Superluminal neutrinos - follow-up

The OPERA collaboration, or at least a large subset of it, has a revised preprint out (and apparently submitted somewhere), with more data on their time-of-flight studies of neutrinos produced at CERN. Tomasso has a nice write-up here. Their previous preprint created quite a stir, since it purported to show evidence of neutrino motion faster than c, the speed of light in vacuum. The general reaction among physicists was, that's really weird, and it's exceedingly likely that something is wrong somewhere in the analysis. One complaint that came up repeatedly was that the pulses used by the group were about 10000 nanoseconds long, and the group was arguing about timing at the 60 ns level. You could readily imagine some issues with their statistics or the functioning of the detector that could be a problem here, since the pulses were so long compared to the effect being reported. To deal with this, the group has now been running for a while with much shorter pulses (a few ns in duration). While they don't have nearly as much data so far (in only a few weeks of running), they do have enough to do some analysis, and so far the results are completely consistent with their earlier report. Funky. Clearly pulse duration systematics or statistics aren't the source of the apparent superluminality, then. So, either neutrinos really are superluminal (still bloody unlikely for a host of reasons), or there is still some weird systematic error in the detector somewhere. (For what it's worth, I'm sure they've looked a million ways at the clock synchronization, etc. now, so that's not likely to be the problem either.)

Update:  Matt Strassler has an excellent summary of the situation.