News:

VIP Donors can remain anonymous while still enjoying all the benefits.......

Main Menu

Cube 122 - Let the race begin?

Started by sixty_something, June 26, 2010, 06:13:45 AM

Previous topic - Next topic

sixty_something

one vote per user .. results of the poll are shown after you have voted

click on the spoiler below for a copy of the GNU analysis


Spoiler
From RED's perspective, the GNUbg analysis shows:

Cube analysis
0-ply cubeless equity  +0.601 (Money:  +0.601)
 0.800 0.001 0.000 - 0.200 0.000 0.000

Cubeful equities:
1. Double, pass         +1.000
2. Double, take         +1.007  ( +0.007)
3. No double            +0.842  ( -0.158)

Proper cube action: Double, pass

In Match winning Chances (MWC):


Cube analysis
0-ply cubeless MWC  54.64% (Money:  +0.601)
 0.800 0.001 0.000 - 0.200 0.000 0.000

Cubeful equities:
1. Double, pass         57.72%
2. Double, take         57.77%  (  0.05%)
3. No double            56.50%  ( -1.22%)

Proper cube action: Double, pass

---------------------

From WHITE's perspective we see the complementary vales:


Cube analysis
0-ply cubeless equity  +0.601 (Money:  +0.601)
 0.800 0.001 0.000 - 0.200 0.000 0.000

Cubeful equities:
1. Pass         -1.000
2. Take         -1.007   -0.007

Proper cube action: Pass

SUMMARY:

This is so close, it is hard to call it a true pass. For example, in the game itself i took the cube and it was not marked as it fell under the threshhold of -0.008 for marking a decision doubtful. In the pip count WHITE has a 2 pip lead. So, effectively the two players are even. WHITE will have a slight advantage once bearoff begins, but that may be two or more rolls away, while RED may get its first pip off on this roll - most certainly the next.

RED will win about 60% of the time according to MWC, but WHITE is certainly not out of it. I will leave it to others to give us a better analysis using a rollout.

[close]

you may enter the ID's below into GNUbg or Snowie to evaluate your answer yourself:

Position ID: 33YGEAC+ewMBAA  Match ID: MBmgAAAAAAAA

should the image in the poll question (a linked url) disappear, use the attached image to view the board position and/or repair the link

Education is a method whereby one acquires a higher grade of prejudices.
-- Laurence J. Peter US educator & writer (1919 - 1988)

A little inaccuracy sometimes saves tons of explanation. -- Unknown
e-mail me

diane

ok - thanks for doing this...now I can show you why I am for embedding the image in either the poll AND the first post OR the first post.

You can see that before I embedded the image into your first post, when I hit reply - I lost all sight of the board..which is what I want to avoid.

When the post is embedded into the first post - I can see it while I am replying.

Now you can 'see what I see' and I hope this clears that up.

Ok - off to think about the poll  ;)
Never give up on the things that make you smile

stog

Spoiler
bad cube take (no diagram in reply btw)
[close]

stog

yes diagram is visible if u sroll down in reply but then u see unspoilered comments  ;)

diane

Never give up on the things that make you smile

Zorba

Spoiler

A pip count is essential here. Fortunately this one is not too hard to do manually, using the point numbers and flip if you count White's.

Red from home to outfield: 10+9+16+10+11=56 pips
White from home to outfield (flip point numbers top/bottom): 5+4+6+12+12+15=54 pips

So the race looks about equal here; with Red on roll you can subtract 4 pips and see that Red's slightly ahead then: 52 to 54. In itself this is not good enough to double.

However, there's a 11/36 (almost 1/3) chance to hit White and add 10 pips to his race count right away, which would make it an instant pass.

Furthermore, White's bearoff formation is much worse than Red's with 5 checkers stacked on the acepoint and an open fivepoint with two checkers on 6. These two liabilities also combine to make it worse: a 5 in the bearoff has to be played 6/1, stacking yet another checker on the already overloaded acepoint.

Three checkers on the fourpoint, with only two on the 3pt and 2pt is also a small liability.

Then there's the fact that White has longer to go in the outfield before he can start bearing off; this is also a downside, as can be seen f.i. when White rolls a 2-2 next turn: instead of a good roll that takes three checkers off, it's just average, moving 8 pips closer to home.

All in all, you can add quite a few pips to White's pipcount to adjust for all this. Using the Keith Count wastage method (see http://bkgm.com/glossary.html#K):

4*2 + 1 + 1 = 10 pips extra for White; 4+1 = 5 pips extra for Red. I'll add in an extra 2 pips penalty for White's extra crossover and general bearoff ugliness, so overall, we get the following adjusted pipcount:

Red 61, White 66: Red's is like 8% ahead and on roll. Close to a double just on the race alone.

Add the almost 1/3 chance at gaining an immediate 10 pips in the race, and potentially (much) more when White dances, and this must be a very clear double. The question is the take.

