This is an almost trivial application of special relativity. This is even less than a trivial application of special relativity, it's literally a third grade "

This is an almost trivial application of special relativity. This is even less than a trivial application of special relativity, it's literally a third grade "distance = rate x time" problem: Skipping over the relativistic stuff, which is quite frankly unnecessary (or rather, it's necessary, but the CERN authors apparently already included the corrections), the only two equations of import are (re-named vars for lack of LaTeX): 1) time_baseline = dist_baseline / c 2) time_observed = dist_baseline / (c+v) where v is the speed of the satellite. In other words, if two things are moving towards each other, then you should use the combined velocity to compute the time. Duhhh... Now, the fact that the difference in these formulas turns out to almost exactly explain away the problem is certainly suggestive, I'll be honest about that. But if this was all there was to the story, it would be massively embarrassing for someone at CERN, this is the type of error I mentally ruled out immediately because these people should be way too careful to be making elementary screwups like this. This is the type of error that freshmen lose points for on a mechanics 101 test, not one that eludes dozens of physicists working around the clock for a month. So I'm skeptical here; it's not that the author of this paper is wrong (though there are credibility red flags in the paper, even beyond the fact that he's not a physicist [he's a computer guy], like the fact that he defines gamma as the inverse of the commonly used relativistic quantity, something that indicates only a passing familiarity with the field), it's that I'm not sure that the CERN guys could reasonably be assumed to have made this mistake. Then again, maybe everyone else looking over the work also assumed that they'd dotted their i's, and didn't bother checking the trivial stuff. It's certainly happened before...