-22 at 2 away 6 away

is Player 2 score: 0 pip: 160

6 point match

pip: 167 score: 0 is Player 1

XGID=-b----E-C---dE---bbe----B-:0:0:1:22:0:0:0:6:10

to play 22

1.	XG Roller++	13/11(2) 6/4(2)	eq: +0.146
	Player: Opponent: 53.04% (G:15.81% B:0.74%) 46.96% (G:13.16% B:0.63%)
2.	XG Roller++	24/22(2) 6/4(2)	eq: +0.141 (-0.005)
	Player: Opponent: 52.57% (G:13.68% B:0.49%) 47.43% (G:10.27% B:0.35%)
3.	XG Roller++	24/22 13/11 6/4(2)	eq: +0.107 (-0.039)
	Player: Opponent: 52.49% (G:14.11% B:0.61%) 47.51% (G:12.91% B:0.51%)

eXtreme Gammon Version: 2.10, MET: Kazaross XG2

First off I have written about reply rolls before.  See here for replies to 61P.

http://bigmoneybill.blogspot.com/2012/02/replies-to-61p.html

13/11(2), 6/4(2) is the most common way to play double twos.  It unstacks both heavy points and forms the basis for a strong offensive game.  You have both priming potential and attacking potential.

The exceptions are generally when you can hit a blot profitably elsewhere (obvious) and also if you in priming danger than the more balanced play 24/22(2), 6/4(2) is preferred.

Here 61P forms the beginning of a prime but your prime is better and his is nothing to panic over.  The six point is still stacky.  Basically your prime and offense is superior so just pay to that.

So in "A" where the score is even the play for money 13/11(2), 6/4(2) is best.  Note the evaluation above has the plays as tied but I don't believe it.  XG for money gives 13/11(2), 6/4(2) as best by .01.  The balanced play 24/22(2), 6/4(2) is reasonable since opponent is priming you but as already stated you have the start of a better priming game already so why not just pursue that?

"B" shows a score where you are significantly trailing.  The key concept here is that as the trailer you should steer the game towards more gammons.  Gammons win you 4 points and are more valuable for you than the opponent.  Thus you must play offensively here 13/11(2), 6/4(2).  I am not posting the evaluation here but it's about .06 error at this score compared to a virtual toss up in "A."

"C" on the other hand shows you with a big lead (2 away 6 away).  Conceptually you are playing "gammon save" mode here.  Your first priority is to not get gammoned and therefore 24/20(2) is best although 24/22(2), 6/4(2) is very close and essentially tied.  I actually like the balanced 24/22(2), 6/4(2) best as this makes offensive progress and cuts down the gammons a bit.  I can see how at a more extreme score like 2 away 4 away where the opponent's gammon value is a perfect 1.0 that 24/20(2) would then be the clear winner.

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

#42.  54 to play from the air.  Anchor or hit?

is Player 2 score: 0 pip: 147

7 point match

pip: 163 score: 0 is Player 1

XGID=-b----E-C--AeD---a-c-Ab-bA:0:0:1:54:0:0:0:7:10

to play 54

1.	XG Roller++	Bar/20 21/17*	eq: +0.170
	Player: Opponent: 55.58% (G:14.02% B:0.76%) 44.42% (G:15.39% B:0.42%)
2.	XG Roller++	Bar/21 13/8	eq: +0.123 (-0.046)
	Player: Opponent: 52.79% (G:13.60% B:0.49%) 47.21% (G:11.39% B:0.25%)
3.	3-ply	Bar/21 11/6	eq: +0.016 (-0.153)
	Player: Opponent: 50.16% (G:12.26% B:0.46%) 49.84% (G:11.49% B:0.24%)

eXtreme Gammon Version: 2.10, MET: Kazaross XG2

The player at the board chose to anchor with Bar/21 and 13/8 and I can certainly understand why.  When being blitzed our first priority is to survive and by making the anchor we have just accomplished that.  Hitting is counter-intuitive for that reason and I would also add scary!

