User ID:

Remember me
Lost password?

Thursday Thorns
When Corks Go Bad
           —Random Chance or Big-Time Problem?

By Charles Olken

I was an Econ major in college, and an integral part of the requirements for studying with the learned professors of supply, demand, disposable income, Fed policy and Keynesian theory was a thorough grounding in the vagaries of statistical analysis.

It has been said of that latter topic that there are “lies, damn lies and statistics”, and if you have been following the so-called Budget and Debt Ceiling battles back in that great wastepaper basket in the east we call Washington, D. C., you have been inundated with all three categories of numerical bomb-throwing.

So, when I found myself reaching for my old “Stat 101” textbook tonight, I had to first pass through a certain period of disbelieve. Yet, upon pinching myself, it became clear that I did want to look back at those old measures of random chance.

It all has to do with wine, of course, or in this case, with corked wines—those bottles who corks impart a musty, sometimes sharp, oyster shell quality to a given wine and pretty much rob it of its primary character.

Here is what happened. Here at CGCW, we pull about 400 corks every month. Back about fifteen or twenty years ago, the incidence of “corked” wines ran close to 5% by our measure—sometimes higher and sometimes lower. And the cork supply industry finally wised up and started looking for ways to correct the problem. If for no other reason, we can thank the screw-cap and plastic imitations of cork for pushing the cork folks in the right direction.

For a few years now, maybe six or so, we have been finding “corked wine” incidences running somewhere near one per cent with some months a little higher. But, just recently, the percentages have seemed to jump back towards two to three per cent of bottles tasted. And we find ourselves on edge again about the problem.

The question that has us running to our old texts on random chance is this: In a world in which things happen at a one per cent rate of samples tested, does a short term increase to two per cent or higher represent the limits of random chance or are we looking at a return to the bad old days? For the moment, we are giving cork the benefit of the doubt. But we have been put on notice, and if we are looking at a negative trend, it will become clear over time.

Cork, you made me love you. Don’t let me down.


The CGCW Experience - Take the Tour

Meet the New CGCW

For thirty-five years, Connoisseurs’ Guide has been the authoritative voice of the California wine consumer. With readers in all fifty states and twenty foreign countries, the Guide is valued by wine lovers everywhere for its honesty and for it strong adherence to the principles of transparency, unbiased, hard-hitting opinions. Now, it is becoming the California winelover’s most powerful online voice as well. And, our new features provide an unmatched array of advice and information for aficionados of every stripe.


Statisical significance
by Martin S. Frant
Posted on:10/27/2013 1:14:56 PM

I told my son that your statistics looked like the classical case of soldiers who died in the Prussian Army from horsekicks. 

He replied as follows:

I think for a large n, this can be approximated pretty well by a normal distribution. Then we have a really useful formula: std dev of the proportion = SQRT[ p(1-p)/n] = SQRT[(.01)(.99)/400] ~ (.1)/20 = .005 . So yeah, a shift from p=.01 to p=.02 is just about significant  at the 95% level (2 s.d.), and a shift to .03, way past the 99.9% level (4 s.d.).




Three Sigma Limits
by Charlie Olken
Posted on:10/27/2013 3:13:51 PM

Martin, that is  pretty smart third-grader you have.

I am not convinced, however, that the seeming increase in bad corks is not just the result of random chance, and that results that may seem beyond three sigma limits in any one month are sufficient to conclude that the cork industry is letting us down again.

And Martin, I used to be a part-time academic several decades ago, and I apologize for acting like one here when clearly I have not gone near this kind of stuff for eons.

Leave a comment below, but please limit your comments to 1,200 characters or less. We find it helpful to make a copy of our comments to be sure that they fit. In that way, you can edit them if they run long.

(Please note: your e-mail address will not be visible after posting)



Note: Refresh your browser to see your latest comments.

Having technical problems with the comment system? Click here.