Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Step up? 2

56 views
Skip to first unread message

Tim Chow

unread,
Sep 17, 2015, 9:26:12 PM9/17/15
to
XGID=---BbBBBB-B--B--acbc-bb-A-:0:0:1:41:0:0:0:0:10

X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O O | | O O O X |
| X O O | | O O O |
| O | | O |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| | | |
| X X X | | X X O X |
| X X X | | X X O X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 128 O: 121 X-O: 0-0
Cube: 1
X to play 41

---
Tim Chow

Chase

unread,
Sep 17, 2015, 9:51:00 PM9/17/15
to
I guess the step up play is 24/20 13/12, which looks pretty dangerous
and introduces a couple of nice fly shots for O as well. On the other
hand, sitting back and waiting to be smothered is plenty dangerous too.
I've no clue, but I would play 13/8 and hope for something better later.

13/8




---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

Bradley K. Sherman

unread,
Sep 18, 2015, 1:49:34 AM9/18/15
to
Tim Chow <tchow...@yahoo.com> wrote:
>Score is X:0 O:0. Unlimited Game
> +13-14-15-16-17-18------19-20-21-22-23-24-+
> | X O O O | | O O O X |
> | X O O | | O O O |
> | O | | O |
> | | | |
> | | | |
> | |BAR| |
> | | | |
> | | | |
> | | | |
> | X X X | | X X O X |
> | X X X | | X X O X |
> +12-11-10--9--8--7-------6--5--4--3--2--1-+
> X:128 O:121, Cube:1, X to play 41

13/8

--bks

bananab...@gmail.com

unread,
Sep 18, 2015, 2:04:45 AM9/18/15
to
'Looks pretty dangerous' translates to 'many blots worried about gammons'?

Stick asks...isn't your dmp play clear? If so, then make it.

Stick

michae...@gmail.com

unread,
Sep 18, 2015, 7:51:13 AM9/18/15
to
I am not sure if playing 24/20 13/12 is the DMP, but I will go in another way to check this position.
Firstly I am not too much worried of playing 13/12. O needs a 53 or 54 to hit and even if this happens she should worry more than us in loosing a gammon.

Now about playing the 4
Not stepping up to 20: O makes her 5 point and full prime with 13,14,43
Stepping up to 20: It adds two additional rolls 42,32.
What's 2 additional rolls compared to the benefit of stepping up? So definitely

24/20 13/12


badgolferman

unread,
Sep 18, 2015, 9:19:26 AM9/18/15
to
I would probably play 24/20, 3/2 OTB and take my chances, but I'm not
feeling lucky today so I'll try 7/6, 7/3 instead.

Tim Chow

unread,
Sep 18, 2015, 11:21:38 AM9/18/15
to
On Friday, September 18, 2015 at 2:04:45 AM UTC-4, bananab...@gmail.com wrote:
> 'Looks pretty dangerous' translates to 'many blots worried about gammons'?
>
> Stick asks...isn't your dmp play clear? If so, then make it.

This is a great example of why I find the DMP rule mostly useless. The DMP was no clearer to me here than the money-game play.

---
Tim Chow

Chase

unread,
Sep 18, 2015, 1:45:57 PM9/18/15
to
To me, this simply means you need to focus your study on DMP play. If,
as advertised, the DMP play is the best play 95 percent of the time at
other scores, I sure as heck want to know what the DMP play is as often
as possible.

In this position, the DMP play wasn't clear to me either, so the rule
didn't help me find the right play. But I'm constantly coming across
positions where I'm very confident I can spot the DMP play (my DMP PR
has dropped about 2 points; granted it was not very good to begin with).
No rule helps on every play, but this one comes in handy quite a lot,
and it's easy to apply over the board.

I'd say the same thing about Woolsey's rule by the way. Even if your
doubling judgement is way off (meaning you're not apt to make the right
decision to begin with), it can be helpful. And it will become more
helpful as you continue to improve your judgement. And, again, it's
extremely easy to apply over the board. That's my kind of "rule."

Anyway, just my perspective. As always, do what works for you.

Tim Chow

unread,
Sep 18, 2015, 6:40:22 PM9/18/15
to
On Friday, September 18, 2015 at 1:45:57 PM UTC-4, Chase wrote:
> To me, this simply means you need to focus your study on DMP play. If,
> as advertised, the DMP play is the best play 95 percent of the time at
> other scores, I sure as heck want to know what the DMP play is as often
> as possible.

