I had been debating what word to use for the end of the title. Anomaly is clearly polite for just the sake of politeness. Debacle seems stronger than I’m willing to defend in a short posting, though would be happy to push such an interpretation in person. “Business” seems to be evasive. So mistake it is.
Matt Strassler now has significant details posted on a blog entry of precisely what happened.
This makes it clear that high-level issues such as the data analysis and the like were given a lot of thought, but that there was little or no low-level hardware examination of the experiment either before the announcement or in the wake of the excitement of the announcement. To me, as an experimentalist, this is very hard to understand. If I have an instrument which has made a shocking discovery (one that I and everyone else working on the project suspects is wrong, but have some slight hope that might be right) I’m going to want to tear the whole thing down and touch every piece, rebuild it, remake every connection, redo every calibration, and in the process possibly find (or just repair) the problem. This didn’t happen within the OPERA group.
The problem was eventually tracked to the now-infamous loose fiber optical connector and the data are understood to now make sense. Anyway, do read the account of the error and its discovery…
