I work with sensitive instruments, though nothing like what these guys probably use, so I have some idea of the challenges associated with calibration, vibration, temperature and humidity changes, static electricity, etc., etc.
It is not nearly as simple as most people think to make extraordinarily precise accurate measurements. In fact, most digital devices “lie” to people.
For instance, I have a digital hygrometer (relative humidity) that reads out in 0.1% increments. Which leads the average person to think it’s accurate to 0.1%.
In fact, it is actually only accurate within a plus or minus 2% range. And that’s from about 20% to 80%. Above or below that middle range the accuracy drops off quickly. And the 2% assumes recent calibration.
I believe the same is true of many instruments.
On our side we used the scales to weigh samples of mail (usually 10 pieces) which would give us the factor necessary to estimate the accuracy of the mailer's count.
These scales were responsible for verifying the validity of weights and piece counts on about $40 billions in postage per year. 20 years of that would come to just under $1 trillion.
The two issues for us were the precision of the reading (that is, how many hundreds of millions of current readings would be made) in how much time ~ and how to protect the scale platform from pressure differentials created by breathing and, to put it bluntly, ausfahrting!
The solution was simple ~ after so many hundred millions of measurements of current you just cut it off ~ just like that ~ and take what you get. During the act of weighing the sample, the acceleration of the mass of the sample pieces exceeded the pressure differentials from breathing, and ausfahrting ~ but ausfahrting could be controlled by telling the clerk "face the readout" while weighing the samples ~ that way, although the scale was at roughly tail high, the direction of any emissions would be away from the scale.
Differences would be obscured over the course of doing hundreds of samples per tour ~ with little deviation from the expected values. Aggregate revenues would be protected.
The answer is that precision measurements are accurate only to the extent required by the process to achieve satori in real time. This applies to all systems.