I don't agree with this logic. The money-game play is the best play 95% of the time at other scores, too. I sure as heck want to know the money-game play is as often as possible. If I know that, then I can also have a good handle on DMP as well, since the DMP play will usually be the same as the money-game play.

A large part of backgammon is training one's intuition. If I'm going to be playing money games a lot, I believe that I should be training my money-game intuition primarily.

Having said that, I do agree that one way to improve one's play is to learn how to estimate wins and gammons, and how many more games one play wins compared to another. This is something I want to, and do, train on. But I do this by studying money-game and match-play positions, rather than by spending a lot of time on DMP specifically. When I get a position I don't understand, I first check the money-game play, and then I look at the bot's pseudocubeless wins and gammons. This allows me to learn something about all scores at once.

I did once try spending some time focusing on DMP exclusively. It was interesting, but if anything it hurt my money-game play because it caused me to make blunderful loose plays more often than I used to. I also lost some of my feeling for the cube.

> In this position, the DMP play wasn't clear to me either, so the rule
> didn't help me find the right play. But I'm constantly coming across
> positions where I'm very confident I can spot the DMP play (my DMP PR
> has dropped about 2 points; granted it was not very good to begin with).
> No rule helps on every play, but this one comes in handy quite a lot,
> and it's easy to apply over the board.

It's good that you find that the rule works for you. It doesn't work for me.

> I'd say the same thing about Woolsey's rule by the way. Even if your
> doubling judgement is way off (meaning you're not apt to make the right
> decision to begin with), it can be helpful. And it will become more
> helpful as you continue to improve your judgement. And, again, it's
> extremely easy to apply over the board. That's my kind of "rule."

I agree that Woolsey's rule can be useful for some people some of the time. It was not helpful when I was doubling too aggressively. It was helpful in guiding me when to stop dialing down my doubling aggression. Nowadays I don't find it helpful because it's much more helpful for me to focus on the specific types of situations in which I double too aggressively and don't double aggressively enough, rather than try to apply some sort of blanket principle.

---
Tim Chow

Tim Chow

unread,
Sep 18, 2015, 6:46:06 PM9/18/15
to
On Friday, September 18, 2015 at 6:40:22 PM UTC-4, I wrote:
> I did once try spending some time focusing on DMP exclusively. It was
> interesting, but if anything it hurt my money-game play because it caused
> me to make blunderful loose plays more often than I used to. I also lost
> some of my feeling for the cube.

Oh, by the way, just to add to this list, I also developed bad habits in the endgame by bearing off insufficiently aggressively and losing my feeling for when I needed to start running off the gammon.

---
Tim Chow

Chase

unread,
Sep 18, 2015, 8:17:32 PM9/18/15
to
On 9/18/2015 6:40 PM, Tim Chow wrote:
> On Friday, September 18, 2015 at 1:45:57 PM UTC-4, Chase wrote:
>> To me, this simply means you need to focus your study on DMP play. If,
>> as advertised, the DMP play is the best play 95 percent of the time at
>> other scores, I sure as heck want to know what the DMP play is as often
>> as possible.
>
> I don't agree with this logic. The money-game play is the best play 95% of the time at other scores, too. I sure as heck want to know the money-game play is as often as possible. If I know that, then I can also have a good handle on DMP as well, since the DMP play will usually be the same as the money-game play.

It sounds like you do agree with the logic. You just think money play
serves as a better point of comparison than DMP. Fair enough.


> A large part of backgammon is training one's intuition. If I'm going to be playing money games a lot, I believe that I should be training my money-game intuition primarily.

Again, fair enough. I rarely play money games, so perhaps that's a
source of our different perspectives. That said, even if I'm better off
knowing the money move than the DMP move, I'm much more confident in my
ability to figure out the DMP move, without having to factor in gammons
and upcoming cube decisions. It's simpler, and I like simple.

