Deflategate, Data and Debate: A Football Scandal as a Learning Opportunity for Winegrowers
WARNING: This post is long, may be offensive to Patriots fans and also includes math.
Sometimes little things in this world drive me unjustifiably mad. It’s like they itch at my soul. Litter left within sight of a garbage can. People wearing too much perfume in a tasting room or nice restaurant. Anything I’ve inadvertently seen or heard about on MTV in the past decade. I’m a Colts fan and so another thing that drives me really crazy is how great the Patriots have been for so long. I’ve been following Deflategate far too closely for my own good. But there’s one thing that really gets my goat: underutilization of information. We, as humans, often make claims and voice opinions that are not informed by readily available data.
There have been some defenders of the Patriots claiming that the change in temperature from indoors – where the footballs were inflated – to outdoors, would have deflated them. I wouldn’t be surprised if Dark Lord Belichick, in all his wisdom, understands this and decided to take advantage of it, feeling that it would not be against the rules to inflate the balls to the minimum level acceptable and let nature take its course. If that were the case, I would say it’s on the officials to deal with that mental juke move, not technically cheating. And with Bad Bill, it’s all about the technicalities – ask Ravens fans.
But the data proves that would be impossible. Before I go on, though, let me just promise you that THIS POST DOES HAVE TO DO WITH WINEGROWING! I’m just taking the scenic route through a place much colder than wine country: Foxborough. So, how do I know that this was no accident? Well, let’s revise that to, “Why do I strongly suspect that this was no accident?” THE DATA!
Brush off your high school chemistry books and look up the formula for the relationship between pressure, volume and temperature: (PV)/T = k. P stands for pressure, V for volume, T for temperature in kelvins and k for a constant. We can remove the constant, since it is not a variable, and instead write this equation as an equilibrium between the variables to yield (PoVo)/To = (PmVm)/Tm. In this equation the lower-case “o” stands for Officials. It indicates that those were the levels of the variables right after the officials weighed the balls. The lower-case “m” stands for Measured or Minions. It indicates that those were the levels of the variables the next time the balls were measured or, alternatively, after Bold Bill Belichick’s minions got to them.
Now, I am addressing only the claim that a temperature shift could have caused the pressure change. I know the Colts would have lost the game even had the game been played with bowling balls. And I know there are many other issues to argue over. I am only addressing the claim that temperature could explain the footballs’ low level of inflation. After all, this is a wine blog, not a football blog.
So here is the data that we have or can impute:
What we’ve heard from the media is that the balls must be inflated to between 12.5 PSI and 13.5 PSI. We’ll assume that they were inflated to the bare, legal minimum.
We’ve also heard that 11 of the 12 balls measured were underinflated by 2 PSI to 10.5 PSI.
We can safely assume that Vo = Vm, since the balls would have had to have been very noticeably deflated for any significant volume change to have occurred.
We can assume that the footballs were inflated at 72 degrees which, according to Google is 295 degrees kelvin. Checking Weather Underground’s site, it looks as though the temperature for the game could have fallen as low as 45 degrees Fahrenheit or 280 degrees kelvin. Both of these temperature assumptions are as generous as a reasonable person could be and better information is likely available to the NFL that would probably shed a worse light on the situation.
With this information, we can solve the formula for Pm to find what this shift in temperature would have done to pressure.
(12.5 * Vo)/295 = (Pm*Vm)/280
11.9*.95*Vo = Pm*Vm
11.3*Vo = Pm * Vm
Now, we know that the volumes stayed constant, so we can continue by cancelling out Vo and Vm, which are equal.
11.3 = Pm
The effect of temperature would have reduced the PSI to 11.3, not to 10.5. Claims to the contrary ignore the data. Of course, one can make a variety of other arguments to defend the Patriots. But the claim that temperature made this change is not borne out by the data available, even with the most generous assumptions made regarding room temperature and game temperature.
Now, the NFL may have even more data available. If these checks are made in every game and any records are kept they can look at the probability that such deflation could occur. Even if officials only record the number of balls that failed to pass muster, that would be enough. If they actually record levels of pressure in the footballs, you’ve got a treasure trove of data to use to figure out the probability that this could have happened without human tampering.
So, what in the world does this have to do with vineyards?!
This same failure to use readily available data occurs on a regular basis across vineyards in California. It’s one of those itches that drives me crazy. We gather so much data in this business – and should probably gather more – and then we fail to use it, instead making decisions based on our gut (also known as conjecture). The list of examples could go on and on, but let me just throw out three, in Q&A format:
How high will my yields be this year?
Growers need to be able to estimate their yields in advance for their own financial planning and, more crucially, so there are no surprises for their buyers. I have heard from many growers who do cluster counts and cluster weights and use a rule of thumb who tell me these measurements never work as well as their own gut-based estimates. That is because we all have different vineyard sites and, unfortunately, different methods of sampling and measuring.
Beyond the procedural fixes of creating consistency, one should come up with his own constants for these calculations. Growers are already keeping track of their cluster counts and weights. Microsoft Excel makes it very easy to track this data year after year and find a correlation between these variables and final yields. That way, you are multiplying your cluster weight by a number that, when multiplied by your cluster count estimate, brings you much closer to actual yields than any rule of thumb or gut feeling. Each year, you produce more data to further refine this number for your specific vineyard. You can do it by vineyard, variety, clone, block, whatever. You can even add in readily available weather data and account for the timing of the sampling.
When will my grapes be ready for harvest?
Once we’re close to harvest this becomes easy to estimate and is heavily influenced by upcoming weather. But we can do a decent job of estimating this in advance, using a spreadsheet and the dates of operation performed on the vineyard, combined with weather data. It is actually pretty easy.
Should I plant this variety or this variety?
I am asked this question often and my typical response is to inform the questioner that I actually make my living answering questions like that. One of the best investments a grower can make is to spend the few thousand dollars necessary to thoroughly answer that question. But, if you’re unwilling to do that, let me exhort you to at least do the following:
Plot the change in acreage for the relevant varieties in all areas relevant to your market;
Plot the change in price on the same graph for the same areas;
Compare them to each other.
If you see, for instance, that prices for, I don’t know, Viognier in the Sierra Nevadas region is rising, but planted acreage is rising at the same rate, you probably want to avoid planting Viognier there. You may ask, “But if prices are rising while supply is rising, doesn’t that indicate that there must be rising demand for the market to accommodate rising yields, without prices levelling off?” No.
Vineyards, as we all know, are non-bearing for three years. In the very common case that farmers are planting to respond to price rises in real-time, the future holds bad news for growers, since supply increases will continue even after planting slows and after a supply-demand equilibrium is reached. That creates a high probability that prices will drop to an extent that could eliminate profitability. This dynamic is even further exaggerated if you are buying land to plant, since the increasing grape prices have undoubtedly pushed land prices up, necessitating greater leverage and greater financing costs to make an acquisition.
Finally, one last note on using data. If you have infrared, aerial photography done of your vineyard, you have a treasure trove of untapped data. Those are not actually photographs but measurements of infrared light that is then converted into numbers. Each number is represented by a different hue or color and converted into a map. All of that raw data is bound to have great predictive applications for watering strategies, Brix, yields and harvest timing. I would love to explore those applications, if anyone reading this is interested and would want to partner up with me. Contact VFA>>>