PWC: Progress Report 4

After 900 1-point matches, my PR is 5.07.

This is up from 5.04 last time, so a bit of a setback over the last 100 matches. I went through a difficult stretch just after my last progress report where my PR rose sharply over the course of 60 or so games. I don't know if the decisions during that time happened to be tougher than usual or if I was just playing poorly. I'd guess it was more the latter than the former, but probably a mix of the two. In any case, I played fairly well toward the end and managed to nudge my PR back down.

The good news is that XG reports my PR over my last 500 games as 4.93, which gives me some hope that I can reach and sustain a PR lower than 5.0 over the long haul. For now, I continue to grind away.

I haven't put enough time into learning the best replies to all the openings. For the most part I'm solid, but there are a few areas where I am guessing quite a lot. I was quizzing myself on replies to 52S last week, for example, and found that several of the proper replies surprised me. Learning these is just a matter of putting in the time, of course, but it's a bit boring, so it takes some discipline.

Position 6

is XG Roller++ score: 0 pip: 98

1 point match

pip: 73 score: 0 is Chase

XGID=-BABbCD-C--------a-bccbaa-:0:0:1:61:0:0:0:1:10

to play 61

eXtreme Gammon Version: 2.19.211.pre-release

The 6 is forced, then find a 1.

Over the board, I dismissed 6/5 since it leaves a shot on 66 next turn. Then I had a choice between 3/2 and 2/1. I chose 3/2 for reasons I don't recall and can't reconstruct. Perhaps I simply hallucinated. In any case, 3/2 is a surprisingly large error, one I'm pretty sure I wouldn't have made if I'd approached the problem in the right way.

My poor rolls next turn are 64 and 61, both of which leave a blot, so I should be looking to make those rolls play as constructively as possible. 2/1 is obviously better than 3/2, then, since it allows me to remake the 2pt and have a five point board with either roll.

Here's the rollout. Note the second best move (8/2 6/5) is almost as good, even though it pays off to 66 next turn. My play (8/2 3/2) is much worse, even though it avoids blotting on 66 next turn. Since I will roll 64 or 61 four times as often as 66, planning for those rolls is much more important than avoiding the bad 66.

is XG Roller++ score: 0 pip: 98

1 point match

pip: 73 score: 0 is Chase

XGID=-BABbCD-C--------a-bccbaa-:0:0:1:61:0:0:0:1:10

to play 61

1.	Rollout1	8/1	eq: +0.725
	Player: Opponent: 86.25% (G:4.46% B:0.02%) 13.75% (G:0.47% B:0.00%) Conf.: ± 0.002 (+0.723...+0.727) - [66.1%] Duration: 33.3 seconds
2.	Rollout1	8/2 6/5	eq: +0.725 (-0.001)
	Player: Opponent: 86.23% (G:5.00% B:0.02%) 13.77% (G:0.41% B:0.00%) Conf.: ± 0.002 (+0.723...+0.726) - [33.9%] Duration: 37.4 seconds
3.	Rollout1	8/2 3/2	eq: +0.708 (-0.017)
	Player: Opponent: 85.41% (G:3.79% B:0.02%) 14.59% (G:0.49% B:0.01%) Conf.: ± 0.002 (+0.707...+0.710) - [0.0%] Duration: 36.7 seconds
4.	1-ply	8/7 8/2	eq: +0.161 (-0.564)
	Player: Opponent: 58.04% (G:5.84% B:0.02%) 41.96% (G:7.61% B:0.11%)
1 1296 Games rolled with Variance Reduction. Dice Seed: 77140590 Moves: 3-ply, cube decisions: XG Roller

eXtreme Gammon Version: 2.19.211.pre-release, MET: Kazaross XG2

Position 5

is XG Roller++ score: 0 pip: 109

1 point match

pip: 127 score: 0 is Chase

XGID=-A-a-CBCB-B-b-a--b-cbB-bb-:0:0:1:21:0:0:0:1:10

to play 21

eXtreme Gammon Version: 2.19.211.pre-release

This 21 isn't all that useful. I can't hit Brown's outfield blot. I can't make a new blocking point. And attacking on the 3pt leaves a double shot from the bar against Brown's stronger board. I figured I'd make the most of it with 10/8 7/6. This leaves 2 shots in exchange for creating an extra builder/attacker for next turn, which seems like a decent trade off.

What I didn't consider is that there won't be a "next turn" if Brown rolls any 6. He'll just run into the outfield and have a pretty easy time of it from there. Once you appreciate that, hitting loose on the 3pt, and depriving him of a "winning" 6, becomes a lot more attractive.

