News:

want some cool sounds while you browse the boards.... see the new collapsable web radio section on the LHS, below birthdays, on the Front page

Main Menu

Match 1, Game 3, Move 9, Forum cube

Started by blitzxz, May 21, 2009, 11:40:51 AM

Previous topic - Next topic

blitzxz

Take or pass?

Score:
Green(forum) 2, White(factotum) 0, match to 5

Pips:
Green 140 White 163

GnuID:
HM/IDCCw3eAAEw:ARmgAAAAEAAA

ah_clem

Spoiler
It's a pretty clear take.  A painful one, but playing on is the right way to go.  While it's tempting to just pass and continue the match at even score (2-2), we have pretty good winning chances here (somewhere north of 40% I'd reckon) and we win the match if we win the game.  If we lose, we'll have to win two games in a row to survive, which is bad but not unreachable.  Play on, even though it hurts.


[close]

Zorba

Spoiler

What is our takepoint (TP, % of wins needed to take a cube)?

If we drop the cube, it,s 3-away both (-3,-3) for 50% MWC.

If we take and win, we win the match for 100% MWC.
If we take and lose a single, it's -3,-1 Crawford for 25% MWC.
If we take and lose a gammon, we lose the match for 0% MWC.

Our lonely backchecker is vulnerable against the strong 4-prime with builder. We've already exposed a second blot that factotum wilk hit 10/36 of the time. We may anchor, but it's not guaranteed. Also, it may be a low anchor with serious gammon risks. Because of our awkward position (primed and stripped), we may expose another blot anytime. So I'd say there's a serious gammon risk here, clearly more than average (the opening position has ~27% gammon rate). Let's say 1/3 of our losses will be gammon losses.

This means that when we lose, 2/3 of the time it's single for 25% MWC and 1/3 of the time it will be a gammon loss for 0% MWC, overall that averages 17% MWC when we lose.

Now we can do the risk/gain calculation: the risk is the difference between dropping this cube, and taking and losing. Dropping is 50% MWC, taking and losing is 17%%, so the risk is 50-17 = 33% MWC.

The gain is the difference between taking and winning, or dropping the cube: 100-50 = 50% MWC.

Our takepoint is now risk/(risk+gain) = 33%/(33+50)% = 33%/83% ~= (2/6)/(5/6) = 2/5 = 40% MWC.

Ouch! We need 40% wins to take this cube if my estimates are correct. That is way higher than in a money game with the same gammon rates. A clear effect from the matchscore, combined with gammon wins.

Can we win 40% here? Hmm, I tend to think we can't, although it looks pretty close. So I pass. Practically speaking I also fear the team is more likely to make mistakes from here than factotum is, reducing our winning chances further. It looks like a tough position to play for our side, and easier for factotum. PASS.
[close]
The fascist's feelings of insecurity run so deep that he desperately needs a classification of some things as successful or superior and other things as failed or inferior. This also underlies the fascist's embracement of concepts like mental illness and IQ tests.  - R.J.V.

Luck is my main skill

stog

Spoiler
ok you persuaded me to change my original "take"  zorba -thx- and i almost understand why too! -(ty for the full explanation)
[close]

lewscannon

Spoiler
You're all worthless and weak! :smile: It's a risky take, granted, and Zorba provides excellent back-up to his opinion. However, this position doesn't scare me and I don't like giving this game up. Smart cube by factotum, as either he'll get away with an even match or has a shot at winning the whole thing with a gammon, if our dice go south
[close]

diane

Spoiler
I went back to my tried and tested strategy of 'doing what lews does', it has worked well for me in the past.... :laugh: :laugh: Seriously, I didn't want to cube this, and now I want to give it up even less....it looks hard and potentially messy, but it isn't as bad as it seems. And how could we possibly make more checkerplay errors than Fact - with all this brainpower thinking it through [as long as we aren't deciding moves by coin toss  ;)]
[close]
Never give up on the things that make you smile

ah_clem

Spoiler

Quote from: Zorba on May 21, 2009, 03:06:19 PM

Practically speaking I also fear the team is more likely to make mistakes from here than factotum is, reducing our winning chances further. It looks like a tough position to play for our side, and easier for factotum. PASS.

