PWC: Progress Report 6

I've made some decisions on how I am going to proceed with Phase 2, and I'm easing back into study mode after a short break.

One note: I've never cared for the term money game, for several reasons. First, I don't play for money. Second, money game isn't very descriptive; it implies that money is at stake (when it may not be) while not really saying anything about the game itself. Anything can be a "money game." Third, it's typically understood that money games are played with the Jacoby rule in effect. I don't like the Jacoby rule and won't be using it in my study, so stepping away from that connotation is a plus.

Likewise, I've never cared for the term unlimited games, which has become the replacement term for money games. It's a better term than money games, but it doesn't roll off the tongue well, and it too is not all that descriptive. A game of backgammon is not "unlimited." Nor is a session of games "unlimited." The length may be indefinite or unspecified, but it is not unlimited; it will end at some point.  

In any case, I'll be using the terms session play (SP) and session games (SG) from now on. I find them easier on the ears, and more descriptive of what I'll be doing. For my purposes, a session is one or more distinct games, using the cube, with gammons and backgammons counting, and no Jacoby rule in effect.

Having gotten that out of the way, I've decided on a simple approach for Phase 2: I'll be playing session games and analyzing them as usual. At the same time, I'll be on the lookout for instances where the best move doesn't win the most single games. These are cases where the SP move is likely different from the DMP play. In this way, I can work on my SP while still striving to improve my DMP skills, and studying the differences should give me a better feel for how and when to take gammons into account.

I've only got a few SP games under my belt so far, but I've noticed that studying in this way is slower than it was with DMP. I'll still be giving progress reports after every 100 games, but they will be less frequent, since I'm not getting in as many games each day.

PWC: Progress Report 5

Graduation Day!

After 1000 1-point matches, my PR is 4.93.

This is down from 5.07 last time and, with that, Phase 1 of Project World Class (PWC) is officially a wrap.

To celebrate, I'm going take a bit of a break. I've been sticking with it pretty consistently, and I think pushing on immediately is apt to lead to burn out. Besides, I need to work out how I am going to approach Phase 2.

Phase 2 will be a transition to money (unlimited) play, with the goal of reaching a money PR of 5.0. However, I don't want to focus solely on money games. I still have a lot of room for improvement in my DMP play, and I want to keep pushing my PR down toward 4.0 on that front (though this is not officially a Phase 2 goal).

What I have in mind is an approach where I am working at money and DMP in tandem, comparing the effects of gammons and the cube for each move. I have a few ideas on how I might go about this, but I need to work out the details. I need an approach that is quick and simple, otherwise I risk getting bogged down with the technological side of things.

Until then... Woohoo. Go me.

Position 9

is Player 2 score: 0 pip: 159

1 point match

pip: 112 score: 0 is Player 1

XGID=-abBC-DBaA--bB---a-dbA-b--:0:0:1:32:0:0:0:1:10

to play 32

eXtreme Gammon Version: 2.19.211.pre-release

If I ever decide to post a "Blunder of the Day," this would be a good candidate.

I wasn't sure what to do and decided to swing away with 13/8*. This is a bad idea for a couple of reasons, and I should have known better. First, my play leaves four blots scattered about, and that should always be a red flag. I've got a solid advantage here, but bad things can happen in a hurry where blots are concerned. There needs to be a huge upside to take on even a small risk of things blowing up in your face. There's really no such upside in this position. In fact, the second reason my play is so bad is that there is actually a huge downside to hitting. I really don't want Brown making a second anchor in my board, especially with the extra timing that comes with the hit. That's just going to foul up the whole wretched mess.

I should have simply played 21/18 9/7. It may look risky, but it's not really. Brown has so many checkers back that there's not a lot he can do to me. If I get hit, I just make another run for it. I'm far enough ahead that I can afford to do this several times. As long as I'm not exposing several blots, I should be okay.

