Double to DMP with only 14 hits and 11 pip deficit?

[boop]

well the title describes my confusion quite well i think -

Double to DMP with only 14 hits and 11 pip deficit?

plus ... lots of numbers are awful.

what gives?

b p

[ah_clem]

At 2a 2a the cube should have been turned a long time ago.

The simple rule of thumb for cube action at 2a 2a is:

Ship the cube the first chance you get.

At 2a2a, in almost every situation where you double at the first opportunity it is either positive or neutral equity-wise. In a few positions it will be a negligible error, but it's not worth memorizing the exceptions.  Just ship it and save your mental cycles for checker play.  If you win the opening roll, play like it is DMP.

Use this strategy and you will never lose your market. Looks like you lost your market here.  Don't make that mistake again.  This is one of the easiest to apply cube rules.

[boop]

hi ah_clem ... I was using the "if the opponent is much weaker than you, maybe you can wait a bit" method.
(Peter Bennet used it against me and won the match in a live game last month :-(   )

By move 4 I was on the bar against a 3 point board and so wasn't too disappointed with a central cube.

My opponent was a sub 1450 player so I also imagined that a gammon for me or 2 singles in later games would still be quite likely.

Obviously GNU disagrees with my thinking but it can't just be the "cube asap" rule can it?

[ah_clem]

Here's what i would suggest:

While you have the match open in gnu, go look at all the cube decisions leading up to where you finally pushed it.  How much equity did you give away by holding the first ten times?  My hunch is not all that much until the position displayed.  But you wouldn't have lost any had you cubed right away.

In the position posted, any six looks like a market loser (15 shots), so that should be a tip off to cube.

As for optimal strategy for asymmetrical player ability, you'll have to ask someone other than me if you want a really informed opinion.  But my take is that most poor players don't understand DMP as well as "normal" scores so might as well go the DMP route from the get go.

All that said, I do hold the cube sometimes at 2a2a against fenceposts when it's clear that I'm the underdog - give them a chance to make mistakes and leak equity.  But you have to remember to ship it at the first whiff of market losers.

[boop]

ok, so looking at the match in GNU

1st move I rolled first so no double option.
5 optional doubles (no equity loss)
2 no doubles
1 double take on move 9 (my not doubling cost me -0.035) .

On move 10, the move in question I lost a whopping -0.141 by not doubling. That's the bit I don't understand seeing as I'm not favorite to hit/escape with a pip lead.

Hmm - maybe it's the hit + escape probablity that gives me the edge?

I make the hit 14 rolls (42, 22 don't play)
A roll of 12 or more is needed to escape that checker with only 3 possible rolls - 66 55 33 (44 doesn't play)
A roll of 10 would leave me odds on to escape the single checker and only 1 pip behind. (not decisive)
Adding these together ... 14 + 3 + (-0.2 or something   ) = <50%

also taking into account my opponent had consistently played lots of terrible moves ... i didn't want her to have a win with a small series of jokers. GNU wouldn't have taken into account her previous and telling huge mistakes I imagine. Am I right there?

I still feel no double was right. Does anything above back me up?

[ah_clem]

The rollout in the diagram shows you having about 53% chance to win this game.  That alone is enough to ship it at 2a2a.  meanwhile, you have 15 shakes that hit and these look like cashes. Plus your opponent has six checkers buried which represents a long term liability unless she can fill in the other two points.  If it turns into a race that's quite a bit of wastage, so your race deficit is not as big as you think.

But I do think you may be on to something. The 'double right away' strategy assumes that your opponent will double as soon as she is over 50% GWC, but your opponent has shown that she is not going to double until she's lost her market.  That changes the calculus, but I'm not enough of a game theorist to say exactly how.  Cashing gets you to 68% MWC, and that's better than playing on with the cube at 2 with 52% GWC.  If provoking a wrong pass is in the offing, that changes even more.

So, if your opponent tends to  mishandle the cube, either by taking clear drops or by passing clear takes, then it might be better to hold for a while to get into a position where you can provoke a cube error.

I generally don't play opponents below about 1650, so I usually just ship it at 2a2a.

[boop]

thanks that makes more sense ... still need to chew it over. Markets are (very) slowly making more sense to me.

... it was a fibs tourney so I wasn't choosing to play a complete novice to up my rating 

[NIHILIST]

I've always been amused by this issue over the years.

For instance, how does one know who the better player is ? RollingFool is playing Nack Ballard, it's 23-23 in a 25 point match, who's the better player ?

When I was playing a lot I played with a bunch of guys in chouettes for years. In some cases I still don't know who the better player is.

Obviously if you are playing your opponent in a tourney for the first time and have never met him or seen him play before, and its a short match you might as well cube.

If you think your opponent is the better player and less prone to incorrect play than you are, you might as well cube.

