Tommaso Dorigo

Happiness from a macro

Today I drove to Padova at lunch time, after visiting for the first time my new house. During these last few days I was unable to do enough analysis work, but a few ideas built up in my mind and I needed to flush them out -either by proving them useless or by exploiting them to the fullest.

One thing I needed badly to try was a pretty trivial check of the Z signal I've recently extracted. As my twentyfive readers know (I do not mean to be the equal of A.Manzoni with that citation, only to pay him homage), for having learnt it in a recent post, I have managed to obtain, after almost three years of fight, a signal of Z decays to pairs of b-quark jets in data collected by a dedicated trigger, which selects events with tracks likely to have originated from b-quark decay. But in Run 1 (1992-1996) at CDF I had done the same using data collected by a muon trigger, a less efficient way but the only one available then.

Today I wanted to see whether the signal was there in a trigger similar to the one used in Run 1: this one is called B_MUON, and collects events which have one muon candidate and one track likely to come from b-quark decay.

And I did it! I used the exactly same recipe, to verify it and double-check any systematics involved in the procedure. I had a dataset of 6.2 million events which had been sitting there in one of my disks for almost 6 months, and quickly processed it, then applied my macro...Which crashed mercilessly! I was only able to debug it and launch it before leaving my office at 7PM (no, I am not a crazy scientist as Gordon Watts, who can forget to go to sleep and keep working all night - I have a go***mn life out there, and not even a lusty Z signal is going to take it away from me!).

Of course, at home I asked Mariarosa for 20 minutes of mercy, and logged on the University server to check the results of the macro... It worked! And sure enough, I saw a signal completely compatible to what I had found in the dedicated sample, and with the exact characteristics I had guessed beforehand - a slightly offset mass, due to the energy lost to the non-interacting muon, and a signal to noise ratio equal to the other one, but a fourth of it in absolute size due to the less efficient trigger.

Tonight I am sooo happy! I now know that not only the signal is there to stay: I also showed myself that my background parametrization is likely affected by very small systematics. That means a very small error on the jet energy scale -I'd guess around 2.5% now (need to do more tests for that) and most probably less than 1% in two more years- and a hefty contribution to all Run II precision measurements of the top quark mass at CDF!

I also am quite confident that CDF will bless my preliminary plots (we call them PR plots if we want to just show them around with no permission to disclose the full analysis) in time for Moriond QCD, where I will have the honor of including them in my talk on Tevatron searches for the Higgs boson.

Gordon, you can tell your colleague to start working double shifts if he wants D0 to beat my measurement! :-)

New house!!!

Today is the day... And I think back at last year...

One year ago, I spent three months in search for an apartment to buy in Venice city center. My wife's family house -a big, old, beautiful historical house in campo S.Margherita, with a portal of the 14th century (see picture on the right)- had just been sold for a hefty sum, and we finally had the money to buy a house of our own. We needed at least 1300 sq. feet, three bedrooms, two baths, and since in Venice you walk, it had to be not far from Piazzale Roma (where our car is parked) and the train Station, to allow for quick escapes on weekends and faster daily commutes to Padova. On the map below, the green dot is where we live now, the red dot is the location of the new house.

Prices in Venice are obnoxious. Hard to find anything for less than 650$ per square foot. So the search had to be accurate. Also, there is not much in the market with the characteristics listed above... But I did my job thoroughly, and after three months I had seen 21 apartments, and narrowed it down to four possible purchases. In the end, we got lucky, since the one we really wanted was a bit too pricey for us but we were given help by my mother-in-law. We ended up buying the most expensive one, which fulfilled all of our requests: 1500 sq.ft., four bedrooms, a nice setting on third and fourth (and last) floor, and a 300 ft. terrace thrown in to boot. The net price: 900,000 US$, to which one must add 3% taxes and 2% fees.

That figure gives me goose bumps and three main feelings. Thinking that to earn that sum with my current salary I should keep working for fifty years is just one of the sensations I get. Another is joy for the luck we had to actually have that money to throw at one of the most important purchases of one's life. The last one is rage for the incredible price housing in Venice has achieved, due to "selling out" the town to wealthy foreigners, kicking out most originary citizens to the dull periphery inland.