How does it look? Let's do a rough calculation:

1/3 White gets hit, 10 pips setback, facing a fourpoint board to enter, and very small hitting chances. Let's give white 12% winning chances from here.
2/3 White doesn't get hit, and he's like 8% behind in a rather short race. Give White 30% here.

Overall that gives: (12+30+30)/3 = 23.3% winning chances for White. Some freak gammons for Red from here and it looks like a pretty close call, but most likely a small take.

The double is crystal clear though and a huge blunder not to ship it here. Let your opponent pay to see the end of this game!
[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

sixty_something

that's a great analysis, Zorbro
Spoiler
BTW, i didn't include the pip count on purpose .. it is so close and reasonably easy to "feel", if not calculate .. like you, for me the decision to cube by RED is almost automatic given the opportunity to hit WHITE and potentially take a two and a half roll lead in the bear off
[close]

diane and stog, i was a couple of hours (or more) late in attaching the image to my first post .. maybe that was why it appears to have not been visible and then seems to have appeared .. i agree, diane, that an image MUST be attached to that original post whether or not it is embedded in the poll question itself

all things considered, i still like the embedded image in the poll question .. i think it makes for much improved readability .. IMHO, the added hassles for the creator are not too difficult to overcome -- but all that is a different subject for a different place ;)
A little inaccuracy sometimes saves tons of explanation. -- Unknown
e-mail me

stog

i still see no pic when i click reply -- unless i scro;; down - which then means i see un-spoilered posts
it was a good analysis z. ---
firefox s'snap attached

sixty_something

not seeing an image while posting a reply is just a "feature" of our interface isn't it? i never suggested that "feature" could be fixed, but now i better understand what you all are talkkng about

for me, the improved readability is all about the embedded board image visibilty WHILE answering the poll question and reviewing the votes .. having the image immediately available and visible there is ALL THERE IS to my suggestions .. it is a very simple and intuitive approach to asking and answering questions about a board position .. but here we go digressing again .. so, let's please move this discussion to On the online match format .. i will cross link to this thread there
A little inaccuracy sometimes saves tons of explanation. -- Unknown
e-mail me

RickrInSF

i would just settle for an image as clear as that one!

diane

Quote from: sixty_something on June 26, 2010, 02:13:57 PM
not seeing an image while posting a reply is just a "feature" of our interface isn't it? i never suggested that "feature" could be fixed, but now i better understand what you all are talkkng about

for me, the improved readability is all about the embedded board image visibilty WHILE answering the poll question and reviewing the votes .. having the image immediately available and visible there is ALL THERE IS to my suggestions .. it is a very simple and intuitive approach to asking and answering questions about a board position .. but here we go digressing again .. so, let's please move this discussion to On the online match format .. i will cross link to this thread there

I am trying really hard to be helpful and not arbitrarily against change here...but please...what does all this mean???

Quotenot seeing an image while posting a reply is just a "feature" of our interface isn't it?
NO

Quotefor me, the improved readability is all about the embedded board image visibilty WHILE answering the poll question and reviewing the votes .. having the image immediately available and visible there is ALL THERE IS to my suggestions
ok - so where do you suggest, or endorse, or actually do...putting the embedded image into the first post - which is the only place it can go to see WHILE you are replying.

and if this is digression move it - but please answer it.
Never give up on the things that make you smile

sixty_something

np, diane .. i appreciate the feedback .. i just worry about boring the socks off others .. for those of us involved in this conversation, i worry about beating a dead horse .. but here goes one more attempt at getting this another step down the road

the issue you've raised regarding not being able to "see" an image at the time of a REPLY is one i had never even considered before .. as i said, it had just become an accepted "feature" to me .. so, once again it seems, we appear to be talking about the same thing, but somehow misfiring on communication

it seems clear to me now that we have three places an image can occur:

1)  IMAGE ATTACHED as a .jpg or .png file to the original post associated with the poll question
2)  IMAGE EMBEDDED WITHIN POLL QUESTION itself as i have done with this problem
3)  IMAGE EMBEDDED WITHIN ORIGINAL POST , as you, diane, have apparently done below (which, btw, i just discovered)

other than using HTML coding, which stog says we cannot do, i can't think of any other way to display the board image ..  so, let's look at the three above in some detail

we have traditionaly resorted to 1) IMAGE ATTACHED for most problems and forum matches ,, more often than not, this has been the only way an image has been made avaialbe .. i objected to it because of having to scroll down through posts (often without spoilers) because i see posts in LIFO order, a different order than you and apparently most others

additionally, the step of enlarging the attached image so i could actually read it is almost always required .. on occassion that step can take up to two minutes on my machine .. i have no clue why it takes so long, but on occasion it does .. most images enlarge in place almost immediately when i double click them,  but some require opening MS Photo Editor and then the file itself .. two minutes is an eternity when i just want to see the damn board to answer the poll question .. thus, my continuing frustration with 1)

on the plus side for 3), is that the file is permanently attached to the post and resident on the FIBS server .. that image will never disappear

