Search This Blog

February 27, 2014

Baileys 2-20-14 v JQ

Below are selected errors from a live club match against John Quinn. Matches are to 5 points in the losers bracket.

I lead 3 away 5 away and have 53 to play. But don't let the score fool you. The correct play is same as for money.


is JQ

score: 0
pip: 157
5 point match
pip: 161
score: 2

is Bill
XGID=-b--B-D-B---eE---b-d---bB-:0:0:1:53:2:0:0:5:10
to play 53

1.XG Roller++24/21 13/8eq: +0.134
Player:
Opponent:
55.90% (G:17.27% B:0.92%)
44.10% (G:12.85% B:0.46%)
2.XG Roller++8/3 6/3eq: +0.111 (-0.023)
Player:
Opponent:
54.83% (G:19.32% B:0.92%)
45.17% (G:13.47% B:0.77%)
3.XG Roller+13/10 13/8eq: +0.102 (-0.032)
Player:
Opponent:
54.85% (G:18.02% B:0.91%)
45.15% (G:13.08% B:0.65%)
eXtreme Gammon Version: 2.10, MET: Kazaross XG2


I made the 3 point with 8/3, 6/3.

The computer likes up and down 24/21, 13/8.

This is the type of position that should be fundamental. Unfortunately I am struggling to grasp the essence of what's going on.

I "know" that when opponent makes the 2 point he has an attacking formation and I should therefore be reluctant to split into the attack. Therefore when I can do something constructive like make an inner board point this seems like an obvious play.

I must be overemphasizing this point as correct is to split now. Why? Some guesses:

  • The 5 plays well. It unstacks the midpoint while reinforcing the stripped 8 point and leaves no shots.

  • The 4 point is more of a priming formation than attacking formation so I should strive to fill in the bar or 5-point rather than the 3 point. By placing an extra man on the 8 point I give myself some numbers to perhaps make either the five or bar next time.

  • While I don't want to be blitzed, now is an ideal to split (?). He only has eight men in the zone so it is not too dangerous. Also, my man sits on the 4 point which is less ideal to attack than a man on his 5 point. (Note that with 52 or 53 split is correct while with 54 two down is correct. Hence my speculation about the 5 point being the spot he'd rather attack on.)

  • Splitting hopes to anchor next time, and also places some pressure on opponent's outer board. JQ can't simply pull down from the stacked midpoint without leaving a decent amount of indirects and possibly even a direct shot. So I am influencing all sectors of the board - a balanced play.

  • Finally splitting doesn't leave a blotted 8 point which could be hit by a seven from JQ.


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

Very next play.  How to use these double twos?

is JQ

score: 0
pip: 141
5 point match
pip: 153
score: 2

is Bill
XGID=---BBbC-A---cE--bb-d---bB-:0:0:1:22:2:0:0:5:10
to play 22

1.XG Roller++24/22(2) 13/11(2)eq: -0.177
Player:
Opponent:
46.88% (G:9.99% B:0.32%)
53.12% (G:11.61% B:0.36%)
2.XG Roller++24/20(2)eq: -0.216 (-0.038)
Player:
Opponent:
44.39% (G:8.00% B:0.22%)
55.61% (G:8.27% B:0.24%)
3.3-ply24/22(2) 8/4eq: -0.283 (-0.106)
Player:
Opponent:
42.91% (G:8.43% B:0.23%)
57.09% (G:10.58% B:0.31%)
eXtreme Gammon Version: 2.10, MET: Kazaross XG2



I rather quickly moved 24/20(2).  This makes the best forward anchor and lessens the gammons.  What could be wrong with that?

And yet 24/22(2), 13/11(2) is clearly correct.  I *think* these are the reasons:

  • Opponent has 2 point made.  The priming threat isn't that great since JQ can't use the men on his 2 point to build points in front of me.  So there is less need to advance all the way.

  • Also I am down in the race.  So I want more contact (stay back) and I want more contact (by making the 11 point to block him).

