The following is an excerpt from a book I am working on intermittently. I do not know whether the project will ever see the light, and it just occurred to me that I could share a tiny bit of it with you in my blog. Enjoy!

----

My office in the CDF portakamps is the second on the left after “the crossover”, a corridor which connects the two main aisles. The one next to mine was occupied during those years by Fotis Ptohos, who had been a Harvard University graduate student since tempus immemorabile. That was not such a peculiarity: it was generally understood that in Harvard, Ph.D. studies would routinely last ten years or even longer, a fact confirmed by the joke that circulated in the Harvard University High-Energy Physics Laboratory (HUHEPL) in Cambridge, according to which the lab in the basement of the HUHEPL building was full of G-10: no, not the material used for electronic boards: rather, 10^th-year graduate students!

 Fotis was tall and lean, and he had long, light-brown hair which were thinning above the forehead. He carried round glasses over intense blue eyes, and an irregular blondish beard. He was quick with a joke, and I could often hear loud bursts of laughter through the thin wall dividing our offices, any time Paolo Giromini paid him a visit without anything to bitch about. Despite the good-humored nature of Fotis, I had had few chances to interact with him in the past: social life was close to nonexistent for those doomed graduate students of CDF not lucky enough to be part of a large group.

One day, though, their voice appeared louder than usual. They were stationing in front of my door, and they sounded to me like bavarians at the Oktoberfest. I had been spending the afternoon in the attempt to find a bug in a program which was to produce in my intentions a fancy histogram, but was instead returning a core dump, crashing without appeal or explanation. I was frustrated and, with no progress to show for two hours of investigations, I decided to ascribe it to the noise. I thus got out of my door, staring with a reproachful, frowning face at the duo of jokers. This was the first time I spoke to Paolo Giromini.

It took them exactly one minute to get me to join their laughter. Of course, they were having fun of some of our colleagues –in the benign way people from Tuscany are used to discuss anything from their enemies to their siblings; despite being Greek, or maybe because of it, Fotis appeared perfectly at ease with those customs. To me it was a new thing to see a professor of physics letting go in the company of students with comments on his colleagues, their attitudes and their skills, comments which were ranging from inappropriate to mocking. For Paolo it was business as usual: he was just like that – he would joke about anybody. Be it the Pope or Sheldon Glashow, nobody was immune. 

The issue at hand was the infamous "Secvtx scale factor", a number describing the different efficiency of the Secvtx b-tagger when applied to data and simulation. Paolo and Fotis explained to me that they had recomputed it using a much larger Monte Carlo dataset, which had cost Fotis months of unforgiving, repetitive work. With millions of simulated bottom-antibottom production events, the contribution to the total uncertainty due to the Monte Carlo statistics was now much reduced, and the departure from unity of the scale factor  was unquestionable. Less so was the lack of any dependence of its value on the energy of the b-quark tagged jet, but there was objectively no reason to hypothesize it from the plot obtained by the Frascati group. The factor was 1.25, give or take a tenth.

It was the first time I was hearing about the whole issue, but it did not take me long to realize that my attendance of the top group meetings had been too scarce during the last year or so: this was a number which had a large impact in my own analysis!

    “So, let me see if I understand this” I told Paolo. “You mean to say that despite all the tunings and tweakings of our detector simulation, if one looks for a signal yielding b-quark jets in a simulation made with the Herwig Monte Carlo generator, one is going to see 25% less than what one would see in the data, other things being equal ?”

    “Well, 20% less, one divided by 1.25. The factor applies to the data, or its inverse to the simulation. And Herwig or Pythia make no real difference.”

    “Ok, ok. So if my simulation predicts I should see 40 Z boson decays into b-quark jets, I should rather expect to see 50, correct ?”

    “No, no. That would happen only if your selection required one Secvtx b-tagged jet in the event.”

    “But my selection requires both jets to be b-tagged with Secvtx! Wait... Am I underestimating my signal by 1.25 squared then ?”

    “Well, yes, correct. There is no correlation between the scale factor of two b-tags in an event. We see no indication that there be any.”

    “But then my Monte Carlo prediction becomes .... 1.25 squared... Hmm 1.2 squared is 1.44, plus 0.12... 1.56 times 40... 63 events! Sixty-three! It agrees with the data!”

     “Aspetta, what’s the excitement ? Are you seeing Z decays in the data ?”

    “Sure, I have isolated a sample where I have an excess of 91 events, giveor take 20. Now 91 was really off from 40, but it is less than 1.5 standard deviations away from 63, the number I compute if I take into account your scale factor! Thank you!!”

    “Ah! Bravo guaglione.”

    “So you appear to have an independent confirmation of the scale factor then. Please explain how you find the signal. Do you have a plot ?”