the workaround i continuously have proposed is 2) which is clearly demonstrated in this problem .. IMO, it greatly enhances readibility .. yes, it takes a little extra effort on the part of the creator of the poll, but it is, i assert, doable and worth the effort to learn and use .. a drawback to 2) is that it requires a web based server like flckr, which i use .. thus, the image is subject to being temporarily unavailable or even lost .. it is the latter reason why i have consistently also attached a JPG to the original post .. thus, in the past i have used 1) and 2) together

now, with 3) and your example of embedding the board image in the original post itself we have a third option with its own limitations, ie. it requires scrolling down through more recent posts (now in a LIFO order for everyone) to see when composing a reply .. welcome to my world stog -- your comments below are indicative of the problem i have had since day one looking for the attached image in an original post .. 3) also requires access to a web based image server which means it has the same disadvantages as 2)

hopefully i have sumarized the three situations correctly .. it is apparent to me that NONE ARE BETTER THAN THE OTHERS .. nor do any stand ALONE .. some are simpler, some more elegant, but all are limited

so, why not compromise and give up the search for a perfect solution, as it appears there is none?

WHY NOT USE ALL THREE?

USING ALL THREE BOARD IMAGES is my suggestion or endorsement

when i create my next problem post (soon), i will actually do it to see how it looks
A little inaccuracy sometimes saves tons of explanation. -- Unknown
e-mail me

diane

Quote from: sixty_something on June 27, 2010, 02:20:48 PM
on the plus side for 3), is that the file is permanently attached to the post and resident on the FIBS server .. that image will never disappear

a drawback to 2) is that it requires a web based server like flckr, which i use .. thus, the image is subject to being temporarily unavailable or even lost ..

It seems you are missing one point from these remarks.  Don't put the image in FLKR, put it in the gallery here, that way it is available for as long as this board is here.  In the other experiment, I have demonstrated it is possible, and possibly easier than putting it somewhere else.

Give it a try.
Never give up on the things that make you smile

sixty_something

i have noticed that, diane .. i just haven't done it yet .. i just used flckr as an example most folks could relate to .. maintaining all board images at the FIBSboard server may well be the best place for all this and Gb of disk space is relatively cheap, i suppose .. i have been advocating that for a long time .. i was unaware until your posts how much work stog had done here to give us access to storing images .. nonetheless, i am sure it all adds to maintenance and backup issues for stog, but i trust he will let us know when that becomes and issue

so, thanks for continuing to push the dialog and trying new things , diane .. even the angels can do no more
A little inaccuracy sometimes saves tons of explanation. -- Unknown
e-mail me

stog

#14
there may be issues in using the 'gallery' for the fibsboard matches - please see the 'embedding images' thread in the moderator board

and see example here, which uses the attached pic - simply copied to pollheader http://www.fibsboard.com/index.php?topic=3236.0

ah_clem

Spoiler

I count 56 pips for red and 54 pips for white.  If this were a straight race, no way is it a double.  But it's not a straight race - red has a ~30% chance to whack a blot and win that way.  This should be enough to make it a double without having to do any calculations, but for the sake of the argument, here they are anyway:

Red is on roll, so that's worth about 60%.  Then add 2 percent per pip lead and we're at 56% to win the straight race.  So if Red misses he's ~56% to win.  He misses about 70% of the time, so his racing wins amount to ~39%. 

What about the 30% that hit?  Well, he doesn't win them all, but he wins most.  It looks smilar enough to a well timed ace point game that I'll give white about 25% here.  So Red wins another 30 x .75 or 22.5% this way.    Adding up the two, we get about 62% to win, which usually isn't enough to double but 30% of the rolls are market losers.  Hence, double away.

Doing the math, the double is closer than it seemed at first glance, but I still think it's correct.  For one thing, the "math" above is not completely accurate and misses many subtleties.  For instance, if it turns into a race, red will start taking checkers off immediately while white has a couple of rolls before beginning the bearoff.  White's stack on the ace point will cause wastage so a straight pipcount overestimates White's chances.  Then there's the gap on the five.  Red's board looks much better bearoff-wise.  These three observations tip it in favor of the double.

As for the take, 25% to win if hit together with forty-something chance to win when not hit adds up to more than the ~25% to win necessary to accept the cube.



[close]

ah_clem

Spoiler


Quote from: Zorba on June 26, 2010, 01:15:34 PM

White gets hit, 10 pips setback, facing a fourpoint board to enter, and very small hitting chances. Let's give white 12% winning chances from here.

Interesting.  I had pegged White's chances after being hit at about 25%, similar to a well-timed ace point game. You estimate it as less than half that. 

Where did I go wrong?  Let me see if I can work it out myself: 
1) the hitting chances in an ace-point game are higher than here - not many rolls for red that force a blot. 
2) Many rolls force White to break contact.
3) Once White comes off the bar, it's not well-timed anymore

Add them together and it can't possibly be 25%.  I'm still not sure how to arrive at twelve percent though, and this part of the equity is key to the take/pass decision.

[close]