The second infn school of statistics took place this week in the nice "green island" of Ischia, in the gulf of Naples, Italy. Organized by the INFN section of Naples, the school aims at training Ph.D. students and post-graduates in the foundations and the applications of the statistical methods most used nowadays in particle physics, nuclear physics, and astrophysics.

As one of the members of the scientific advisory committee (together with Luca Lista, who took care of the local organization with the local organizing committee, and Riccardo Faccini and Giovanni Punzi), I took part in the definition of the program and the identification of the most suitable lecturers. Although the previous edition (2013, in Vietri) had already been quite good, we had spotted some points where improvements could be obtained, and I think we succeeded.

I believe the level of the material was appropriate to the background of the students, and the lessons were quite informative and clear. This of course is ultimately the merit of the lecturers themselves. These were Roger Barlow, Glen Cowan, Jan Conrad, Helge Voss, and Mario Pelliccioni.

Without giving much detail, here are the contents of the lectures: Barlow gave an introduction to probability theory and discussed the most important distribution functions, which are commonly encountered in data analysis. Cowan discussed estimators, error propagation, and the method of maximum likelihood. Conrad focused on interval estimation, in particular in the presence of nuisance parameters. Voss gave an overview of multivariate algorithms and their applications. Pelliccioni discussed software tools commonly used for limit-setting and measurements at the LHC.

All lecturers also participated in interactive sessions with the 75 students, where they offered simple problems and applications related to the material they covered during their frontal lessons. I believe this is an important part of any course of statistics, as there is a gap between knowing something in statistics and being able to apply it successfully to a specific problem. Hopefully this gap was bridged by the hands-on sessions.

The venue of the school was the Hotel Continental in Ischia. The location was good and so was the food - too much of a good thing, in my opinion. I had to fight with myself to avoid having to register my belly as excess luggage on my flight back. But maybe this is due to my compulsive behaviour when confronted with quality food free of charge...

The afternoon of Wednesday was spent touring the island with a boat. The weather was good but the air was not too warm, so when the boat stopped some 150 meters from the beach of Sorgeto, well-known for its hot-water springs, I decided not to join some 30 students who plunged in the water to reach the beach. That was a wise decision on my part, judging by the proceedings...

The issue was that to reach the beach, the students had to surpass a set of rocks which looked flattish but were covered with a peculiar species of clams, endowed with very sharp edges. As the sea was not perfectly calm, almost none of the students were able to pass the rocks without bruises and cuts. On their coming back, a few of them required the help of others, and were dragged aboard with life buoys drawn by ropes from the boat.

Above: distressed students are helped returning to the boat

Despite the incident, the tour of the island could be completed without too much trouble, and nobody was hurt seriously enough to require a doctor. But this be a warning to those of you out there who will participate in the "Top 2015" conference or the CMS Week this coming September: you will also probably end up touring the island, and if you decide to swim to the Sorgeto springs, you've better take a detour and aim to the left of the spring, where no rocks are present...