Hmmm.  That's an interesting line of analysis.  Although it may be theoretically correct to take this cube, the fact that we're collectively a not a very strong player means that we may have a better chance of winning by passing.  We're looking at a thorny position that will leave many opportunities for mistakes - perhaps we're just not up to it?

I'm fairly confident that this is a theoretical take, but like Zorba I'm not sure we have the ability to back it up by making the right moves ... so is it correct to make a deliberate cube error here?

I tend to drop cubes from gnubg when a back game is in progress or developing - I just don't know how to play those positions (some say that neither does gnubg, but it's better at it than I am) and I find I win more matches by "incorrect" cube handling.  I probably also lower my error rate by eating a single cube error rather than making a long string of checker play errors.

Maybe I should adopt the same approach here and change my vote....

[close]

blitzxz

Spoiler
Quote from: Zorba on May 21, 2009, 03:06:19 PM
What is our takepoint (TP, % of wins needed to take a cube)?

If we drop the cube, it,s 3-away both (-3,-3) for 50% MWC.

If we take and win, we win the match for 100% MWC.
If we take and lose a single, it's -3,-1 Crawford for 25% MWC.
If we take and lose a gammon, we lose the match for 0% MWC.

Our lonely backchecker is vulnerable against the strong 4-prime with builder. We've already exposed a second blot that factotum wilk hit 10/36 of the time. We may anchor, but it's not guaranteed. Also, it may be a low anchor with serious gammon risks. Because of our awkward position (primed and stripped), we may expose another blot anytime. So I'd say there's a serious gammon risk here, clearly more than average (the opening position has ~27% gammon rate). Let's say 1/3 of our losses will be gammon losses.

This means that when we lose, 2/3 of the time it's single for 25% MWC and 1/3 of the time it will be a gammon loss for 0% MWC, overall that averages 17% MWC when we lose.

Now we can do the risk/gain calculation: the risk is the difference between dropping this cube, and taking and losing. Dropping is 50% MWC, taking and losing is 17%%, so the risk is 50-17 = 33% MWC.

The gain is the difference between taking and winning, or dropping the cube: 100-50 = 50% MWC.

Our takepoint is now risk/(risk+gain) = 33%/(33+50)% = 33%/83% ~= (2/6)/(5/6) = 2/5 = 40% MWC.

Ouch! We need 40% wins to take this cube if my estimates are correct. That is way higher than in a money game with the same gammon rates. A clear effect from the matchscore, combined with gammon wins.

Can we win 40% here? Hmm, I tend to think we can't, although it looks pretty close. So I pass. Practically speaking I also fear the team is more likely to make mistakes from here than factotum is, reducing our winning chances further. It looks like a tough position to play for our side, and easier for factotum. PASS.

Great analyzes by Zorba but I don't agree with the desicion and voted for take.  :) This is because we have big advantage in pips and if we get over that prime we are immediate favorites to win (not by much though). And we will have several throws to hope for that joker. I don't use match winning chances as well as Zorba but my general borderline rules for these kind of positions are:

fivepoint prime, no builders and one checker behind on edge - take
four point prime, max one builder and one checker behind not on edge - take

Then I try to adjust these to score and pips. Score turns the scale for pass but big pip lead turns is back to take.
[close]

playBunny

 
Rollout, 1296 games, 2-ply throughout

Spoiler
Not an awful lot in it. I went for the take because the pipcount was ours and we only had one man behind the 4-prime. That wasn't looking to be extending too soon so escaping had good chances. I thought the main downside to our position was going to be the awkwardness of getting the midpoint and, especially, the 15-point home.

Cubeless equity  +0.330 (money:  +0.203)

1 .. Double, take ..... 0.958
2 .. Double, pass ..... 1.000 .. (+0.042)
3 .. No double ........ 0.872 .. (-0.085)

Proper cube action: Redouble, take
[close]