Yet the hitting play Bar/20, 21/17* is correct.  Julius notes that

• hitting in the outfield reduces the checkers in the offensive zone to 7.  
• takes away half the opponent's roll
• catches up in the race (you are up 9 pips after hitting)

The phantom fear factor is the most important consideration here in my opinion.  When being blitzed we want the security of the anchor and is hard psychologically to give that up immediately after attaining it!  As noted above the opponent will only have seven men in the zone after the hit and  this is very important.  It will be difficult for opponent to work you over with this limited material to work with.  Another consideration not mentioned above is that you are not favored to be hit back.  In other words, most of the time the opponent misses and even when he does hit you are favored to bounce right back in from the bar, so the risk is noticeably muted.  On the other hand by sending a third man back you gain big in the race.  Finally, note that as opponent has ace point made and you don't you have priming potential while your opponent does not.

***********

#43  54 from the air.  Anchor or hit?

is Player 2 score: 5 pip: 154

11 point match

pip: 161 score: 3 is Player 1

XGID=-b----E-D--AdC--aaad--b-AA:0:0:1:54:3:5:0:11:10

to play 54

1.	XG Roller++	Bar/20 24/20	eq: +0.270
	Player: Opponent: 54.50% (G:13.18% B:0.48%) 45.50% (G:8.80% B:0.28%)
2.	XG Roller++	Bar/16*	eq: +0.229 (-0.041)
	Player: Opponent: 53.82% (G:15.84% B:1.12%) 46.18% (G:14.90% B:0.67%)
3.	XG Roller++	Bar/21 13/8	eq: -0.346 (-0.616)
	Player: Opponent: 41.84% (G:11.05% B:0.62%) 58.16% (G:23.21% B:1.20%)

eXtreme Gammon Version: 2.10, MET: Kazaross XG2

The player wrongly chose to hit with Bar/16.*  Contrasting with the prior problem, this time it is correct to anchor with Bar/24, 20/24.

According to Julius you should look to get out of a blot hitting contest when you have obtained an advantage.

• Obtaining an advanced anchor that creates dilly builders is a great positional asset
• Obtaining a defensive anchor when you are down in boardage is generally a good idea
• Obtaining a racing lead or even game is desirable 

While I don't necessarily disagree with the above comments, the same basic reasoning applies to the previous problem yet hitting was right there.  So I think there is more to the story.

As always backgammon is about weighing risk and reward.  First off when you anchor the race is already close so you don't gain much by hitting, yet you risk a lot.  The opponent has more heat on you in this position (8 in zone after the hit rather than 7) and has both blitzing and priming potential.

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

#44.  36 to play from the bar.  Hit, anchor or something else?

is Player 2 score: 1 pip: 162

7 point match

pip: 171 score: 3 is Player 1

XGID=-a---aE-C--AeC---cAda---AA:0:0:1:36:3:1:0:7:10

to play 36

1.	XG Roller++	Bar/22 24/18	eq: -0.168
	Player: Opponent: 46.82% (G:9.66% B:0.31%) 53.18% (G:13.15% B:0.43%)
2.	XG Roller++	Bar/22 11/5*	eq: -0.206 (-0.038)
	Player: Opponent: 46.31% (G:12.03% B:0.55%) 53.69% (G:15.78% B:0.88%)
3.	XG Roller++	Bar/16	eq: -0.501 (-0.333)
	Player: Opponent: 41.05% (G:9.18% B:0.37%) 58.95% (G:20.87% B:1.42%)

eXtreme Gammon Version: 2.10, MET: Kazaross XG2

I got this one wrong. My play was Bar/22, 11/5* hitting. My own rule of thumb for these positions is to always hit when there is only a one point board. I suppose I will have to throw away this rule now.

The correct play was Bar/22, 24/18.