  • The blot on the 8 point isn't under much danger no matter what play I make.  JQ will be reluctant to break anchor to hit.  Not too relevant here as applies to both plays with 22 but I wanted to note it.

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

Later in the same game.  I still lead 3 away 5 away and have been doubled.  Take or Drop?

is JQ

score: 0
pip: 76
5 point match
pip: 88
score: 2

is Bill
XGID=-ABCCBB----------cbcBbbc--:0:0:-1:00:2:0:0:5:10
on roll, cube action?

Analyzed in XG Roller++ No double Double/Take
  Player Winning Chances: 78.47% (G:1.15% B:0.01%) 78.60% (G:1.14% B:0.01%)
  Opponent Winning Chances: 21.53% (G:0.41% B:0.00%) 21.40% (G:0.36% B:0.01%)
  Cubeless Equities +0.574 +1.008
Cubeful Equities
     No double:+0.867 (-0.047)
xg Double/Take:+0.914
     Double/Pass:+1.000 (+0.086)
Best Cube action: Double / Take
eXtreme Gammon Version: 2.10, MET: Kazaross XG2


I dropped. Down 12 pips I figure this would be a drop in a straight race and it's not obvious that the contact favors me. Barring boxes I am not getting a shot next time. I will leap out next roll and then JQ can point on me or pick and pass or leave me alone. This seems like unfavorable contact. Gammons are low but it feels like I don't win much from here. So I dropped.


This position is actually a take. And it's strictly score based. The position is indeed a small drop for money. But at this score the takepoint is about 19% so by the book I should take.

The real question is how to remember that at this score the takepoint is lower than normal? And why is that? Once again I don't know the answer. What follows is speculation.


If I drop the score is 3 away 4 away. This is one of those special scores where the opponent has trememdous cube leverage. His gammon value is almost 1.0, which leads to frequent fast double (as trailer) and fast passes (as the leader). The leader could easily get gammoned for the match. This is a score the leader would just assume avoid. Which is a long winded way of saying "what the hell, if I lose its 3 away 3 away which isn't too bad. 3 away 4 away isn't much of a picnic anyway."

Winning gets you to Crawford. Worth noting but this can't really be the reason. At least I don't think so. After all, winning also gets you to crawford from 3 away 4 away and also 3 away 6 away and 3 away 7 away where the respective takespoints are 21.5%, 22% and 23%. I suppose I should add that these are "Stick's takepoints" which assume less recube vig than the computer does. Stick's umbers don't quite conform to the computer match equity tables.

Another possibility comes from Mochy's 5 point match advice. Where he simply states the trailer at 4 away and 5 away wants to win 4 points while the leader at 2 away and 3 away wants to win 2 points. This position is exactly what the leader wants - a two point game. Maybe. Maybe. A con to this line of thinking is the same holds true for 3 away 6 away and 3 away 7 away yet the takepoints are higher. So once again I am at a loss.

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

I lead 3 away 4 away and am on roll. Roll or Double?

is JQ

score: 1
pip: 166
5 point match
pip: 134
score: 2

is Bill
XGID=----cBD-CB--cBb-Bb-ca-a---:0:0:1:00:2:1:0:5:10
on roll, cube action?

Analyzed in XG Roller++ No double Double/Take
  Player Winning Chances: 71.40% (G:10.97% B:0.28%) 71.67% (G:11.00% B:0.29%)
  Opponent Winning Chances: 28.60% (G:4.17% B:0.14%) 28.33% (G:4.18% B:0.15%)
  Cubeless Equities +0.572 +1.381
Cubeful Equities
     No double:+0.739 (-0.053)
xg Double/Take:+0.792
     Double/Pass:+1.000 (+0.208)
Best Cube action: Double / Take
eXtreme Gammon Version: 2.10, MET: Kazaross XG2

I rolled. I really wasn't thinkin of cubing unless and until I cleared the 15 point. I am reluctant to cube knowing JQ can recube aggressively for the match. Especially since my takepoint on a recube at this 3 away 4 away score is a very high 40%.

