RUSS ROBERTS

READ

The Data in the Wells Report

05.13.15

I’m biased. I’m a Patriots fan. I wanted the Wells Report to exonerate the Patriots, their staff, and their quarterback. It did not. The question remains as to whether it’s accurate.

There are many damning facts in the Wells Report. Texts. Unexpected bathroom adventures. A reference to a needle and a reference to a deflator. Are there innocent explanations for those facts? Maybe.

But it wouldn’t surprise me in the least to discover that Brady encouraged Patriots employees to push the envelope on the low side of the legal psi limit. He apparently likes the ball softer than harder. Maybe they had an “understanding.” No smoking gun, no explicit texts, but winking and nodding. Very possible. Maybe “probable” or “likely” as the Wells report concludes. There’s circumstantial evidence pointing that way. (And if there was any evidence pointing in an opposite direction, the Wells Report chose not provide it.)

But as a data guy, I’m interested in the data the Wells Report provides and whether it reinforces the circumstantial evidence. What I’ve tried to do here is bring out some issues I have not seen discussed elsewhere related to the data. We’re going deep into the weeds.

According to the report, the Patriots footballs before the game were measured to be very close to 12.5 and the Colts very close to 13.0. (The rulebook allows 12.5 to 13.5). At halftime, due to prompting by the Colts, the refs again measured 11 Patriots balls and 4 Colts balls. Here are the measurements of the Patriots balls (from p. 68 of the report):

PatsHalfTime

(Blakeman and Prioleau are two referees. They used separate gauges both brought to the game by Walt Anderson, the head referee. One gauge, what the report calls the Logo Gauge, consistently measured roughly .35 psi higher than the Non-Logo gauge. The Wells Report presumes that Prioleau must have been using the Logo Gauge at halftime. Very reasonable.)

These raw data are where the NFL got into trouble. Remember the very first reports that ESPN leaked from the NFL office? That 11 of 12 footballs were measured to be under-inflated. That the NFL was reportedly furious with the Patriots for cheating. Look at the data. All of the balls are below the legal minimum. Many of the measurements show balls below 11 psi. One is at 10.5 psi. That’s TWO POUNDS of pressure below the minimum. Seems like an egregious violation.

The problem is that as far as I know, no one in the NFL had ever measured the psi of footballs at halftime of a game in January. As a result, they didn’t realize that balls lose pressure when it’s cold outside. As the Wells Report states, the Pats balls, assuming they started the game at 12.5, should actually have been between 11.32 and 11.52 at halftime simply due to weather. The measurements in the second column are mostly in that range. So those balls are mostly in compliance once you take into account the effect of the weather.

(By the way, the investigation was allegedly started by a Colts interception of Brady. D’Qwell Jackson who made the interception reported that the ball felt soft. And indeed, when tested, as it was three times, the measured pressure was 11.45, 11.35, and 11.75. (P. 70 of the report.) You can imagine how the refs must have reacted–the minimum pressure is 12.5 psi! The Patriots are cheating! But the interception occurred an hour or so after kickoff. The physics would predict that even if the ball had started at 12.5, it would fall in pressure after leaving the toasty locker room and being kept on the field for an hour. The intercepted ball actually is evidence on the surface that the Patriots were NOT cheating. And we now know that the NFL had told the referees that the Colts were suspicious of the Patriots even before the game started. I say “on the surface” because the Wells Report doesn’t say when at halftime it was tested. If it was toward the end of the half, it should have warmed considerably from being in the locker room for 13-15 minutes. But the Report makes a big deal about the fact that it was marked with tape, the measurements were written on the tape and the ball was put aside and not reintroduced in the game, presumably because it was presumed to be very damaging evidence against the Patriots. This is consistent with the argument that the refs and the NFL viewed any number under 12.5 as damning when in fact it would depend on how long it had been on the field and in the locker room before being tested.)

So as Mike Florio points out, the real question is which gauge was used to measure the balls before the game. If Walt Anderson used the Logo Gauge (the one that measures high), then the pressure of 8 of the 11 balls measured at half-time is above 11.32 psi which means the under-inflation can be explained by physics. To make it an apples-to-apples comparison, you’d have to add .3 or .4 to the measurements shown above in column 1. Over 70% of the balls fall then fall within the range predicted by the physics and the others are not so far away. That pretty much ruins the case against the Patriots.

Walt Anderson believes he did use the Logo Gauge. According to the Wells Report:

Anderson is certain that he checked the footballs prior to the AFC Championship Game with one of the two gauges that he brought with him to Gillette Stadium. Although Anderson‟s best recollection is that he used the Logo Gauge, he said that it is certainly possible that he used the Non-Logo Gauge.

