A very good post on this by Matt Strassler from earlier this year:

http://profmattstrassler.com/articles-and-posts/particle-phy...

There were apparently two problems:

1. An improperly calibrated clock 2. A fiber optic cable was not fully plugged in. This led to a little bit of light leaking out from the cable, which lead to a (very small) delay in recognition of the arrival of a pulse on the fiber.

Yes, modern physics experiments are pretty complicated and lots of things can go wrong!