Chase

unread,
Sep 18, 2015, 8:19:47 PM9/18/15
to
I haven't spent any amount of time studying bear-off play, but I think
you make a good point -- one I will keep in mind when I get around to
working on it.

michae...@gmail.com

unread,
Sep 19, 2015, 8:40:49 AM9/19/15
to
I was wondering why the DMP is the correct play (presumably) 95% of the times.
If this were true then we should 95% of the times ignore whatever increase in gammon losses a move results to, as long as it increases single wins.
Surprisingly however the DMP criterion is not far from the truth.
I did some research based on the MET table and here are my results:

Unlimited games at score 0-0:
You can exchange A% more single wins for about 2A% as many gammon losses

Matches when both players are more than 4A, either you are Trailing or Leading: Same as above
Matches between 4A and 2A:depends on the score and whether you are trailing or leading, but it never falls below 1.1A%

NB.From preliminary observation, cube possession doesn't seem to affect the percentages. But I could be wrong on this.



Tim Chow

unread,
Sep 19, 2015, 2:35:03 PM9/19/15
to
On Saturday, September 19, 2015 at 8:40:49 AM UTC-4, michae...@gmail.com wrote:
> I was wondering why the DMP is the correct play (presumably) 95% of the times.
> If this were true then we should 95% of the times ignore whatever increase
> in gammon losses a move results to, as long as it increases single wins.

Gammonish positions tend to be gammonish both ways and tend to cancel each other out.

A less formal way of saying it is that the best play is just the best play. I don't know the history of the game, but my guess is that the gammon rule was introduced to keep the game interesting even when it became very lopsided. People probably expected that the gammon rule wouldn't make much difference except in such lopsided cases. I'm sure they underestimated the effect that gammons would have on "ordinary" play but by and large the instinct was correct.

---
Tim Chow

Tim Chow

unread,
Sep 19, 2015, 2:44:21 PM9/19/15
to
On Friday, September 18, 2015 at 8:17:32 PM UTC-4, Chase wrote:
> It sounds like you do agree with the logic. You just think money play
> serves as a better point of comparison than DMP. Fair enough.

Well, I thought the "logic" involved the logical inference to the conclusion, "You need to focus your study on DMP play." I don't dispute the *fact* that the DMP play is usually the correct play at all scores.

Usually, the best play is usually the best play, 95% of the time, independent of score. Certainly, one needs to develop the ability to detect the best play in this 95% cohort. But how does the observation that the best play is the DMP play help me find it? It's just a change in vocabulary. If I used the French or German word for "best" instead of the English word, it wouldn't help me find it. It only helps if the DMP play is easier to find than the best play. Generally, I don't find that to be the case. If I can find the DMP play then I can find the best play, and vice versa. After all, they're usually the same! The exceptions are those situations where the DMP play is blindingly obvious (like holding onto a last-ditch ace point anchor without worrying about gammons) *but* there are a lot of gammons to consider. In such positions, one has to try to weigh wins against gammons. I agree that *that* skill is important to develop. But that's entirely different from Stick's "DMP rule."

In the position at hand, when I said that I wasn't sure what the DMP play was, I doubt that studying DMP would help me any more than studying money games. I just don't understand the position well enough to know what's going on. Yes, I need to study. No, I don't agree that shifting to DMP specifically, as opposed to studying wins and gammons and general strategy, will fast-track me to the right answer in positions like this one.

---
Tim Chow

Tim Chow

unread,
Sep 19, 2015, 3:03:13 PM9/19/15
to
Again, I think that what one has to judge is O's priming potential versus her attacking potential. I have included a rollout of the original position as well as XGR++ evaluations of three variants:

1. Move X's blot from the 24pt to the 23pt and change the roll to 31.
2. Move O's spare from her 8pt to her 11pt.
3. Both of the above.

The rollout says to step up in the original position. #1 gives X better defense against O's prime so he's less eager to step up. #2 gives O better attacking chances, again making X less eager to step up. #3 combines both factors and now XGR++ doesn't want to step up at all.

1. Rollout¹ 24/20 13/12 eq:+0.003
Player: 51.83% (G:9.71% B:0.57%)
Opponent: 48.17% (G:17.21% B:0.67%)
Confidence: ±0.013 (-0.011..+0.016) - [100.0%]

2. Rollout¹ 13/8 eq:-0.096 (-0.099)
Player: 47.44% (G:7.46% B:0.29%)
Opponent: 52.56% (G:11.88% B:0.32%)
Confidence: ±0.014 (-0.111..-0.082) - [0.0%]

¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller

eXtreme Gammon Version: 2.19.208.pre-release

---------
Variant 1
---------

XGID=---BbBBBB-B--B--acbc-bbA--:0:0:1:31:0:0:0:0:10

X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O O | | O O O X |
| X O O | | O O O |
| O | | O |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| | | |
| X X X | | X X O X |
| X X X | | X X O X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 127 O: 121 X-O: 0-0
Cube: 1
X to play 31

