Is this a record

[pirate22]

Today with various partners,and opps--- for 2 hours ply IMP

I was not allowed to play a hand-either my pard bid and all passed

or opps over bid and went down---after 2 hours 15 minuites

i gave up with a plus 75 imps no one about.

I have asked who ever is responsible to shuffle the cards

[matmat]

i don't usually do this.

 

you played for 3 hours, not two

you played 31 boards.

you declared 3 hands, not 0.

you finished with +52 imps, not +75 imps.

 

what does this have to do with shuffling? what does this have to do with general bridge discussion? do you really think that after decades of existence of this game this is even close to some sort of a record?

[awm]

I once played a club game (face to face bridge) with a new partner. Out of 26 boards, I declared zero hands. My partner declared 21. We finished with a 55%, and partner somehow decided from this that I was a very good player. Perhaps he liked the way I turned the dummy.

[Free]

It's very easy to break this 'record' by playing very slow.

[1eyedjack]

It is likely that the "+75 IMP" discrepancy arose by taking the figure off the on-line display rather than by a later retrieval from MyHands when all the hands have completed the movement.

[daveharty]

I posted about a similar game I had recently:

 

http://www.bridgebase.com/forums/topic/42355-strange-statistical-anomaly

 

A week or two after that game, I played in a club game where I declared 11 out of 24 boards. Maybe I was overcompensating.

[aguahombre]

It is not a statistical anomaly when my results are inversely proportional to the number of hands I declare :angry:

[peachy]

It is likely that the "+75 IMP" discrepancy arose by taking the figure off the on-line display rather than by a later retrieval from MyHands when all the hands have completed the movement.

 

Could well be. But the hours?

[hrothgar]

No only is this not a record, it doesn't even appear to be that unlikely:

 

From the sounds of things, you played 31 boards.

You ended up declaring 3.

 

For simplicity, I'm going to assume that roughly 1:10 boards get passed out.

I'm going to throw these out of the mix, leaving a dataset in which

 

1. You played 28 boards

2. you declared 3

 

If I remember my math, it would seem appropriate to model this as a binomial distribution with 28 trials and a probabily of success = .25

(All other things being equal, you probably expect that you will declare one out of four hands that don't get passed out. If you're playing a particularly conservative or highly aggressive system, please let me know and I can adjust things accordingly)

 

I'm attaching a PDF for the described distribution. (From the looks of things, there is roughly a 4% chance that this will happen)

 

0.000317479271413

0.002963139866525

0.013334129399361

0.038520818264821

0.080251704718377

0.128402727549403

0.164070151868682

0.171883016243381

0.150397639212958

0.111405658676265

0.070556917161635

0.038485591179074

0.018173751390118

0.007455898006202

0.002662820716501

0.000828433111800

0.000224367301113

0.000052792306144

0.000010753988289

0.000001886664612

0.000000282999692

0.000000035936469

0.000000003811444

0.000000000331430

0.000000000023016

0.000000000001228

0.000000000000047

0.000000000000001

0.000000000000000

[Free]

Don't ever talk statistics with Richard, he'll pwn you. :D

[matmat]

Don't ever talk statistics with Richard, he'll pwn you. :D

 

but he does like his significant digits :)

[hrothgar]

but he does like his significant digits :)

 

Reducing the precision would have required a second line of code...

 

foo = binopdf([0:28], 28, .25)'

 

which translates as

 

Create a variable named foo, which stores the PDF for a binomial distribution with N = 28 and P = .25.

The PDF should be evaluated at integers between 0 and 28.

Transpose the result so its a column vector rather than a row vector

 

<<Got to love having this kind of stuff sitting on the desktop>>

[Cyberyeti]

This sort of thing used to be pretty common in schools bridge where it was not uncommon for one player to be much better than the other, and it was also not illegal (or just never punished) for only one member of the partnership to play transfers.

 

We had a 20 board match where I played something like 15 (and our two at the other table played about 8 each), so partner got either 0 or 1 and the only imps we dropped while piling up well over 100 was 3N-1 in both rooms.

[helene_t]

but he does like his significant digits :)

82.65291% of all statistical surveys overstate their own accuracy.

 

But hey Richard, tabulating a binomial distribution can be done with a spreadsheet, you need to do something more complex for your Matlab advertisment! Hand-hogging by certain players could overdisperse the number of declared hands by a particular player. Then again, some players may think "now it is partner's turn to declare", thereby underdispersing the distribution. How would we model that? Could we construct a model that would allow us to determine how much of the overdispersion is caused by aggressiveness rather than hogging? Aggressive bidders would hog on behalf of their partners as well so it should be possible.

[gwnn]

10% passouts is laughable.

[hrothgar]

10% passouts is laughable.

 

Please feel free to suggest a more appropriate number and redo the calculations...

[aguahombre]

Due to today's agressiveness (opening bids), I will venture a guess of one pass-out per 50 boards at our table. One per 80 at big clubbers' tables.

 

No science involved in this guess.

[Zelandakh]

82.65291% of all statistical surveys overstate their own accuracy.

93.87% of online statistics are made up on the spur of the moment.

[hrothgar]

Due to today's agressiveness (opening bids), I will venture a guess of one pass-out per 50 boards at our table. One per 80 at big clubbers' tables.

 

No science involved in this guess.

 

Anyone have BridgeBrowser handy?

[Bbradley62]

Due to today's agressiveness (opening bids), I will venture a guess of one pass-out per 50 boards at our table. One per 80 at big clubbers' tables.

 

No science involved in this guess.

Only 2 of my 675 BBO hands from the past month were passed out, and I don't open all that aggressively.

[gwnn]

0.052 for 4 out of 31 (I think 0 is the most likely number of passouts).

0.917 for >4

0.0307 for <4

 

http://stattrek.com/Tables/Binomial.aspx

 

0.0803, 0.865 and 0.0551 for 28.

 

Something's odd - why did hrothgar get other numbers? :(

[hrothgar]

0.052 for 4 out of 31 (I think 0 is the most likely number of passouts).

0.917 for >4

0.0307 for <4

 

http://stattrek.com/Tables/Binomial.aspx

 

0.0803, 0.865 and 0.0551 for 28.

 

Something's odd - why did hrothgar get other numbers? :(

 

PDF versus CDF?

 

fwiw, I used the web site to estimate the probabily for

 

N = 28

P = .25

X = 3

 

and it spit out 0.038520818264821

[gwnn]

Oh. I was looking for 4, not 3 declared hands. Why? I don't know.

[hrothgar]

Oh. I was looking for 4, not 3 declared hands. Why? I don't know.

 

Zero's a bitch, don't cha know

[mtvesuvius]

FWIW I played 440 hands on BBO last month, and passed two of them out. One with a robot, one with a regular partner. This is not counting the Passouts in Robot Reward tournaments.