So, today I am getting the keys of the house. This is unfortunately not the end of the story: although it is in almost perfect conditions, we decided we will remodel it. It came in steps: at first there was this corridor Mariarosa did not like. Then we realized it would be nice to have an open-space livingroom, which could be achieved by taking down a part of a wall. Then the main bathroom seemed a bit smallish... In the end, little will be left standing as it was! A friend of mine made blueprints for the new main floor (the upper floor will luckily need no remodeling!). So we are looking at 6 more months of pain and some 130,000 US$ more!

But today is the day no less. After ten months spent dealing with the seller, who gave us quite some trouble and procrastinated his leave, we finally can set foot on our new property!   

The hot chihuahua

In a recent post I briefly discussed about the hot chihuahua of Meet the Fockers, that humped missionary-style a Einstein puppet in a scene which -although pretty hilarious- could no doubt be considered the lowest point of this years' celebrations for The World Year of Physics - yet.

Yesterday evening I was browsing TV channels when I saw the trailer of the Italian version of the movie. I was not surprised to see poor Albert humped again in the trailer, given the relevance of that scene with the movie plot (I'm being sarcastic here).

But I was indeed surprised when, as the last scene of the trailer, just before fading, we were treated by another sex act of the relentless dog with poor Albert's puppet, this time -I have to get graphic now- sitting on his face, while unmistakably paying the same homage to him.

It appears that the latter scene was cut from the American version (but I would appreciate feedback on this if I missed it). If that is the case, it makes me wonder even more than I did in the other post, when I suggested that the more "natural" (for a dog!) act of humping Einstein doggy-style would have been considered too outrageous by the public: I of course very well know that there is a difference between what is considered obscene in different countries.... but in a country where none other than the President states under oath that a fellatio, unlike intercourse, is not sex, is it not funny that censors cut the former and keep the latter ?

The say of the week

I used to keep a web page with my stuff, on home.fnal.gov, a site owned by the Fermilab laboratory. I had a cool little box in the page, titled "The say of the week". I collected there sentences which I found amusing, and a new one made it there each week.

Then, one day, I put there a link: "this week's say is brought to you by G.B*** senior! Click to get it!". If you clicked on it, you saw the pic, where he was showing an eloquent finger pointing up. The picture had a sentence below it: "A bird in the hand is worth two in the b***" - asterisks were not there then.

A few days later, my site was obscured, and I was told some taxpayer had complained about improper use of DoE funds. I had to show I felt sorry for the mishap, and the say of the week was gone for good.

A couple of years later, I reinstalled the whole site in a Padova University site (Italians are much looser on these sorts of things), and I happily reinstated everything. But by then I had too many things to take care of, and there have since been very few additions to my little list.

Starting this week, I've decided I am going to recycle the sentences I had there. I will be more cautious, since the site before you, dear reader, is no less public-owned than the one which gave me trouble. But now, without further ado, here is this week's say:

"I believe sex is the most beautiful, natural and pure thing money can buy" (Steve Martin).

The buried lottery

Particle physics is cool, but being a particle physicist these days is tough. First of all, we live in our own little world, disconnected from the outside, incapable of making headlines, and aching for it: the general public is not interested in science in general, and subatomic interactions do not even score high among scientific attractions. Moreover, we suffer from a chronically tight budget, much smaller than the vast amounts of money thrown into space with Mars probes and other cosmic gizmos (hey don't get me wrong - I do love the Hubble!). Congressmen just don't get it why they should give away money for just another particle whatchamacallit, only a bigger whatchamacallit. (That's how the SSC died, sadly). It just does not excite them enough.

But I have a solution to both these problems at once! The idea hit me one day, when I was considering the smallness of the probability of a proton-antiproton collision producing a rare phenomenon such as the creation of a top quark pair inside our detector. What is our detector, buried 50 feet underground at Fermilab, if not a giant lottery ? The unpredictability of the quantum process of two particles colliding makes it a perfect randomizer. A buried 5000-ton lottery! Here are some dividends one might assign to detected signatures of processes:

• production of a jet with energy above 20 GeV: 1/100
• production of a B meson: 1/1000
• production of a jet with energy above 50 GeV: 1/10,000
• production of a W boson with decay to leptons:1/1,000,000
• production of a Z boson with decay to leptons: 1/10,000,000
• production of a top quark pair: 1/10,000,000,000
• production of a Higgs boson associated with a W, with same-sign trilepton decay (say e+e+mu-): JACKPOT!