1. XG Roller++ 23/20 13/12 eq:-0.011
Player: 52.10% (G:9.10% B:0.44%)
Opponent: 47.90% (G:16.07% B:0.57%)

2. XG Roller++ 13/12 13/10 eq:-0.060 (-0.049)
Player: 49.55% (G:7.34% B:0.26%)
Opponent: 50.45% (G:12.43% B:0.26%)

eXtreme Gammon Version: 2.19.208.pre-release

---------
Variant 2
---------

XGID=---BbBBBB-B--Ba-abbc-bb-A-:0:0:1:41:0:0:0:0:10

X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O O O | | O O O X |
| X O O | | O O O |
| | | O |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| | | |
| X X X | | X X O X |
| X X X | | X X O X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 128 O: 124 X-O: 0-0
Cube: 1
X to play 41

1. XG Roller++ 24/20 13/12 eq:-0.222
Player: 46.56% (G:9.37% B:0.54%)
Opponent: 53.44% (G:20.06% B:0.66%)

2. XG Roller++ 13/8 eq:-0.284 (-0.062)
Player: 43.10% (G:7.20% B:0.28%)
Opponent: 56.90% (G:13.01% B:0.35%)

3. XG Roller++ 24/23 13/9 eq:-0.290 (-0.068)
Player: 44.04% (G:8.72% B:0.48%)
Opponent: 55.96% (G:19.08% B:0.81%)

eXtreme Gammon Version: 2.19.208.pre-release

---------
Variant 3
---------

XGID=---BbBBBB-B--Ba-abbc-bbA--:0:0:1:31:0:0:0:0:10

X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O O O | | O O O X |
| X O O | | O O O |
| | | O |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| | | |
| X X X | | X X O X |
| X X X | | X X O X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 127 O: 124 X-O: 0-0
Cube: 1
X to play 31

1. XG Roller++ 13/12 13/10 eq:-0.138
Player: 47.19% (G:7.58% B:0.27%)
Opponent: 52.81% (G:12.54% B:0.27%)

2. XG Roller++ 23/20 13/12 eq:-0.222 (-0.084)
Player: 46.56% (G:9.37% B:0.54%)
Opponent: 53.44% (G:20.06% B:0.66%)

eXtreme Gammon Version: 2.19.208.pre-release

---
Tim Chow

michae...@gmail.com

unread,
Sep 19, 2015, 3:26:37 PM9/19/15
to
On Saturday, September 19, 2015 at 9:03:13 PM UTC+2, Tim Chow wrote:
>
> Again, I think that what one has to judge is O's priming potential versus her attacking potential.

I don't understand how you use this criterion.
Re the Original position if X won't step up attacking potential is very low, whereas priming potential has say a value A.
If he steps up attacking and priming is essentially the same.
Could you please explain how you go from there?

Tim Chow

unread,
Sep 19, 2015, 3:37:29 PM9/19/15
to
Usually, stepping up makes X more vulnerable to being attacked, whereas staying back makes X more vulnerable to being primed.

Assume that strategically speaking, X would like to escape. Then, if O's priming potential is stronger than her attacking potential, X will usually want to step up. If O's attacking potential is stronger than her priming potential, then X will usually want to stay back.

Another way to put it is, does X prefer to step up and risk getting attacked, or stay back and risk getting primed?

---
Tim Chow

jimt...@gmail.com

unread,
Sep 24, 2015, 8:38:44 PM9/24/15
to
I don't know what the DMP play is.

michae...@gmail.com

unread,
Sep 25, 2015, 9:20:02 AM9/25/15
to
On Friday, September 25, 2015 at 2:38:44 AM UTC+2, jimt...@gmail.com wrote:

>
> I don't know what the DMP play is.

It's when you risk it all to win a game. Usually this would be the case when losing a gammon won't matter. e.g. in a 1 point match.

Tim Chow

unread,
Sep 25, 2015, 1:05:03 PM9/25/15
to
On Thursday, September 24, 2015 at 8:38:44 PM UTC-4, jimt...@gmail.com wrote:
> I don't know what the DMP play is.

"DMP" stands for "double match point," or equivalently a 1-point match. All that matters is who wins. Gammons and backgammons don't count and there is no cube in play.
---
Tim Chow
0 new messages