Pressure gauge discrepancies undermine Wells report
When the NFL dusted off the Ted Wells bat signal in January for an investigation regarding the question of whether the Patriots tampered with footballs used in the AFC championship game, it was believed that the raw measurements of air pressure inside the footballs taken at halftime would point clearly to tampering. Lost in the text messages between Larry and Curly and the question of whether Tom Brady was the Moe Howard periodically clunking their heads together to ensure that the footballs were suited to his preferences is the possibility that the raw measurements don’t point to tampering.
The erroneous report from ESPN that 10 of 12 balls were 2.0 PSI below the 12.5 PSI minimum cemented the early narrative that the amount of air missing from the footballs clearly suggests tampering. The actual numbers, standing alone, say otherwise.
It’s possible that the actual numbers suggest no tampering at all. Which could be the biggest problem with the 243-page report.
Here’s where we try (key word: try) to take something that’s pretty complicated and make it somewhat understandable.
First, the officials had two pressure gauges available -- and those pressure gauges generated very different measurements.
One gauge had a Wilson logo on the back. The other didn’t. One had an obviously crooked needle. The other didn’t.
The gauge with the Wilson logo and the longer, crooked needle typically generated higher readings, in the range of 0.3 to 0.45 PSI.
The measurements taken at halftime of the AFC title game by the two available gauges demonstrated this reality. Here’s the gap in PSI for each of the 11 Patriots footballs, based on the two gauges: (1) 0.3 PSI; (2) 0.35 PSI; (3) 0.35 PSI; (4) 0.3 PSI; (5) 0.35 PSI; (6) 0.35 PSI; (7) 0.45 PSI; (8) 0.45 PSI; (9) 0.4 PSI; (10) 0.4 PSI; and (11) 0.45 PSI.
Second, referee Walt Anderson doesn’t recall which gauge he used to measure PSI at the start of the game.
The absence of a documentation regarding the air pressure in the Patriots footballs prior to kickoff can be justified by Anderson’s clear recollection that he ensured each ball was set to 12.5 PSI. However, Anderson doesn’t clearly recall whether he used the gauge that generates the higher measurement or the one that generates the lower measurement.
It’s an important point because the gauge used before kickoff determines the starting point for the halftime analysis. If the pressures were set by the gauge with the logo and the long, crooked needle, that’s the gauge that should have been used at halftime. If it was the other gauge that was used before the game, that’s the one that should have been used at halftime.
The Wells report concludes that Anderson used the gauge that generates the lower measurement before kickoff, despite Anderson’s lack of specific recollection as to which gauge he used. The reasoning for the decision to assume Anderson used the gauge without the Wilson logo appears in the paragraph contained at the bottom of page 116 of the report.
Frankly, the explanation doesn’t make much sense. If anyone understands it, please let us know.
Here’s the one thing that does make sense: Without knowing which gauge was used to set the pressures before the game, it’s impossible to know which set of readings taken at halftime is the accurate set of readings, and which set of readings should be thrown out.
Third, knowing the gauge that was used before kickoff is critical to proving tampering.
At page 113, the Wells report states: "[T]he Ideal Gas Law predicts that the Patriots balls should have measured between 11.52 and 11.32 psi at the end of the first half, just before they were brought back into the Officials Locker Room. Most of the individual Patriots measurements recorded at halftime, however, were lower than the range predicted by the Ideal Gas Law.”
As those of you who were visiting PFT frequently in the early days of #DeflateGate may recall, the Ideal Gas Law refers to the formula that determines the changes in gases based on various factors, including but not limited to volume, pressure, and temperature. And the Wells report concludes that all Patriots footballs should have measured between 11.52 and 11.32 PSI at halftime.
But that observation hinges on the question of which gauge was used to set the PSI prior to kickoff. If the gauge that generates the higher numbers was used, the measurements of the Patriots footballs taken by that gauge are mostly consistent with the 11.52-11.32 PSI range at halftime: (1) 11.8; (2) 11.2; (3) 11.5; (4) 11.0; (5) 11.45; (6) 11.95; (7) 12.3; (8) 11.55; (9) 11.35; (10) 10.9; and (11) 11.35.
Based on those readings, three of the footballs were above the predicted range, five were in the predicted range, and three were below the predicted range.
By assuming that the gauge that generates the lower readings was used before the game began, the readings taken by that same gauge at halftime show that one ball was above the predicted range, two were in the predicted range, and eight were below the predicted range. Which is more consistent with the conclusion that some degree of tampering occurred.
So, basically, the scientific proof of tampering hinges on a literal coin flip between the pressure gauge that generated a higher reading and the pressure gauge that generated a lower reading. Apart from the very real problems inherent to the NFL using pressure gauges that generate such dramatically different readings for a key postseason game, the justification used to assume that Walt Anderson used before kickoff the gauge that makes tampering more likely doesn’t feel like the outcome of a scientific experiment. It feels like an effort to work backward to justify a predetermined conclusion.
None of this changes the fact that Larry and Curly exchanged text messages that point to a pattern of tampering during and prior to the 2014 season. None of this changes the fact that Tom Brady a/k/a Moe Howard may have a smoking gun or two lurking somewhere on his cell phone.
Regardless, scientists don’t dabble in probability or inference; even with the 50.1-vs.-49.9 standard of proof that applies both to civil lawsuits and the NFL’s High Court of Cheating, expert witnesses typically must base their opinions on a reasonable degree of scientific certainty. The fact that two gauges were available to Walt Anderson, the fact that those gauges generated such dramatically different readings, and the fact that Anderson specifically doesn’t remember which of the gauges he used to set the pressure prior to kickoff makes it very difficult to conclude with any degree of certainty whether the accurate measurements taken at halftime are the ones that suggest tampering, or the ones that don’t.
While the full scope of the report suggests that something fishy happened, it was believed all along that a conclusion of tampering would be backed up by sound, scientific evidence that tampering occurred. Despite 243-pages of polish, the scientific evidence in this specific case is significantly undermined by the fact that the NFL was using a clearly defective pressure gauge prior to one of its most important football games of the year.
UPDATE 1:25 p.m. ET: As to Anderson’s recollection regarding the gauge used to set the PSI levels before kickoff, his “best recollection” is that he used the gauge with the logo, which generates higher numbers. But he conceded that it’s “certainly possible” he used the other gauge, and the Wells report concludes that he did.