Of course, the numbers are actually different (larger, for rare processes) if one considers the subset of collisions that actually get recorded (ones which get an identification number), since the online trigger system makes some selection on the fly exactly with the purpose of enhancing these tiny figures. But one would still be looking at numbers as small as one in a million for tight top pair candidates, and one in a billion for golden Higgs production.

Hence my idea. How to

1. get funds for running and upgrading our experiment ?
2. attract the interest of people, media coverage ?
3. show them how cool particle physics is ?

Simple! People love gambling, especially in the US. And in order to understand how big their chances are of winning a few bucks they will go a long way into understanding the details of the natural system at Blackjack or probabilities of slot machine dividends. So I propose that we sell tickets with event numbers printed on them before we start data taking (events recorded by the detector are numbered by run and event number), say at one dollar per event. We then assign dividends for particular events, based on the selection cuts they would pass at analysis level. These dividends should be honest enough to make the game enticing, but still allow for some margin of gain for the experiment. Let's live on our own funds, and screw the congress!

As soon as events are collected, people can verify whether the one corresponding to their ticket is a J/psi candidate or what, and get paid in case they won something. (Sure, we'd need a stable reconstruction program... Hehm!)

Imagine what a boost this would give to the knowledge of the general public! You would start hearing the guy next to you at the bar boasting about the event he got, maybe showing a colorful event display to everybody. "See this purple bar here ? This is a 60 GeV electron! A golden W candidate, I'm telling ya! And the darn track here, go**amn failed by an inch the requirement for a second electron! Would've won a million...". It would really be fantastic. The first Higgs candidate event would win several million dollars, and would get media coverage in no time. But people would learn physics, that's the best thing.

Of course, people who bought the events would be their sole owners. If you were to publish an analysis based on a restricted set of events, and wished to show an event display, you'd need permission from the owner, and maybe cite his name in the Acknowledgements section. "The authors wish to thank John Doe for kindly providing Fig.3"...

And you'd have to invite lottery participants to the trigger meetings as if they were shareholders... You could not change prescales (a prescale is a device that limits the rate of collection of a particular set of events, to keep the total output rate within the writing capabilities of the system) on your triggers without changing dividends, and things could get ugly. For once, we'd have trigger meetings with people screaming at each other for a good reason, MONEY! (People do scream at trigger meetings sometimes... And it just happens in return of minor reductions of their dataset size!).

This idea is too cool to let it go... I need to talk to Ray Orbach at the Office of Science of the DoE, pronto!

Sick

Well, today I resisted the temptation of staying in bed and went to work, but produced very little... I had a high temperature all day, coughed a lot, and my jobs kept crashing rather annoyingly.

I doubt I'm going to make it tomorrow... It's bad, because tomorrow is meeting day, and I should be following the Jet Energy and Resolution meeting in videoconference from FNAL. Moreover, we are giving the final touches to the evaluation of systematic uncertainties to the jet corrections, to be blessed next week: most high Pt physics results CDF will show at winter conferences in February and March depend critically on these - no blessed corrections, no results!

However, I don't want to give the impression that I am absolutely necessary for this process to complete: Florencia and Anwar will be at the helm. That's the main reason for not having a single convener in a working group.

Maybe I'll make an attempt at going to work in the afternoon...The meeting is 10AM at FNAL, which is 5PM here. We'll see how strong I am tomorrow.

Americans against war

Just read it on a web newsreader: 55% of polled american citizens now think it was not a good idea to wage war to Iraq (Washington Post-Abc), and 52% believes the war was a mistake (Usa Today-Cnn).

Better late than never, fellows!

I do hope that some of the change in public opinion was triggered by disgust towards the pictures showing horrific abuse and torture of Iraqi prisoners, rather than by cold accounting for pros and cons of the military action per se. The picture on the right, taken by US troops in the Abu Ghraib prison, is so disturbing I made the thumbnail very small. Use at your own risk.

