(Continuing from: The Gaussian Episode 1)
I therefore considered myself lucky... I practically had one of the best student of my course with me, someone I enjoyed working with, serious but capable of laughing and joking, who also lived quite close to my house, and I had one of the best calculators on the market …what could go wrong?
Pietro and I got to work, we dedicated time to testing the program for calculating the coefficients of the Gaussian with invented numbers, the time for the laboratory wasn't much so we talked extensively about the things to do, the sequences of actions so as not to waste time, the opportunity to divide the roles in certain times (who was at the trolley and took the speed and who wrote the samples), in short, we tried to imagine everything that could happen during the laboratory test, finally I made sure I had material available more than enough for drafting the report (graph paper, pencils, erasers, rulers…everything you need).
The day of the lab everything went pretty much as we'd imagined, except that it took us longer than expected to get used to the cart and readings, as the assistant didn't seem very cooperative, resulting in us taking fewer samples of what we had anticipated. Everything went well until the day of drafting the report, where we had a nasty surprise…even if we redid the math over and over again the Gaussian refused to appear in our graphs, in its place there was something else…something that I will go into further after you. We immediately blamed ourselves and felt frustrated about what had happened, it was not possible, we must surely have done something wrong, and we began to examine the possible causes, perhaps we had taken too few samples, perhaps we had written them down hastily and had noted wrong values, who knows, we even came to suspect that we had moved too quickly around the measuring bench and/or had perhaps breathed too much, thus altering the speed measurements of the trolley. We resigned ourselves to asking (shamefully) for a new test to be carried out and we put up with the inevitable sarcastic comments of the assistants, we began to get worried, there simply WASN'T a way to have a third laboratory test (just mentioning the possibility produced a flow unstoppable head shaking and unpleasant comments like “cossa **** gave' fatto???” (quote from the original in Venetian dialect: “what the **** have you done?”), the second was the last test, the LAST, then a Gaussian had to come out of our relationship, or we could choose "a course of study more suited to our (giggle/grin) inclinations and abilities". It goes without saying that Pietro and I faced the second laboratory measurement as if it were an initiatory test, to be passed at any cost under penalty of death, that second day we moved (obviously ONLY when the trolley was NOT moving) like silent and merciless ninjas, even breathing was controlled, our movements synchronized, we must have looked creepy, because that's the only day I don't remember giggles or comments from technicians. The number of samples taken was TRIPLE of the previous time, and every single value noted had been checked separately by both. And the day of the report came, it should be remembered that we had been advised that each of the two tests had been carried out in a different measuring bench, therefore we could not use the data from the previous session in the new report, they were "completely irrelevant". To our surprise and terror the unfortunate event occurred ... in practice there was no way to see a correspondence between the sampled data and the shape of a Gaussian, what came out, in both cases, was something completely unexpected: it looked like to the hump, or rather to the humps of a camel: TWO beautiful humps, or more properly two "peaks of statistical probability" with a non-negligible interlude, a sort of "saddle" of data. No matter how many times we repeated the grueling entry of several hundred samples into the calculator, we always materialized in front of us a thick, “stubby” Gaussian (the pattern resulting from the data) in the center of the actual distribution and samples that had precisely two peaks. Exhausted and depressed we gave ourselves another day to think, it was Friday, in the late evening we simply gave in to tiredness and thought that after all... we had the whole weekend to "manage" the situation.
The other two days were lived amid growing frustration, guilt, fatigue, after another half day we finally resigned ourselves to falsifying the data, if they wanted a Gaussian we would somehow get it out, so we began a long work of pseudo-random tampering of the samples, in particular those concentrated in the two distinct "peaks" with a view to bringing out a single peak.
The problem was that, with the technological means of the time, it was a titanic job, in fact for each modification of the data we were forced to 1) calculate the parameters of the resulting Gaussian 2) draw both the Gaussian and the set of counterfeit data on graph paper and 3) compare the resulting opprobrium with some black and white copies of "exemplary" reports delivered in previous years that we had managed to find by roundabout routes, and about the authenticity of which I began to have growing doubts.
We spent the whole weekend in this continuous work of falsification-representation-comparison, looking for something plausible and presentable, we were exhausted, two men on the verge of a nervous breakdown.
Then came the day of presentation of the laboratory forgery work to the professor/technician, one had to show the prepared report, discuss it very briefly, and obtain the pass for the actual exam.
Pietro and I were tired, worried, and felt like abject scammers, but nevertheless we introduced ourselves, showed the work, and in no time at all we obtained the coveted pass, with some small protests for the "low quality" of the data reported (the drawing on graph paper hadn't seemed "neat" and "detailed" enough, but after a few comments a pass had followed, in short, the minimum wage of sarcasm.
As I was about to leave (with that feeling of having stolen something from someone and having gotten away by a hair's breadth), one detail caught my attention: one of the students had irritated the professor, and a loud lecture ensued. which was impossible not to hear.
I knew the student involved by sight and found him unpleasant, he was the classic nerd, always impeccable in language, form, manners, even the clothes, he seemed to come out of a department store catalogue, His comments on others' performances were always detailed, vaguely sarcastic, even biting … well … a real pain in the ass.
The nerd's work made me understand what the professor meant by "careful work", he had found a beautiful type of graph paper, a larger sheet format and with a higher resolution than normal, The reported dots looked machine made, not a point larger than the others, there was not a trace of erasing with the eraser, not a stain, not a wrinkle of the paper… nothing… an immaculate work…..but….HERETIC: the nerd had deliberately chosen to report first The experimental points (the threshold percentages of the samples taken), and then he had traced with an ability worthy of Giotto a curve that empirically interpolated these points, the resulting curve was completely empirical, and was not even remotely similar to a Gaussian, it was perfectly outlined what I had only glimpsed: a splendid and clear pair of camel-like humps, wonderful, true, unmistakable… even aesthetically pleasing.
The professor initially flew into a rage, then, considering the nerd's reputation (a whole semester in the front line of the lectures, with three hundred people fighting for that place, which he invariably managed to win) unexpectedly gave a second opportunity for the heretic: "go away, come back tomorrow with something that eliminates this opprobrium and brings out something plausible".
I don't know what happened the next day but I had the opportunity to later see the nerd go all the way to graduation, and before me, so I think he too did what Pietro and I had done... only better, and probably faster.
Time passed quickly and I forgot what happened until several years later, and precisely until the drafting of my degree thesis…
(To Be Continued: The Gaussian Episode 3)
