The Wells report physics argument, based on multiple experiments as well as theoretical modeling, runs as follows.[9]:Appendix 1, 63–68 Several conjectured sources of variability (differences in game use, alleged "vigorous rubbing" by the Patriots before play, leakage during the game, and variations in football volume) can be set aside as they have no discernable effect. Based on documented habit, as well as the recollections of referee Walt Anderson, the Patriots balls were (as usual) set around 12.5 psi, and the Colts balls around 13.0 psi, before their games. The ideal gas law shows that footballs inflated in a warm environment will drop in pressure in a cold environment; however, a football is not a thermos, and the footballs would have rapidly started to re-inflate when taken to the officials' locker room for halftime testing. (Wells estimates that the Patriots balls had 2–4 minutes to repressurize before measurements began; the measurements themselves spanned an estimated 4–5 minutes.[9]:70) Most or all of the Patriots and Colts footballs should have mostly warmed up to room temperature and substantially reinflated by the time each measurement occurred. In addition, there is a minor problem of variability: to give one example, it is hard for a temperature-based theory to explain why Patriots ball #2 tested fully 0.6 psi lower than Patriots ball #1, when both ball #1 and #2 started before the game at the same psi. If temperature is the major factor, the Patriots psi should stay roughly the same, or gradually increase, as subsequent balls are tested; instead, the psi of the footballs changed substantially from one ball to the next. (However, Wells acknowledges that the variability test does not, on its own, reach the level of statistical significance; therefore the conclusion is based on the magnitude of the drop, rather than the variability.)
Besides temperature-based deflation and the timing of the measurements, the condition of a ball's surface (wet vs. dry) also has a small but detectable effect on the measured pressure; there can also be minor measurement error caused by the gauges. During halftime, the referees used two gauges on each ball: the same Non-Logo Gauge that Wells believes to have been used by Anderson before the game to confirm the pre-game pressure, and an additional Logo Gauge. The Logo Gauge appears to consistently run at least 0.35 psi above the (accurately calibrated)[9]:Appendix 1, 45 non-Logo gauge, but both were determined to be extremely consistent and precise. In particular, the Logo gauge is inaccurate (it runs high) but is precise (it consistently runs high by the same amount every time), and therefore can be used as additional confirmation that the non-Logo measurement is correct (with the exception of Colts ball #3, below). Wells believes that Blakeman and Prioleau used the Non-Logo and Logo gauges respectively in the Patriots halftime tests, and that the two of them switched gauges with each other for the Colts halftime test.