Well, that’s awkward for the NFL’s case against the Patriots. But the NFL and Exponent (the consulting firm hired by the NFL whose work is found in the 100 page appendix at the end of the Wells Report) are pretty confident that Anderson in fact used the lower-reading Non-Logo Gauge, Here’s why. Exponent tested both gauges against a Master Gauge, a very accurate, well-calibrated gauge. They found that the Non-Logo Gauge was very close to measuring the true pressure. The Logo Gauge measured .3-.4 psi too high. That means that if Anderson had in fact used the Logo Gauge, the Patriots actually submitted balls to Anderson that were at 12.1 psi instead of 12.5 psi. Could the Pats have an inaccurate gauge, also? That would be unlikely argues Exponent. But what clinches the argument according to Exponent is the Colts balls. According to Anderson’s pre-game measurement, they were all around 13.0. The Wells Report interviewed the Colts ball boys who said their goal was 13.0. And that’s what Anderson found when he tested the balls. So obviously Anderson used the accurate Non-Logo Gauge before the game.

Here is how Exponent justifies their conclusion (italics in the original):

According to information provided by Paul, Weiss, personnel from both the Patriots and the Colts recall gauging the footballs for their teams to pressures at or near 12.5 psig and 13.0 psig, respectively, prior to providing the balls to Walt Anderson. Each team used its own gauge to adjust the final pressures before presenting the balls to the referee, who used a gauge different from either used by the two teams to measure the pressure in the footballs. Walt Anderson recalled that according to the gauge he used (which is either the Logo or Non-Logo Gauge), all of the Patriots and Colts footballs measured at or near 12.5 psig and 13.0 psig, respectively, when he first tested them (with two Patriots balls slightly below 12.5 psig). This means that the gauges used by the Patriots and the Colts each read similarly to the gauge used by Walt Anderson during his pre- game inspection.

It has been shown that the Logo Gauge consistently reads higher than all other gauges analyzed in this investigation. As a result, it is very unlikely that the Logo Gauge would have read similarly to the gauges used by each team. Therefore, it is most likely that the gauge used by Walt Anderson prior to the game was the Non-Logo Gauge, which read similarly to the Master Gauge and other gauges tested during the investigation.

That’s very clever. Surely both teams couldn’t both be using gauges that are overestimating the true pressure, right? If that’s right, then the evidence against the Pats is pretty strong.

Except for one problem. There’s no reason to think the Patriots and Colts gauges are accurate. Exponent repeatedly tested the Non-Logo Gauge and the Logo Gauge alongside brand new gauges (called Exemplar Gauges in the report) and they tested all three types of gauges against the Master Gauge which is presumed to be accurate. What they discovered is that while Exemplar Gauges were generally accurate, not only was there a large discrepancy between the Non-Logo and Logo Gauge but that discrepancy grew during the course of the investigation. At the end of the investigation, the Logo Gauge was .75 psi higher than the Master Gauge compared to .35 at the start of the investigation. The Non-Logo Gauge which matched the Master Gauge at the start of the investigation was found to be .05 to .1 psi below the true pressure. So the gap between the two gauges had roughly doubled. (And take a look at Figure 9 in the appendix of the Wells Reports. The Logo and Non-Logo Gauges don’t act anything like the new Exemplar Gauges in responding to changes in ambient temperature. Very weird.) Old gauges are increasingly unreliable.

Exponent points out correctly that it is very unlikely that a similar loss of accuracy occurred on the day of the AFC Championship game. There just wasn’t enough usage of the gauges that day compared to the thousands of tests Exponent ran during the weeks of the investigation.

But what Exponent ignores is that the gauges used by the Patriots and Colts equipment managers may have been used hundreds if not thousands of times in preparation for practice and games over the years. There is no reason to think they were brand new. And there is no reason to think they were accurate even when they are new. Look at Figure 4 in the Appendix of the Wells Report–when the true pressure is 13 psi, brand new gauges of the same model as the Non-Logo Gauge registered a range from 12.5 to 13.2! Even if the Colts tried to get to 13.0 psi, maybe they had a gauge like the Logo Gauge of Walt Anderson. Or maybe they just had an inaccurate one. That clouds the clever explanation for why Anderson must have used the lower-reading Non-Logo Gauge despite his recollection. Gauges are inaccurate to begin with and get funky over time. Who knew? Now we do.