Julius:

• When you have more backmen than your opponent (3-2) you should try to emphasize defense by making an advanced anchor
• You should try to avoid having a blot hitting contest on your opponent's side of the board
• Obtaining less blots with even boardage is an advantage

I'm not sure I agree with the first point.  After all, one fundamental of Magriel is that you can take more risks when you are behind and have an anchor.  After all, going from bad game to worse game isn't as bad as going from good game to bad game.

OK so I am quibbling.  Here are my reasons for what they are worth:

• You are actually outboarded (1.5 to 1)
• Hitting diversifies opponent's numbers and is unlikely to work.  Observe most 4's 5's and 6's hit you back whilst 1's and 3's make the offensive five point.
• Making the anchor is good no matter what the opponent rolls.  I think this is the key.  You accomplish something for sure by making the anchor while hitting is unlikely to work.  So you are risking too much to bypass a for sore benefit in making the anchor.
• Also, as Julius notes, hitting on wrong side of board (exchange of hits costs you pips)
• Hitting with wrong man (you would prefer to hit off the six point to put those guys to work)

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

#45.  42 to play from the air.  Lost of options here.  what is your choice?

is Player 2 score: 3 pip: 116

11 point match

pip: 118 score: 5 is Player 1

XGID=--aBABCBB---AAb-ab-eaba--A:1:-1:1:42:5:3:0:11:10

to play 42

1.	XG Roller++	Bar/23 8/4	eq: +0.246
	Player: Opponent: 65.27% (G:24.85% B:1.64%) 34.73% (G:12.09% B:0.82%)
2.	XG Roller++	Bar/23 6/2*	eq: +0.030 (-0.217)
	Player: Opponent: 58.81% (G:28.36% B:2.58%) 41.19% (G:19.86% B:2.31%)
3.	XG Roller++	Bar/23 12/8	eq: -0.001 (-0.248)
	Player: Opponent: 61.06% (G:19.55% B:1.46%) 38.94% (G:17.55% B:0.80%)

eXtreme Gammon Version: 2.10, MET: Kazaross XG2

Julius stated he made the costly error Bar/23, 12/8 here.

With low confidence, I actually came up with the right play Bar/23, 8/4.

Julius notes: the best play makes a solid 5-prime and increases boardage to 4-2

• A solid prime is stronger than a prime with a gap
• an innerboard point is more valuable than an outerboard point
• When both players have one back man, a blot hitting contest may ensue, therefore, you should try to make the strongest inner board

My reasoning had more to do with looking at how the opponent's rolls play.

If you hit with Bar/23, 6/2* than basically all 2's and 4's are good for the opponent.  note that every 2 and every 4 will hit and cover at least blot.  Opponent would be quite happy with this.

On the other hand, by entering and making the board you place immediate pressure on opponent to roll an immediate six.  While it's true that every six either hits and covers a blot or hits two checkers by my count this is less good rolls.

As Julius notes you are making a strong board and have a slotted 5.5 prime.  In an exchange of hits you have the better board and the better prime.  Any subsequent hit by you rates to be very strong, a near winner by itself.

Note than 8/4 is always constructive as well while the hit is only sometimes constructive.

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

#46.  Later in same game as prior problem.  54 to play Escape or Make the point?

is Player 2 score: 3 pip: 118

11 point match

pip: 127 score: 5 is Player 1

XGID=b--B-BCBCA-------b-dbcbB--:1:-1:1:45:5:3:0:11:10

to play 45

1.	XG Roller++	23/18 9/5	eq: +0.450
	Player: Opponent: 73.20% (G:20.72% B:0.62%) 26.80% (G:7.55% B:0.31%)
2.	XG Roller++	9/4 8/4	eq: +0.424 (-0.026)
	Player: Opponent: 71.39% (G:23.00% B:0.61%) 28.61% (G:6.64% B:0.35%)
3.	4-ply	23/14	eq: +0.357 (-0.093)
	Player: Opponent: 70.87% (G:19.33% B:0.58%) 29.13% (G:8.64% B:0.42%)

eXtreme Gammon Version: 2.10, MET: Kazaross XG2

Juluis made the 3 point with 9/4, 8/4 - possibly on autopilot. Even though you need to escape badly you also need the 4 point badly. A tough decision. I may have made the 4 point as well and hoped to roll another 5 later.