And yet the double is clearly correct. The key seems to be JQ's double blotted one point board. Most likely I will abandon the 15 point next roll. John will have to hit and cover a point and then I will have to miss from the bar - else he's behind the eight ball. And this is a relatively small parlay. So since the crunch point is coming and I am a heavy favorite, now seems to be the time to double. It wasn't really on my radar at the table.



February 23, 2014

Baileys 2-20-14 v Dan Smith

Below are selected blunders from our club's weekly tournament.  I hope to record matches on a more frequent basis and analyze them here.
 


is Dan Smith

score: 0
pip: 155
7 point match
pip: 164
score: 0

is Bill
XGID=----bAD-C--AeD---a-eb---B-:0:0:1:21:0:0:0:7:10
to play 21

1.XG Roller++24/22 6/5eq: -0.135
Player:
Opponent:
47.76% (G:9.17% B:0.34%)
52.24% (G:13.97% B:0.45%)
2.XG Roller++13/11 6/5eq: -0.229 (-0.094)
Player:
Opponent:
46.00% (G:8.40% B:0.31%)
54.00% (G:16.05% B:0.54%)
3.3-ply11/9 6/5eq: -0.237 (-0.102)
Player:
Opponent:
45.26% (G:8.69% B:0.30%)
54.74% (G:15.92% B:0.70%)
eXtreme Gammon Version: 2.10, MET: Kazaross XG2


I played 13/11, 6/5 making two points and keeping my back men frozen. This turns out to be a whopper sized error, with the computer preferring 24/22, 6/5.

Here is what I was thinking at the time. Dan has a blitzing formation rather than a priming formation so I don't need to split. By splitting I am just encouraging him to unload those stacks on my head. Also, 13/11 establishes a new point. It eliminates the joker seven.

Perhaps these are the more relevant thinking points (though I admit I am confused on this one)...

Opponent has an anchor and I don't --> argues for spitting

Opponent only has two point board and eight in the zone so attack isn't likely to be too dangerous --> argues for splitting

Opponent's next play is likely to come down from the midpoint... splitting provides more coverage of the oufield --> argues for splitting

The 11 point, while an asset, can never be part of the same prime (since 11 and 5 point are six apart)


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


Later in the same game Dan cubes. Take or Drop?

is Dan Smith

score: 0
pip: 100
7 point match
pip: 131
score: 0

is Bill
XGID=--B-bBC-BB---BA--a-dc-bbaA:0:0:-1:00:0:0:0:7:10
on roll, cube action?

Analyzed in XG Roller++ No double Double/Take
  Player Winning Chances: 65.32% (G:41.29% B:0.48%) 65.38% (G:42.02% B:0.51%)
  Opponent Winning Chances: 34.68% (G:7.46% B:0.37%) 34.62% (G:7.43% B:0.38%)
  Cubeless Equities +0.652 +1.374
Cubeful Equities
     No double:+0.706 (-0.294)
     Double/Take:+1.109 (+0.109)
xg Double/Pass:+1.000
Best Cube action: Double / Pass
eXtreme Gammon Version: 2.10, MET: Kazaross XG2

If Dan doesn't cover and I hit, or if he does cover and I simply enter I am right back in this game. I have a strong offensive position and can win a lot. So I took. I underestimated the gammon losses - which are a staggering 41%. I guess when I lose most of the time I am closed out and in big trouble. In particular, the loose blot in the outfield makes a big difference. Pick up this blot and now the position is a comfortable take.

Something I have been trying to do lately is look at the winning chances in the position. Here they are 65/35% with huge gammons. I am hoping that taking the time to notice this will gradually sink into my brain what a 60/40% position looks like versus 65/35% or 70/30% etc. This knowledge would obviously help evaluate cube decisions over the board if one were able to attain it.

