Nominates Exponent’s “DeflateGate” Study for Research Fraud of the Year 2015 (more for its audacity than its impact)

How Did the U.S. Scientific Community Composed of Millions  of Professionals (including over 800,000 doctors)* Not Catch It?

* (Who Presumably Follow Sports and Dislike/Like the Patriots in Proportions Similar to the General Public)

An athlete of extreme wealth and privilege may not win the sympathy of anyone.  However, if “evidence” can be so blatantly and unethically distorted against such an individual, what can be done to you or those of little to no means when accused of any wrongdoing?

This article does not claim the innocence of any of those accused in an alleged scheme to tamper with game day ball pressures but does claim there is clear evidence of guilt against one party, Exponent.


The premise of the full “DeflateGate” (nickname) report conclusions rely on 5 key points:

1) The Patriots’ balls pressures were underinflated (by whatever cause or force) compared to what is expected
2)  3-4 of the Colts’ balls could adequately serve as the control for the Patriots’ 11 balls in determining (1) though 50% of the data points (pressures) were never recorded
3) The data indicate that the underinflation of the Pats balls were likely due to an intentional deflation
4) This deflation occurred at the hands of Pats employees
5) This intentional deflation likely occurred either at the direction of the Pats QB or with his knowledge

What if the data of the Exponent report does not support (1)?

intentional deflation is one cause of underinflation 
just as
murder is one cause of death 

What if there was no death (or no underinflation), could a murder (intentional deflation) conviction hold?


Using the most forgiving analysis of the work of Exponent, a large number of concerns could be kindly described as numerous serious “flaws” and “weaknesses” which contends is so beyond the scope of accepted research analysis that it likely reflects intentional misrepresentations and

others which are simply obvious distortions.

(1) The report states “Each team provided at least 24 balls . . . ”

The 48+ potential data points pre-game were never recorded.  Exponent essentially fabricated 15 of these data points (11 for the Patriots and 4 for the Colts) for the study, supposedly based on the recall of the officials.  All Patriots’ pre-game balls were assigned 12.5 psi and the Colts’ balls were assigned 13.0 psi pre-game.  Table 3 of Exponent Appendix 2 shows that the Patriots balls were not just assigned the value of 12.5 but actually 12.500 (to the level of 1/1,000 psi) when there was no recorded data.

For halftime, 11 Patriot balls (22 data points) and just 4 Colts balls (8 data points) were measured (psi).
However, Exponent raises serious doubts to the validity of the Colts ball #3 data and eventually drops it in the analyses.  Therefore there are

.                          Pre-game                                                           Half time
Pats                  24+ balls and 24+ unrecorded psis            11 balls; 22 psi measurements*

Colts                 24+ balls and 24+ unrecorded psis             3 balls ; 6 psi measurements*

* 2 gauges used

of the 30 used data points (initially 4 Colts balls and 11 Patriots balls) were included, 15 were never recorded but were manufactured by Exponent with virtually no attention to the essentially complete corruption of any analysis.

Keep in mind that it appears that all parties agree that the Patriots claimed to prefer psis at the very lowest end of acceptable (12.5) and the Colts report a desired psi was around 13.  The HeadSmart Labs analysis (not included in the “Wells Report,” see post below) and even Exponent fully concede that the pressure of the balls were expected to drop during the game.  Again, the Patriots balls supposedly started near 12.5 and were expected to drop during the game which they all did.  They all fell below 12.5 psi and this is well noted in the report: 100% of 22 readings below 12.5 and 100% of 11 balls with 2 readings below the 12.5 reading, which is all expected.

What is not highlighted is that 50% of the Colts ball readings (3 of 6) and 100%  (3 of 3) of readings by one gauge were all below 12.5 at half time.  100% of the 3 usable Colts’ balls had at least one reading below 12.5 at half time  (the most unfavorable spin), but Exponent chooses to report it as “Three of the eight Colts measurements were under 12.5 psig.”  3/8 =  37.5% which is more favorable than 50 or 100%.  They use ball #3 to expand the denominator from 6 to 8 though Exponent suggests that data of ball #3  is corrupted.  They also use the most favorable spin possible: “Each of the Colts balls measured was within the permissible 12.5–13.5 psig range on at least one of the gauges used.” (i.e., 0% were under 12.5) when the worst spin could claim 100% of Colts ball were under 12.5 at the half by at least one gauge.  The equivalent view of the data is not afforded to the analysis of the Patriots’ balls – see (3) below.  Impartial researchers never engage in this type of misrepresentation.

But this is all fairly irrelevant since the balls are expected to drop in pressure.  It is used erroneously in the study to draw  suspicion.  So, 100% of Patriots balls were under 12.5 and 0% of Colts balls under 12.5 per Exponent but 100% of all Patriot and 100% of usable-data Colts balls had at least one measure under 12.5 at the half.  How many readers would realize that though?

In every comparison, the most favorable spin in Colts data is chosen and the least favorable when the Patriots are analyzed.  In at least one case the researchers simply falsify a conclusion about the data presented.

So, a fairer analysis would be to measure the drop in psi to what is expected.  HeadSmart Labs physically did this analysis and determined that the Patriot ball deflations fell within what was expected.