This explains a mystery of the Wells Report. One of the more entertaining parts of the text messages uncovered by the Wells Report comes in the aftermath of the Patriots game against the Jets on October 16th, 2014 when Brady evidently went ballistic because the balls were over-inflated that day. The next day there was the following text between the two Patriots equipment guys at the center of the investigation:

I checked some of the balls this morn… The refs fucked us…a few of then were at almost 16

How was that possible? Wouldn’t the Patriots deliver balls that measured 12.5? Did the refs just add air for spite? How would they end up in the 15 psi range? I think the answer is that the head ref that day had an inaccurate gauge and pumped up the Patriots balls because he thought incorrectly that they were too low.

Here is a telling quote from the Wells Report found in footnote 29:

Anderson believes that he acquired one of his gauges last season, and the other approximately three or four seasons ago.

Exponent was unable to acquire a gauge like the Logo Gauge. (The dozens of new gauges they bought to work with–the Exemplar Gauges–were the same model as the Non-Logo Gauge.) That is probably because that design is no longer being made. The Logo Gauge is an old gauge that Walt Anderson used for a few years and that as a result, it’s no longer accurate, drifting upward in its measurement over the years. It would not be surprising to find that both the Patriots and the Colts use gauges that are not brand new.

So which do you want to believe? Here are your choices:

A: Walt Anderson’s memory is flawed. His pre-game measurements were actually with the Non-Logo gauge, a gauge that was extremely accurate. The Colts and the Patriots also had very accurate gauges that day that were relatively new. The Patriots cheated.

B: Walt Anderson’s memory is accurate. He used the Logo gauge, a gauge that overestimates pressure by quite a bit. The Colts and the Patriots gauges were flawed by the same amount. The Patriots did not cheat.

Taking off my Patriots helmet, I guess I’d say that A is maybe a little more likely than B. But neither is very attractive. The biggest flaw in the Wells investigation is that it didn’t check out the Patriots and Colts gauges. If it’s in the report, I missed it.

What have we learned from the Wells Report? One of the things we have learned is that measuring the pressure of a football is not as straightforward as one might think. The head referee for the game, Walt Anderson, was alerted before the game that there was a possibility that the Patriots might be under-inflating balls. Yet the gauge he brought to the game and that he remembers using, systematically overestimated the pressure in the ball by about .35 psi. This inaccuracy is roughly the amount of tampering the Patriots are accused of. We have learned that temperature and changes in temperature along with moisture have a huge effect on pressure, 2-3 times greater than the magnitude of deflation the Patriots are accused of. And one thing we have learned in the aftermath of the report is that whether the Patriots cheated or not, it does not appear to have much of an effect on the effectiveness of its quarterback. Not only did Brady throw four touchdowns in the second half of the AFC Championship game with balls that the referees inflated to at least 13.0 psi when they discovered that the Pats balls were under-inflated, but other data suggests no difference in Tom Brady’s performance at home or on the road, where opportunities to tamper with the ball would be limited because ball-boys do not travel to away games. On that October 16, 2014 game against the Jets where Brady was furious because the balls were over-inflated, Brady went 20-37 for 261 yards, three touchdowns and no interceptions. His QB rating that day was 103.5. He had a pretty good day at the Super Bowl, too.

Because the NFL had little or no experience with measuring psi in the heat (or cold) of a championship game, it is not surprising that the initial readings, from either gauge brought by Walt Anderson, suggested that the Patriots had been cheating. But a careful review of the measurements should have led them to conclude that the entire process of measuring and complying with the psi regulation was much more complicated than had been previously understood. This is the statement the NFL could have issued based on the Wells investigation:

The circumstances surrounding the balls used in the AFC Championship game of 2015 have taught us that our procedures are inadequate for insuring that all footballs used in the game were within the legal standard of the NFL rulebook. This includes the equipment used by NFL officials to measure pressure and the protocols for insuring that no one tampers with the balls. In light of this, we will be revamping regulations and pre-game protocol to insure that all balls used in NFL contests begin the game within the legal limits of 12.5-13.5 psi.

Despite this uncertainty, there is sufficient evidence to suggest the possibility of foul play on the part of the New England Patriots. They were careless with the supervision of the game balls that day leading to a suspicion that something dishonest took place. Communications among staff members and players suggests the possibility of an attempt to circumvent the rules. But there is no definitive evidence that such circumvention took place. The Patriots will be fined $100,000 and new procedures will be put in place next year to avoid even the appearance of an unfair advantage.

Instead, the NFL decided to tarnish the reputation of a future Hall-of-Famer who some would argue is the greatest player in the history of the NFL. That player is known to even the casual fan as a very intense competitor. I would not be surprised if under the pressure of an impending championship game, he encouraged or allowed staffers to break a rule. It’s a shame that the hard evidence that would make that conclusion definitive is not provided by the Wells Report.

♦♦♦