Spreadsheet: Actual vs Expected PSI Drop Based on Wells Report Data

[upstater1]

This weekend we played official soccer games with a $175 ball. One of the players messing with it on the sideline noticed there was a wobble in it. We kind of pressed it with our hands on all sides and detected a weird shaped bladder (ie pressing on one side, almost like a baby in a pregnant woman) and it was soft elsewhere. i.e. thing was oblong.

Balls for sports are not scientific instruments. It baffles me that we are having a nationwide (worldwide, the Guardian has picked this up!) conversation about PSI and gas laws and temp drops when it comes to expensive balls that have a great deal of variation.

I can totally understand why QBs would reject some balls.

These things are not constructed with tight specifications.

It's a pigskin for heaven's sake!!

 

[Palm Beach Pats Fan]

Can you please send this to Donald Yee?

Done.

I have also e-mailed it to Peter King, FWIW

 

[Froob]

Their ideal gas law calculations match mine PRECISELY.

A football should drop in psi by 1.13 psi based on temperature, with their temp assumptions.

One gauge has pats ball dropping 1.38, one 1.01.

Colts drop was less than 1, violating the laws of the universe. No comments on THAT at all.

Only fair thing to do would be to ban the Colts from time and space. Set a harsh presitent.

 

[Eric Manchester]

Hi guys,

I made a spreadsheet which allows you calculate the difference between the expected and actual PSI drop based on the actual data, allowing you to manipulate assumptions such as the initial reading from each gauge. The justification for my default values are given below.
https://docs.google.com/spreadsheets/d/1QhGMoE80dWZaIFA00XYVDbUEoZ6g0G3CGzF6Muc72yI/edit?usp=sharing

The Wells report was confusing, because of the fact that there were two different gauges, one of which read out values approximately .4 psi lower than the other. We'll call the gauge with the lower read-out Gauge A, and the one with the higher read-out Gauge B. According to the Wells report it is unclear which gauge was used for the initial measurements, but the balls were measured with both gauges at half-time. If you assume that the initial 12.5 starting psi for the Patriots was made with Gauge B, then the initial read-out from Gauge A would have been 12.1 psi. The spreadsheet allows you to change the values for Gauge A and Gauge B, but the starting value for Gauge A should always be 0.4 psi lower than for Gauge B.

The Wells report also stated that given a starting psi of 12.5, the balls would be expected to be between 11.3 and 11.5, or, if Gauge B was used for the initial reading, the balls would have read as being between 10.9 and 11.3 on Gauge A. The spreadsheet allows you to configure the expected psi drop. The default drop is 1 psi, which is on the conservative end of the results from the Wells report.

With these starting conditions, the average ball was 0.002 psi lower than expected, a result that is not statistically significant.

I'd have to go back and check the report, but what you say is fascinating---what if the higher gauge was used originally, and gave a PSI that would have been too low for the other gauge? Then the .3-4 differential would combine with a roughly .5 drop for close to 1 PSI, the amount by which most of them were underinflated.


Two other things I would like to review as well:
1) it said that the PSIs before the game were tested in the shower room area. If showers had been running recently, this would increase heat, which would mean that a drop later would be greater. It might affect both teams' balls the same, BUT we'd have to know, how much time passed between testing (in a few minutes, the room could have cooled considerably), and were they both tested in the shower area?

2) The report said that the artificial drop of the PSI due to outside cold would be gone in 30-45 seconds, so the fact that the Patriots balls were not tested for at least 2 minutes until after they were brought inside supposedly meant the effects of the cold had worn off. However, they also said the Pats tests was completed 4-6 minutes in, and THEN the Colts balls were tested In my mind, if they think there is a significant PSI rise between 30 seconds and 2 minutes, it would seem that the Colts balls might have a chance to rise even more after warming up an additional 3-5 minutes. This might account for why the Patriots PSI drops were roughly double the Colts.' Now, if everything goes back to normal after 45 seconds or so and stays the same, this argument does not hold--BUT what would need to be VERY clear is whether the re-raising of the PSI BEGINS after 30-45 seconds, or is completed. If it only begins, then it continues throughout, and the Colts balls would still be expected to re-rise more, thus once again possibly explaining the discrepancy in the drop figures.

Both of these things, along with many others, should have been considered.

 

[Eric Manchester]

The most revealing part of report is the appendix - in particular pgs 41 onwards

The transient pressure change is key! Balls brought in from the cold to a warm environment increase in pressure quickly. The difference between measuring balls immediately and after 6mins is about 0.5Psi.

Most of the data they are supposing presumes they start measuring the colt balls at 8 mins, hands up who thinks they panicked and measure the Colts balls at 10mins plus.

Also Colts balls as controls? What? Is that pressure machine in any way similar to "normal" football pressures?

Bad science and too many gaps. Hope they nail the NFL with it.

Hi--great, I saw you pointed this out before me, after I posted (the first part, about the PSI difference based on when the halftime measurements were done). This is VERY significant.