10 Q(without A)s on Bottle Variation
There are no great wines, only great bottles of wine
On the 12th of December 2016 Jancis Robinson, OBE, MW posted this on twitter:
Stating "2 halves Climens '88 side by side in original wooden case."
The post -showing two halves of sweet wine from Barsac featuring a golden-like colour, the first, and an amber-like one, the second- ended with the question "Who still doubts bottle variation?".
The answer is: "Me."
And that's the only answer you will find in this piece. The remainder of this article is just a series of open questions. Therefore, if you are looking for answers, my advice is that you do not waste your time in the further reading of this article. Probably the advice remains still valid if you are looking for GOOD questions.
First of all, I would like to start with two brief prefaces.
P1. I am not doubting the existence of bottle variation, but its nature. I will try to explain later why.
P2. I found -pedagogically speaking- Robinson's question quite dangerous. This is my personal opinion, of course. But the way my poor brain perceive the question is: "Who still dare to doubt bottle variation." This is because the question is coming in a form of challenge to doubt, and is coming from a long established and well respected expert who, moreover, brings photographic experimental evidence that would discourage anyone from undertaking the venture. Oh dear brain! I can picture you like a Galileo in front of the Supreme Sacred Congregation of the Roman and Universal Inquisition or like a Minòs solemnly silenced by a "Perché pur gride? Non impedir lo suo fatale andare: vuolsi così colà dove si puote ciò che si vuole, e più non dimandare." No chance you will accept the challenge. You will swallow the sentence, processing it as a given truth and, in turn, manufacturing one of the many ipse dixit bricks that perilously make part of our knowledge's architecture.
And here my questions.
Q1. What sort of variation the two Barsac bottles do represent? Is the posted photo representing elements of a quite wide statistical distribution (Fig.1a) or a very narrow statistical distribution featuring few spurious outliers (Fig.1b)? For example, How do the other bottles in the case look like?
This is important to know, because the case depicted in Fig.1a is representative of intrinsic statistical variability, which we need to accept and live with, whilst the case in Fig.1b is representative of a case of a handful of defective bottles that are ruining an otherwise narrow statistical dispersion.
Q2. In Fig.1, the ordinate F is the well defined occurrence frequency, i.e. how many times we record a bottle having that particular x value. In a normal distribution, we record many bottles around the average value and a decreasing number of bottles as we move far away from the average. The question here is: what the abscissa x represents? Any measurable quantity such as colour intensity or hue, oxygen %, free SO2 %, ... or any hedonistic quantity such as freshness, complexity, likeability. In the first case, we will have an "universal" distribution for that given wine; in the second case, we will have a different distribution for each judge. Is there any statistical filtering technique we could apply to the latter to create a link with the former? The eternal war between objectivity and subjectivity in wine evaluation, which authors such as Tim Crane, Jamie Goode, Roger Scruton, Barry C. Smith and many others have tried to tackle in their works.
Q3. What does it mean "wide" or "narrow" distribution. How we decide whether a given variation is too much, or is acceptable? How we can even decide what should be the "average" or "standard" behaviour? Do we compare it relatively to previous vintages (vertical meter) or relatively to peers (horizontal meter)? Is our reference a fixed point? Does Barsac's terroir give an immutable and unique mark to its wines or the human intervention is more important than nature? Do human styles and tastes vary also with time? Are we just trying to deal with a moving reference in space and time, which leads us in the disconcerting darkness of the most relativistic abyss? I'm getting headache! :-)
Q4. Which variability sources give rise to this statistical dispersion? I can try to sketch a pseudo-answer, just to open further the question.
- Variations in the container (bottle+cork)
- Variations in the content (wine chemical composition)
- Variations in the boundary conditions (temperature, humidity, vibration, light, cosmic rays...)
- Variations in the measurement instruments (analytic or human)
All these four variations can be categorised in two macro classes:
- Intrinsic Statistical Variations (ISV): e.g. the number of holes in a cork (porosity) follows a Poisson statistics, whilst their spatial position follows a Uniform distribution. This kind of variability is due to the intrinsically discrete nature of matter, charges and light, and it cannot be reduced by improving the accuracy of our manufacturing processes.