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


Next game I lead 5 away 7 away and Dan doubles. Take or Drop?


is Dan Smith

score: 0
pip: 160
7 point match
pip: 164
score: 2

is Bill
XGID=----abEaD---bD--ac-c---b-B:0:0:-1:00:2:0:0:7:10
on roll, cube action?

Analyzed in XG Roller++ No double Double/Take
  Player Winning Chances: 63.81% (G:27.80% B:0.77%) 63.77% (G:28.20% B:0.81%)
  Opponent Winning Chances: 36.19% (G:7.01% B:0.27%) 36.23% (G:7.18% B:0.31%)
  Cubeless Equities +0.493 +1.060
Cubeful Equities
     No double:+0.800 (-0.101)
xg Double/Take:+0.901
     Double/Pass:+1.000 (+0.099)
Best Cube action: Double / Take
eXtreme Gammon Version: 2.10, MET: Kazaross XG2


My thinking. This looks scary. nine in the zone. 2 on the bar, and Dan (while he has four back) they are not far back plus I have achieve nothing on the offensive side of the board. Also I have a lead which would tilt a bit towards being cautious. I dropped.

The computer shows this is a take. As far the numbers it is roughly a 65/35 position with 20% net gammons so this works out to a grin and bear it take.

As far as trying to reason through it, let's check Stick's relevant blitzing guidelines.

The race: I am down but it's not overwhelming. This is due to the unusual four men back. Race isn't ussually a big factor in blitz but it is worth noting that if I establish an anchor the race isn't out of the question.

Points Made:

Rule of thumb -- most 2 point boards are takes
-- 3 points or stronger are very dangerous

Here it is a two point board which points towards a take. Another slight factor may be that Dan has the 2 point and not the better 3 point. Meaning Dan is less likely to prime me; it's pretty much blitz or bust.

Rule of thumb -- number of men in the zone

Less than 10 men generally a take
More than 10 men generally a drop
Exactly 10 men it depends on other factors in the exact position

So Dan has nine men which would point towards a take.

Diversification & Efficiency of men in zone

Dan's distro is decent (men spread on three points and not overly stacked) but he has the inferior two point.


Offensive progress:

None. This is very bad for me and points towards a drop. I probably overemphasized this factor.


So add this all up and I guess the general factors point towards a take. In particular the 2 point board and nine in the zone and four men back all point towards take while no offensive progress and match lead point towards drop. All in all I guess a close take, which it apparently is.

Here is one last interesting measure to keep in mind. A friend gave me this one. He says take pip count of the blitzers front 12 men minus 42. 42 pips being a closed board. This is a rough estimate of how many rolls you have to make an anchor.

So here Dan has 99 pips with his front 12 men less 42 pips for a closed board or 57 pips. If you dive by 8 pips for an average roll then I have about 8 rolls to make an anchor before I get closed out (and that assumes Dan rolls perfect to make the offensive points). I am not sure what this really means, but I guess it is some sort of indication that I have a decent chance of surviving this blitz.

Finally here is one last rule of thumb on blitzes from Stick.

He writes most early game blitzes that are Double / Takes are about 65-68% wins with about 25-28% gammons for the stronger side. Here the breakdown is 64% wins and 27% gammons and the position is indeed a take. not that I would know this OTB as far as making an estimate, but geez this Stick guy is good!


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

I lead 5 away 6 away and have 53 to play.

is Dan Smith

score: 1
pip: 183
7 point match
pip: 166
score: 2

is Bill
XGID=-a--abE-C---dD---cAbbB----:0:0:1:53:2:1:0:7:10
to play 53

1.XG Roller++21/18 13/8eq: -0.095
Player:
Opponent:
49.97% (G:11.58% B:0.36%)
50.03% (G:13.67% B:0.46%)
2.XG Roller++21/18 21/16eq: -0.106 (-0.011)
Player:
Opponent:
49.99% (G:10.58% B:0.31%)
50.01% (G:13.79% B:0.45%)
3.XG Roller++8/3 6/3eq: -0.239 (-0.145)
Player:
Opponent:
45.17% (G:11.25% B:0.37%)
54.83% (G:12.60% B:0.48%)
eXtreme Gammon Version: 2.10, MET: Kazaross XG2


I made the 3 point with 8/3, 6/3. I figured this is going to be a long game and making an offensive point, or at least some kind of asset was important.