The rollout:

is Player 2 score: 0 pip: 159

1 point match

pip: 112 score: 0 is Player 1

XGID=-abBC-DBaA--bB---a-dbA-b--:0:0:1:32:0:0:0:1:10

to play 32

1.	Rollout1	21/18 9/7	eq: +0.296
	Player: Opponent: 64.82% (G:35.86% B:9.76%) 35.18% (G:6.50% B:0.52%) Conf.: ± 0.006 (+0.291...+0.302) - [100.0%] Duration: 3 minutes 37 seconds
2.	Rollout1	9/7 6/3	eq: +0.245 (-0.051)
	Player: Opponent: 62.27% (G:37.24% B:9.38%) 37.73% (G:8.32% B:0.59%) Conf.: ± 0.006 (+0.240...+0.251) - [0.0%] Duration: 3 minutes 30 seconds
3.	Rollout1	9/7 4/1*	eq: +0.238 (-0.058)
	Player: Opponent: 61.92% (G:37.29% B:8.06%) 38.08% (G:9.16% B:0.63%) Conf.: ± 0.005 (+0.233...+0.244) - [0.0%] Duration: 4 minutes 01 second
4.	Rollout1	21/16	eq: +0.236 (-0.061)
	Player: Opponent: 61.79% (G:34.09% B:7.10%) 38.21% (G:11.78% B:1.28%) Conf.: ± 0.005 (+0.230...+0.241) - [0.0%] Duration: 3 minutes 23 seconds
5.	Rollout1	9/4	eq: +0.206 (-0.090)
	Player: Opponent: 60.30% (G:35.40% B:8.37%) 39.70% (G:10.02% B:0.90%) Conf.: ± 0.006 (+0.200...+0.212) - [0.0%] Duration: 3 minutes 05 seconds
6.	Rollout1	6/1*	eq: +0.184 (-0.112)
	Player: Opponent: 59.21% (G:37.38% B:9.40%) 40.79% (G:13.42% B:1.81%) Conf.: ± 0.005 (+0.179...+0.190) - [0.0%] Duration: 3 minutes 32 seconds
7.	Rollout1	13/8*	eq: +0.184 (-0.113)
	Player: Opponent: 59.19% (G:37.87% B:11.66%) 40.81% (G:13.28% B:2.30%) Conf.: ± 0.006 (+0.178...+0.190) - [0.0%] Duration: 3 minutes 49 seconds
1 1296 Games rolled with Variance Reduction. Dice Seed: 60451593 Moves: 3-ply, cube decisions: XG Roller

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

Position 8: Take up the gauntlet!

is XG Roller++ score: 0 pip: 126

1 point match

pip: 153 score: 0 is Chase

XGID=--a-B-D-BB--bC--abbe-b--B-:0:0:1:21:0:0:0:1:10

to play 21

eXtreme Gammon Version: 2.19.211.pre-release

I've come to think of a blot on my Op's 9pt as a challenge to my manhood.

The reason is that, between the threat of simply making the 9pt and the variety of combinations which make other key offensive points, a builder on the 9pt is so well placed that it cannot be allowed to persist beyond the immediate turn. You must step up to challenge it directly, or take aggressive action on the other side of the board to create counterthreats of your own. Op has thrown down the gauntlet. You must take it up! Or forever hold your manhood cheap (apologies to the Bard).

I'm being a bit dramatic, of course, but Marty Storer has convinced me that stating rules of thumb as absolutes (Always make the 5pt, Always hit, etc.) is more helpful than stating them conditionally (When in doubt, make the 5pt; When in doubt, hit, etc.). The more definitive the statement, the easier it is to remember, and the more reliably will you apply it. That there will be exceptions goes without saying, but the absolute formulation requires a damn good reason to overturn. And the truth is, if the rule of thumb isn't robust enough to hold up as a near absolute, it's probably not all that helpful.

So, is my gauntlet rule sufficiently robust to be stated as an absolute? I don't know, but I've found it so helpful that (in the spirit of absoluteness) I'm going treat it as such until I'm convinced it is not.

The rollout:

