So, the mean is 237.5 kJ, and I've tabulated the deviation from this. It looks like the Quarq is well within its +/-2% accuracy spec, but of course, this spans a lot of time and probably many variations in weight and equipment. Oh yeah, I'm neglecting the confidence in the mean and such things that probably matter. I'm basically impressed that a device as rugged as this can give such repeatable data day after day.
OK, enough geeking out for a day.