5/3* 7/6 is the best play, and while it comes at the risk of  20 direct shots, it's a relative risk -- relative to the risk of leaving Brown alone to roll his 6. I won't like it if I'm hit back, but Brown will still be trapped behind my broken prime, and I'm a favorite to enter and get some counterplay going.

In the actual game, I was appropriately punished for my misplay, as Brown rolled 66 and cruised home easily.

The rollout:

is XG Roller++ score: 0 pip: 109

1 point match

pip: 127 score: 0 is Chase

XGID=-A-a-CBCB-B-b-a--b-cbB-bb-:0:0:1:21:0:0:0:1:10

to play 21

1.	Rollout1	7/6 5/3*	eq: -0.102
	Player: Opponent: 44.88% (G:8.71% B:0.37%) 55.12% (G:26.73% B:0.40%) Conf.: ± 0.005 (-0.107...-0.098) - [100.0%] Duration: 3 minutes 01 second
2.	Rollout1	10/8 7/6	eq: -0.135 (-0.032)
	Player: Opponent: 43.26% (G:8.05% B:0.28%) 56.74% (G:14.32% B:0.29%) Conf.: ± 0.004 (-0.139...-0.131) - [0.0%] Duration: 2 minutes 02 seconds
1 1296 Games rolled with Variance Reduction. Dice Seed: 73285410 Moves: 3-ply, cube decisions: XG Roller

eXtreme Gammon Version: 2.19.211.pre-release, MET: Kazaross XG2

PWC: Progress Report 3

After 800 1-point matches my PR is 5.04.

This is down from 5.15 last time, so the trend continues. In addition, XG reports that my PR for my last 500 DMP matches is 4.99, which is obviously encouraging. My next progress report will come after I've played 900 matches. I'm hoping to be able to boast a PR of 5.0 or below at that time, which would complete Phase 1 of PWC.

I've identified an area of my game that I need to work on, and that is knowing when to slot the offense barpoint (7pt). This comes up fairly often in the early and middle game when you have a 6 to play and must decide between 24/18 or 13/7. Timothy Chow deserves of a shout out here, as he's posted a series of positions of this type on rec.games.backgammon, and also pointed me to a couple of articles by Walter Trice on Gammon Village. I'm still studying those, and also collecting additional positions from my own matches. Hopefully I can put all this together and gain some understanding, because right now it's hit and miss for me.

Position 4

Here is a position I totally butchered.

is XG Roller++ score: 0 pip: 123

1 point match

pip: 98 score: 0 is Chase

XGID=-ABaBBCC----b--b-bBbcc----:0:0:1:33:0:0:0:1:10

to play 33

eXtreme Gammon Version: 2.19.211.pre-release

Small doubles are often difficult to play, in part because there are typically so many options to consider. That wasn't the problem here, however.  Here, I was so worried about playing safely up front, so as not to leave Brown a direct shot from the bar and let him get back into it, that I failed to consider the elephant in the room: how am I going to get my back checkers home?

The answer, of course, is that the shot I feared leaving was the very solution I should have been looking for. If I play safely, all the pressure is on me. Brown can simply hop into my outfield, hold his outside points, and wait for me break my anchor with a 6.

The right idea is to put Brown on the bar against a five point board. This puts all the pressure on him. He's got to perform immediately or risk being closed out and having me stroll home.

This is an embarrassing blunder to share, but perhaps a little embarrassment will help me avoid such a beginner mistake in the future.

is XG Roller++ score: 0 pip: 123

1 point match

pip: 98 score: 0 is Chase

XGID=-ABaBBCC----b--b-bBbcc----:0:0:1:33:0:0:0:1:10

to play 33

1.	XG Roller++	7/4 7/1 6/3*	eq: +0.391
	Player: Opponent: 69.57% (G:36.44% B:0.08%) 30.43% (G:6.93% B:0.34%)
2.	XG Roller++	7/1 6/3*(2)	eq: +0.342 (-0.049)
	Player: Opponent: 67.12% (G:28.13% B:0.06%) 32.88% (G:5.20% B:0.18%)
3.	XG Roller++	7/4 6/3*(3)	eq: +0.172 (-0.219)
	Player: Opponent: 58.62% (G:17.00% B:0.15%) 41.38% (G:7.81% B:0.30%)
4.	XG Roller++	7/4(2) 7/1	eq: +0.120 (-0.271)
	Player: Opponent: 56.00% (G:13.63% B:0.10%) 44.00% (G:5.86% B:0.22%)

eXtreme Gammon Version: 2.19.211.pre-release, MET: Kazaross XG2