Quote:
Originally Posted by AeroMcAeroFace
If the margin of error is greater than the difference between control results and test results you cannot draw any conclusions other than better testing is necessary to get statistically significant results.
|
Yes, that's a better way of putting it than my explanations.
Measure a baseline, ensure that baseline is repeatable, then start doing some testing, realising that if the test results have less variation from the baseline than the variation in the baseline itself, the results cannot be trusted.
Also, if I have the measurements understood correctly, all the interim measurements (are they instantaneous values?) are largely pointless. Instead, do the average energy consumption over an entire course (I assume the car can calculate that?). That's what we did with a Tesla Model 3 over a 12km loop, windows up and then windows down.