is XG Roller++ score: 0 pip: 126

1 point match

pip: 153 score: 0 is Chase

XGID=--a-B-D-BB--bC--abbe-b--B-:0:0:1:21:0:0:0:1:10

to play 21

1.	Rollout1	13/11 6/5	eq: -0.143
	Player: Opponent: 42.86% (G:11.61% B:1.18%) 57.14% (G:29.83% B:9.36%) Conf.: ± 0.006 (-0.149...-0.136) - [100.0%] Duration: 4 minutes 58 seconds
2.	Rollout1	9/7 6/5	eq: -0.207 (-0.064)
	Player: Opponent: 39.65% (G:8.05% B:0.93%) 60.35% (G:33.35% B:12.93%) Conf.: ± 0.007 (-0.214...-0.200) - [0.0%] Duration: 4 minutes 18 seconds
3.	Rollout1	24/22 6/5	eq: -0.208 (-0.065)
	Player: Opponent: 39.60% (G:10.52% B:1.28%) 60.40% (G:38.30% B:4.69%) Conf.: ± 0.006 (-0.214...-0.203) - [0.0%] Duration: 4 minutes 07 seconds
4.	Rollout1	13/10	eq: -0.221 (-0.078)
	Player: Opponent: 38.94% (G:10.01% B:0.86%) 61.06% (G:28.05% B:7.43%) Conf.: ± 0.007 (-0.228...-0.214) - [0.0%] Duration: 3 minutes 39 seconds
5.	Rollout1	24/23 24/22	eq: -0.235 (-0.092)
	Player: Opponent: 38.24% (G:9.41% B:1.14%) 61.76% (G:31.73% B:3.56%) Conf.: ± 0.005 (-0.241...-0.230) - [0.0%] Duration: 4 minutes 14 seconds
1 1296 Games rolled with Variance Reduction. Dice Seed: 28711547 Moves: 3-ply, cube decisions: XG Roller

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

Position 7

is Player 2 score: 0 pip: 100

1 point match

pip: 121 score: 0 is Player 1

XGID=-B-B-aDBB----Aa--bbbbbcB--:0:0:1:63:0:0:0:1:10

to play 63

eXtreme Gammon Version: 2.19.211.pre-release

At DMP, it's often right, in positions like this, to hit on and slot key points. The hope is to make those points, prime Op's straggler, and get him to compromise his offense. It's a long shot strategy, but the situation is so desperate that there isn't much else to do. And since gammons don't matter, there's little downside to having a bunch of checker sent back. In fact, if you don't succeed in containing the straggler, you want those checkers sent back, as it will sometimes keep you on the bar and prevent you from collapsing your own offensive structure. If you can avoid that, you just might hit a lucky shot and snag a satisfying win.

That was my approach here. I played 8/5* 13/7, hoping to be missed, then to make my 5pt and keep Brown trapped behind my mini-prime until something shakes lose on the other side of the board.

What I overlooked is that I already have a mini-prime and a chance to shake something lose on the other side of the board. As things sit, Brown is vulnerable to an immediate 33. It isn't much, but the swing on that one roll is very nice for me. Playing 8/5* takes away that nice root number and gives Brown no bad rolls in exchange.

The rollout:

is Player 2 score: 0 pip: 100

1 point match

pip: 121 score: 0 is Player 1

XGID=-B-B-aDBB----Aa--bbbbbcB--:0:0:1:63:0:0:0:1:10

to play 63

1.	Rollout1	13/4	eq: -0.722
	Player: Opponent: 13.91% (G:0.79% B:0.01%) 86.09% (G:9.78% B:0.45%) Conf.: ± 0.002 (-0.724...-0.720) - [100.0%] Duration: 2 minutes 00 second
2.	Rollout1	13/7 6/3	eq: -0.737 (-0.016)
	Player: Opponent: 13.13% (G:0.49% B:0.00%) 86.87% (G:9.73% B:0.32%) Conf.: ± 0.002 (-0.739...-0.736) - [0.0%] Duration: 1 minute 50 seconds
3.	Rollout1	7/4 7/1	eq: -0.753 (-0.031)
	Player: Opponent: 12.37% (G:0.62% B:0.00%) 87.63% (G:14.41% B:0.94%) Conf.: ± 0.002 (-0.755...-0.751) - [0.0%] Duration: 1 minute 59 seconds
4.	Rollout1	8/2 7/4	eq: -0.761 (-0.039)
	Player: Opponent: 11.96% (G:0.38% B:0.00%) 88.04% (G:27.65% B:3.88%) Conf.: ± 0.002 (-0.763...-0.758) - [0.0%] Duration: 2 minutes 19 seconds
5.	Rollout1	13/7 8/5*	eq: -0.761 (-0.039)
	Player: Opponent: 11.93% (G:0.50% B:0.00%) 88.07% (G:26.61% B:3.23%) Conf.: ± 0.002 (-0.764...-0.759) - [0.0%] Duration: 3 minutes 00 second
1 2592 Games rolled with Variance Reduction. Dice Seed: 5077922 Moves: 3-ply, cube decisions: XG Roller

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