If Exponent’s predictions are to be trusted on this one point, they state (p 40) “these equations predict that the Patriots balls should have measured between 11.52 and 11.32 psig at the end of the first half.”  If one looks at the table on page 68 of “The Wells Report” portion, 11 out of 22 or exactly half of the numbers are above 11.32.  Furthermore 8 out of 11 measurements by one gauge are above 11.32.  Yes, 8 out of 11!  This is not pointed out by the Exponent authors.  In fact they falsely claim that  “Most of the individual Patriots measurements recorded at halftime, however, were lower than the range predicted . . . ”  When does 50% count as most?  (There is the falsification).  Why would they never discuss the potentially most exculpatory data to benefit the Patriots: the 8 out of 11 favorable numbers.  This would be more consistent with HeadSmart Lab’s experiment (see post below).

In other words, 8 out of 11 (72.7%) Patriots’ balls had at least one measurement within the predicted range on at least one gauge.  Remember, it was the Exponent authors who chose to use  “. . . on at least one of the gauges used” as a standard when analyzing the Colts’ balls in (2) above.  So why is this same standard not used when discussing the Patriots’ balls?  Then 11 out of 22 is described as “most” instead of half and the ill-defined “Master Gauge pressures” correction is introduced to cast further doubt.  

One of the 2 gauges were used pre-game on the Patriot balls.  No one knows which.  If it was the presumed higher measuring ‘logo’ gauge, 8 of 11 numbers were within expected with no direct mention by Exponent. If the other had been used it was 3 out of 11, but again no one knows which was used.  If the “at least one” standard (previously applied to the Colts) is used, again 8 out 11 are within expected for the Patriots. Does the distinct possibility of 8/11 within predicted range appear to be evidence of tampering?  Why is this possibility not even mentioned?

Does the data suggest tampering during this game or like HeadSmart Lab’s experiment actually suggest data potentially fairly consistent with the weather and field conditions and the conclusions have been aggressively manipulated to suggest otherwise?  Can guilt be assigned without evidence of a ‘crime?’ gives no credence to Exponent’s adjustments to “Master Gauge pressures” since there is no clearly adequate explanation of this conversion.  Exponent determined the length of the report and chose in their unlimited space and one opportunity to not explain it adequately and it is therefore rejected.

The heavy reliance of Exponent on a greater degree of change (while eliminating the possibly most favorable change data for the Patriots) compared to the Colts as evidence of tampering is academic nonsense.  The researchers are assigning the Colts balls as an unassailable control when there is no justification for a researcher to ever do this.  The Colts balls were assigned a control status but measurements were only attempted on 4 of their balls (3 balls with usable data) at halftime when there were reportedly 24+ Colts balls for this game and 50% of the data were manufactured for the study.
Exponent contemplates the possibility of the non-logo gauge or the logo gauge being used on the Colts or the Patriots but inexcusably does not factor the possibility of one gauge being used on one team pre-game and the other on the other team pre-game.  There is no written game day documentation presented to eliminate this possibility.

the 4 possibilities include (accepting Exponent’s conclusions regarding the logo vs non-logo gauges and the possible switching of gauges between the officials at half time between measuring the team balls)

team               pre-game             half time
Patriots        a   non logo             Blakeman = non logo, Piroleau = logo
.                    b   logo                  Blakeman = non logo, Piroleau = logo

Colts             c   non logo          Blakeman =  logo, Piroleau = non logo
.                     d   logo                     Blakeman =  logo, Piroleau = non logo

Table 7 follows the assumption by Exponent that the officials switched gauges at halftime between the 2 teams’ measurements and their assumption that the lower numbers represent the non logo gauge.  So there  were 2 gauges used on game day with no recording whatsoever of their use.  There is the belief that both gauges were used at half time, and the possibility of both gauges being used pre-game is eliminated, but there is no game day documentation presented to justify this.  The analysis includes only the possibilities that the non logo or the logo gauges were used on both teams pre-game.  Why?

If the omitted possibility of the logo gauge being used on the Patriots pre-game and the non logo gauge on the Colts pre-game the difference in pressure drops between the Patriots and Colts potentially shrinks considerably (1.01 – 0.73 = 0.28), less than half of the 0.66 and 0.68 possibilities Exponent chose to highlight in table 7.  This greater drop in the home team ball pressure is used to draw suspicion of tampering.  But there are 4 possible scenarios, not 2 and no one knows which one of those 4 is correct since there was virtually no acceptable level of documentation of procedures.

The officials are not researchers and certainly should have been provided a research supervisor/advisor on game day to oversee this ad hoc ‘study.’

So  an a/c scenario and a b/d scenario is considered but a/d and b/c scenarios are mysteriously omitted.

Since the officials made no records of the gauges used, there is no reason to consider only 2 of 4 possibilities other than to hide the most favorable data (b/c) for the home team.  A/d is the worst case scenario for the home team.

DISCLOSURE:  The author is an educator in the health sciences who has owned season tickets to an NFL team, which plays many states and 100s of miles away from Massachusetts and the New England area, for almost 2 decades.
The author has traveled once to see another NFL team play: 
to root for the author’s longtime other favorite team, the Colts against the Patriots in Indianapolis.
(c) 2015

Full PDF of


aka “Wells Report” or “DeflateGate” Report

The Exponent portion begins on the 145th page of this PDF
after the table of contents, the text of this portion is between the PDF page 154 – 243

HeadSmart Labs “DeflateGate” Study

“Pittsburgh, PA (January 23, 2015) — HeadSmart™Labs, a Healy Ventures LLC company, today announced findings from a study on football inflation levels in different climates. The study indicated that the pressure in the footballs used in the AFC Championship game could have dropped 1.95 PSI from weather and field conditions alone.”