Better was escaping with 23/18, 9/5. This run dominates 23/14 since it diversifies your numbers. Moving 9/5 gives you aces to make the four point (and 3's to make the 2 point) while 5's let you escape.

By contrast 23/14 leaves you with 5's to both escape and also to make the 4 point.

Julius explains:

• When you have an anchor with 5 blockers you should be looking for opportunities to escape.
• An ideal time to escape is when your opponent is on the bar
• The correct move is 23/18, 9/5. With 2 checkers on the bar, and opponent having 5 blocking points, escaping with a checker is correct. In addition this secures the 9 point builder and gives you a good ace to play on the next roll. With two checkers on the bar, you have a little more time to attack.

Along the same lines, I would add the most important thing in the position is escaping.  Do the most difficult thing to do first.  It's easier to make the point than escape and escaping is the priority so escape.  As noted above, 9/4 is technically better due to diversification of your rolls.

Finally Stick's rule "always move the back checker" in late stages of a blitz/scramble may apply here as well.

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

#47.  42 to play.  Inner point, Outer point or something else?

is Player 2 score: 0 pip: 141

5 point match

pip: 154 score: 0 is Player 1

XGID=-a---aE-D--AeC-----d-Bb-b-:0:0:1:42:0:0:0:5:10

to play 42

1.	XG Roller++	8/4 6/4	eq: +0.175
	Player: Opponent: 54.02% (G:14.27% B:0.51%) 45.98% (G:11.03% B:0.22%)
2.	XG Roller++	13/9 11/9	eq: +0.148 (-0.027)
	Player: Opponent: 53.40% (G:12.16% B:0.39%) 46.60% (G:9.47% B:0.19%)
3.	2-ply	11/5*	eq: -0.008 (-0.183)
	Player: Opponent: 49.80% (G:12.93% B:0.45%) 50.20% (G:13.29% B:0.26%)

eXtreme Gammon Version: 2.10, MET: Kazaross XG2

The player over the board went for the outer point 13/9, 11/9.

Best was making the inner point with 8/4, 6/4.

According to Julius:

• When you have an advanced anchor, you should be looking for opportunities to improve your offense.
• Innerboard points are better than outer points.
• When you see five or more checkers on a single point, you should be looking for opportunities to unstack.

For my eyes, the stacks on the 8 point and 6 six point are just so ugly it's too hard to pass up the chance to unstack and make an innerpoint in the process.

Note that even though you are leaving a direct six shot (16 numbers or all but 33 hit you) the value of the inner point is more than the outerpoint which leaves no hits.

This problem highlights the importance of having a nice structure. 

Also note that even though you get hit a lot, getting hit isn't so bad.  You are anchored and opponent only has a 3 point board.  Most likely you reenter and play from there. 

The opponent has no points in his outer board (7 to 12 points) so he will have great difficulty rolling this position home.  You rate to get shot(s) later in the game by which time you should have a nice board waiting.

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

#48.  You are on roll.  Cube action?

is Player 2 score: 1 pip: 146

7 point match

pip: 133 score: 0 is Player 1

XGID=a---CBC-B--AeB---c-dBb----:0:0:1:00:0:1:0:7:10

on roll, cube action?

Analyzed in XG Roller++	No double	Double/Take
Player Winning Chances:	73.47% (G:34.41% B:0.31%)	74.11% (G:34.64% B:0.36%)
Opponent Winning Chances:	26.53% (G:3.72% B:0.13%)	25.89% (G:3.93% B:0.14%)
Cubeless Equities	+0.790	+1.674
Cubeful Equities
No double:	+1.037
Double/Take:	+1.504 (+0.467)
Double/Pass:	+1.000 (-0.037)
Best Cube action: Too good to double / Pass
Percentage of wrong take needed to make the double decision right: 7.4%

eXtreme Gammon Version: 2.10, MET: Kazaross XG2

This one I got wrong. My guess was Double / Pass.