The one that's always fascinated me though is what to do if you're convinced that you are the better player. Conventional wisdom holds that you don't turn the cube and let your opponent give himself enough rope to hang himself thru bad play. But if you truly are the better player don't you want to cube and play for the match ?

As an example let's say you've reached an end position in money play where you're on roll and have 1 checker on your 5 point and 1 checker on your 2 point with the cube in the middle. As a favorite at roughly 53-47% you cube and your opponent takes.

Now return to the issue at hand. You are convinced you're the better player. By how much ? By 53-47 ? If so, why not cube ?

How many players over the course of history were convinced they were the better player didn't cube, won 1 point, then lost the match next game when their fencepost opponent gammoned them ?

How often have both players believed they were superior and neither one doubled ?

Anyway, I've never been one who routinely bought into CONVENTIONAL WISDOM. I just turn the cube, especially if I think I'm the better player.


Bob

[ah_clem]

thanks that makes more sense ... still need to chew it over. Markets are (very) slowly making more sense to me.

Here's another way to think about it:  a cash gets you to 68% MWC.  What was her GWC at the time you actually cashed?  If above 32%, congratulations, you held the cube long enough to provoke a cube error. If less than 32%, you would have been better off playing that game for the match than resetting at 2a Crawford.

[dorbel]

Perhaps the confusion arises because players tend to think that a certain percentage of wins is needed to have a cube. If the cube remains live after the cube turn, ie it can be returned later, then you do need a considerable advantage to have a correct cube, but this doesn't apply when the cube is dead after the turn. In this case, any advantage at all is sufficient to justify a cube and sometimes it is correct even when a slight underdog if the position is sufficiently volatile. This position is volatile, because a hit followed by a fan (about 10%) loses the market, in fact red will be too good to double, but any market losing sequences at all would justify it.

plus ... lots of numbers are awful.

Red has no bad numbers here.

[boop]

Here's another way to think about it:  a cash gets you to 68% MWC.  What was her GWC at the time you actually cashed?  If above 32%, congratulations, you held the cube long enough to provoke a cube error. If less than 32%, you would have been better off playing that game for the match than resetting at 2a Crawford.

when I doubled it was a double pass with 67.69% MWC
32.31%  
but she'd already missed 8 double/takes according to GNU.

Can you tell me where you got your 68% from?
Maybe it's that at Crawford 2a1a is 75% minus opponents gammon chances.
Is there a Crawford equity table anywhere?

Red has no bad numbers here.

In the match I came to the conclusion that all these numbers were leading to an opponent's double and a possible pass.

11 12 13 14 22 23 24 25 34 35 44 45  

GNU assures me that they are all takes with about 40% winning chance which is higher than the @ 32% discussed above.
Looking at it that way they're not bad numbers as even if I miss and get hit I've a better than 32% chance of winning.
It's as simple as that huh?

Perhaps the confusion arises because players tend to think that a certain percentage of wins is needed to have a cube. If the cube remains live after the cube turn, ie it can be returned later, then you do need a considerable advantage to have a correct cube, but this doesn't apply when the cube is dead after the turn. In this case, any advantage at all is sufficient to justify a cube and sometimes it is correct even when a slight underdog if the position is sufficiently volatile. This position is volatile, because a hit followed by a fan (about 10%) loses the market, in fact red will be too good to double, but any market losing sequences at all would justify it.

Are there any articles on volatility?
I certainly was aware of her stripped outboard and her wrong home points with the 4 & 5 open but I didn't think of you market losing sequence. Well actually I thought the hit on its own was a market loser but I would cash then and get me to 1a2a Crawford with a high match win probability even without her continuous huge blunders.

So I can see 40% > 30% and double take was correct.

But I felt, and feel that the continued blunders and lack of strategy would give me an extra 10% or so ... maybe getting me to 40% again or even higher.
Her match fibs rating was 1017 and felt like it. Mine was 1840.

wrong?

... and in all honesty even against an equal or better player I'd have struggled to see this as a double take but your answers are helping me see the light.

thanks

[ah_clem]

Can you tell me where you got your 68% from?
Maybe it's that at Crawford 2a1a is 75% minus opponents gammon chances.
Is there a Crawford equity table anywhere?

All match equity tables assume Crawford in the first column and first row.  For instance,

http://www.simborgbackgammonlessons.com/11-point-met-rockwell-kazaross-1

and yes, 2a Crawford is 75% minus the opponent's gammon chances which are about 7%.


There are separate post-crawford match equity tables, but since post-crawford cube action is fairly straightforward most presentations omit it.   Zorba generated post-crawford numbers about ten years ago:

http://www.bkgm.com/rgb/rgb.cgi?view+1095

[boop]

thanks ah_clem