- Process Statistical Variations (PSV): e.g. variation in the diameter of a cork. This kind of variations are due to the tolerance and accuracy of our manufacturing process. It can be reduced if we operate with tighter processing margins.
ISV exhibits high spatial frequency, which means we observe fast variation moving from one sample to another, even inside the same lot (Fig.2a). For example, the number of holes in a slab of cork varies statistically (Poisson) centimetre after centimetre along the slab.
PSV exhibits, instead, low spatial frequency, which means we observe good uniformity inside the same lot, whilst we can have variations from lot to lot (Fig.2b). For example, the amount of SO2 is typically quite uniform in a small vat, but can vary from vat to vat.
We can even complicate things more by considering combined variability sources, like the case of the interface between bottlenecks and corks (which seems to be one of the major causes of premox for traditional closures - see this interesting article for further details).
Do Mrs Robinson's bottles come from different lots, even if they were in the same OWC? Does the prematurely aged bottle show any sign of processing defects in the content or in the container?
Q5. This is partially linked to Q3. Can we define a "Universal" ageing curve? Why we perceive immediately the odd bottle is the amber one instead of the golden one? Even if the photo does not show us the other bottles, we are intrinsically assuming that the golden bottle is our reference. Who is telling us that the golden bottle is following the perfect ageing path, while the amber is prematurely oxidised? Could not be the opposite, with the amber following the perfect ageing curve and the golden one maturing too slowly? Is this just because our brain interprets ageing as "a bad thing", therefore the younger looking must be the better one?
Q6. Even provided a Universal curve exists, is this curve valid just in the ideal storage conditions? How does this "Optimal" curve change when we move from the ideal storage conditions? Does it shift rigidly (Fig.3a), or it is deformed in a way that I could never get the quintessential quality peaks of the ideal curve (Fig.3b), or maybe the slower the better (Fig.3c)?
Q7. Does the "Optimal" curve refer to a standard 750ml bottle? How things change for a half format (Mrs Robinson case) or for a Magnum, Jéroboam, Mathusalem, Melchior and so on and so forth. Is it just a myth or is it true that the smaller the bottle the faster the ageing? If that the case, to which curve we adhere with respect to what shown in Fig.3? Will halves always be under-performing?
Q8. What is, then, the Role of Critics when they indicate optimal drinking window or telling us a wine is "too young" or "at peak" or "over the hill". Is this opinion depending on the judge? Is this opinion referring to ideally stored bottles? Is this opinion depending on bottle size? Do wine critics speak just to wine investors? How useful is this for people like me that get pleasure in drinking the damn thing instead of exercising the fine barbarity of economically speculating on it?
Q9. Similar, but opposite to Q2. In Fig.3 the abscissa "time" is well defined as the interval between corking and uncorking the bottle. But what about the ordinate? What we mean for "quality"? Again, can be this objectively and analytically defined? Or is it totally subjective, and if so, ruining our dreams of an universal truth?
Q10. Does Bottle Variation represent a Bad or a Good thing? A dear friend of mine used to joke that bottle variation is wine critics' best friend, the best parachute for their reputation. Of course this is a joke. The serious reality is that Bottle Variation can be perceived as a good thing. Especially for those that drink artisan or natural wine, bottle variation is fully accepted if not even a wanted intrinsic feature of small-lot handcrafting. Even for people that are distant from the natural wine movement, bottle variation could be an added value, a source of interest and diversity. Or maybe not? For sure if you see wine as an industrial product, I think you will consider variability as an unwanted companion and any measure will be taken to standardise processes and maximise accuracy to limit variations. In this case, I'm sure the use of screwcap could remove alone a good chunk of bottle variation around the planet. Or maybe not?
I'm sure each one of you will have a different opinion. But do we really need a single point of view? Can wine not speak multiple languages at the same time? Can wine not catalyse our different opinions, our own variability towards unity? I like to think, every time I pour a glass, to this liquid form of Deus Ex Machina coming down to sort out this human condition, which is perpetuated since the times of the Tower of Babel. It's maybe thanks to its polyhedral cathartic effect that each one of us can bring her/his own diversity to be part of that indefinable wholeness we all yearn for.