It looks like opponent is down in all phases of the game (behind in race, outboarded, and on the bar perhaps against 4 point board net time). Looks pretty grim - probably pass.


As you can see the position is technically too good which surprises me. What may have tricked me is the fact that it is only a 3 point board. I could see myself thinking I'm only a bit behind in boardage and race so the game isn't hopeless. Also, it's only a 3 point board so i will probably come in next time and have some coutnerplay against the 5 point anchor. So I wasn't positive it was even a drop.

I am unclear as to what exactly Julius is doing in his analysis. He adds Race advantage of 1 + Boardage advantage of 1 (3-2) + Threats of 1 (he counts 17 good numbers) + 1/2 for 11 in the attack zone + 1/2 for having a defensive anchor + 1/2 for gammons to get a total of 4.5 I don't know what this 4.5 means.

He further writes "in general, when your opponent has 1 backman he cannot anchor, so you should continue to attack checker and play for a gammon. If opponent escpaes last backman, then you can consider doubling."

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

#49. You are on roll. cube action?

is Player 2 score: 2 pip: 141

11 point match

pip: 130 score: 0 is Player 1

XGID=aaBBB-Ca----aDa--bbb-Bb-b-:0:0:1:00:0:2:0:11:10

on roll, cube action?

Analyzed in XG Roller++	No double	Double/Take
Player Winning Chances:	83.08% (G:59.95% B:13.25%)	83.06% (G:60.20% B:13.35%)
Opponent Winning Chances:	16.92% (G:3.27% B:0.13%)	16.94% (G:3.35% B:0.15%)
Cubeless Equities	+1.376	+2.879
Cubeful Equities
No double:	+1.487
Double/Take:	+2.770 (+1.283)
Double/Pass:	+1.000 (-0.487)
Best Cube action: Too good to double / Pass
Percentage of wrong take needed to make the double decision right: 27.5%

eXtreme Gammon Version: 2.10, MET: Kazaross XG2

Apparently the player cashed this game when it is obviously too good.

Julius -- "when your opponent has 5 blots and no anchor, you are way too good to double."

Agreed. One way to evaluate whether you are too good or not is to consider what has to happen on the next exchange for your opponent to have a take. I can't really envision anyhting take would give a take. Perhaps if you fail to hit a second checker and then opponent makes a good anchor somewhere he can take. Clearly in this position this is a remote possiblility.

Another strategy is put yourself in your opponent's shoes and see what you would do with a cube. If it's an instadrop then you need to really think twice about doubling as you may be too good.


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

#50. You are on roll. Cube action?

is Player 2 score: 1 pip: 149

7 point match

pip: 151 score: 1 is Player 1

XGID=ab-a-CCB----aE--aA-c-bAbb-:0:0:1:00:1:1:0:7:10

on roll, cube action?

Analyzed in XG Roller++	No double	Double/Take
Player Winning Chances:	81.91% (G:41.86% B:5.69%)	82.26% (G:42.19% B:5.74%)
Opponent Winning Chances:	18.09% (G:4.71% B:0.13%)	17.74% (G:4.47% B:0.14%)
Cubeless Equities	+1.124	+2.448
Cubeful Equities
No double:	+1.267
Double/Take:	+2.286 (+1.019)
Double/Pass:	+1.000 (-0.267)
Best Cube action: Too good to double / Pass
Percentage of wrong take needed to make the double decision right: 20.8%

eXtreme Gammon Version: 2.10, MET: Kazaross XG2

This is another one that looked obvious to me as Too Good. The player mistakenly cashed the game.

Perhaps this is a case of fearing a backgame. Here, though, the opponent doesn't yet have a backgame and even if he makes two anchors the timing is beyond hopeless.

As Julius notes, the opponent simply has too many loose blots to take. And "when you have the potential to trap more than 2 blots, you should play on for the gammon."