The computer prefers making the bar with 21/18 and then either 21/16 or 13/8. I have a hard time understanding this one. The best I can come up with is that by making the bar point I will never be primed. Also I am up in the race so moving the back men makes gameplan sense. Also Dan only has seven offensive checkers so I am hardly in any immediate danger. Perhaps another reason is that the 3 point, while an asset is not ideal. I would prefer to make points in front of Dan, or if I must make an inner point the 4 point is better.

This is another position I don't understand so I am really just guessing as to the reasons.


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

I lead 3 away 6 away and have 54 to play.


is Dan Smith

score: 1
pip: 148
7 point match
pip: 153
score: 4

is Bill
XGID=-b--B-D-C-A-bC-a-bbdb---B-:0:0:1:54:4:1:0:7:10
to play 54

1.XG Roller++13/9 10/5eq: -0.900
Player:
Opponent:
36.04% (G:11.40% B:0.75%)
63.96% (G:22.36% B:1.92%)
2.XG Roller++13/4eq: -1.000 (-0.100)
Player:
Opponent:
32.30% (G:9.83% B:0.52%)
67.70% (G:20.93% B:1.34%)
3.3-ply10/5 8/4eq: -0.992 (-0.092)
Player:
Opponent:
33.74% (G:9.73% B:0.53%)
66.26% (G:21.93% B:1.69%)
eXtreme Gammon Version: 2.10, MET: Kazaross XG2


First off I know I am getting cubed so is there any play I can make where I can take?

Also related is one play less likely to get cubed?


The answer to both questions points to 13/9, 10/5.

My position is fairly desperate so I need to take desperate measures to build an offense before it's too late. Slotting fills the bill. Now if Dan misses I will likely make the 5 point myself when I am right back in the game. Also there is some four duplication going on. 64, 42, 44 are all good rolls anyway.

Instead I made the hopeless 13/4 which duplicates my own numbers. I have ones and threes everywhere to make an offensive point while I also need threes and ones to split on the defensive side of the board. All in all a very poor play.

After this whopper my position is so bad I am supposed to drop! I took and ended up getting gammoned.

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


Later in the same game and I have a 51 to play. Your move?


is Dan Smith

score: 1
pip: 132
                         
7 point match
                          pip: 111
score: 4

is Bill
XGID=-BBBBcC-A-a------Acbbbb-B-:1:1:1:51:4:1:0:7:10
to play 51

1.XG Roller++24/23 8/3eq: -0.985
Player:
Opponent:
16.46% (G:4.21% B:0.05%)
83.54% (G:29.52% B:1.53%)
2.XG Roller++8/2eq: -1.115 (-0.131)
Player:
Opponent:
9.99% (G:1.71% B:0.02%)
90.01% (G:28.14% B:1.39%)
3.2-ply17/16 8/3eq: -1.139 (-0.154)
Player:
Opponent:
9.01% (G:1.41% B:0.02%)
90.99% (G:28.22% B:1.39%)

eXtreme Gammon Version: 2.10, MET: Kazaross XG2



Here I had mentally given up and was in "gammon save mode." I rather quickly played 8/2 safetying one block and holding the ace point. Again my hope is only leave the ace point as Dan is bearing off and run off the gammon.

However I can and should play the brave 24/23, 8/3. You can see from the numbers that this play wins a full 6% more while the gammon rate is only 1% more. A tough play to make over the board but this didn't even cross my radar. I suppose the lesson is to always conside the "DMP" play (the one which wins the most gammons be damned). Then make it unless you can see a good reason not too. In this case, getting gammoned is a great reason not to give up the anchor, except that it turns out you get gammoned much more...