Disclaimer: since my country happily flanked the invasion of Iraq, I do not claim to be any better as an Italian... I did, however, manifest all my outrage against the threatened war well before its start, with two PEACE rainbow flags on my front windows. But I am quite unhappy of seeing how right I was then.

Gloomy 1/24th

Today I got in the 8.02 train just in time, running with my PC bag in one hand and La Repubblica in the other. Running for the train is really something I am starting to avoid, given my mature age and the high frequency of trains to Padova (a 35' ride: see map on the right), but this time I was feeling energic. Anyway, in the train I started browsing the news, and my eye was caught by a 4-column piece about some researchers in Britain claiming that January 24th is the gloomiest day this year.

Their argument runs as follows: Mondays are the worst days of the week; January is (for northern countries) the month with the lousiest weather and the shortest daylight; around the 24th credit card companies send in their bills - and this month, hefty ones containing Christmas vacation shopping!

I found that pretty amusing. I was not feeling depressed in any way, since the sun was shining and my bills were average. Then, at work I managed to have a very unproductive day, trying unsuccessfully to patch a root macro (ended up having to re-process the 21M events of my dataset for a smallish adjustment). At three in the afternoon I started to feel very sick, tired of the few hours of sleep I had (thanks Ilaria), with a sense of oppression to my lungs and a general pain to the bones, and decided I'm incubating a flu. At 6PM I plodded homeward wearily, and now I'm going to get two tylenol cold and run to bed...

The darn British researchers and their 1/24 curse!

Moriond

Last Friday the CDF Speakers Committee selected me to represent the CDF and D0 collaborations at Moriond QCD, a winter conference on particle physics focusing on the strong interactions. I will be giving a presentation there on the new results on "Searches for the SM Higgs at the Tevatron" (SM= Standard Model, the current theoretical framework of particle physics), during the week from March 12th to March 19th. I sort of expected that my request for a talk at a winter conference was going to be accepted (I went as far as to include it in my schedule in a recent post here, and to reserve plane tickets for my travel accordingly), since I've not given talks for a while now - if one excludes a annual Italian symposium called IFAE (Incontri sulla Fisica delle Alte Energie == Meetings on High Energy Physics) where contributions are not distributed by the CDF Speakers Committee. However, I am quite pleased, since the conference will be held in a very nice ski resort in La Thuile (see photo above), a small mountain village close to the border between Italy and Switzerland.

I am not too worried by having to prepare my slides, since the subject of my talk is quite a familiar one. I will read the few most recent papers on the subject, then collect the most recent results from the two Collaborations I am representing, and put the latter together in a logical sequence. However, this talk puts quite some pressure on my schedule for the next few weeks. In fact, it would be quite nice if I could include in the talk a slide or two with preliminary results from my own search of Z boson decays to b quark pairs: a topic which is very closely connected to the search for the Higgs boson, which may show up in our data in the very same signature (two hadronic jets with b-quark content). I do have results, but they are not approved by CDF yet, and the process of approving even one or two "Public Relations" plots will be quite complex.

(I should mention that producing an approved result from the Z search is one of the milestones of the working group I co-lead for March 15th, so the winter conference is nothing but additional motivation for doing what I should be doing anyway!)

The process of producing a result in the analysis of experimental data typically starts when one obtains a "raw" result, which has then to be polished to be presentable. Polishing it does not mean manipulating it, but rather to start attacking it from all possible fronts, imagining all sorts of possible effects that might invalidate it, bias it, or fake it. By doing that, one learns all weaknesses of the result, and manages to estimate the size of all systematic uncertainties affecting the measurement.

So, starting today, most of my time will go into that business for the next few weeks!

An experiment in democracy

CDF just elected its new co-Spokesperson, which will stay in charge for the next two years (for 2005 together with Young-Kee Kim, which was elected last year). The winner is Rob Roser (left, with an appropriate hat ;-), a longtime fellow collaborator, who served CDF in several important roles in the past. I will not say more about him, other than he was not my first choice in the list of candidates.

The election of a new co-Spokesperson in CDF happens every year. From a list of candidates, nominated by their fellow collaborators, a first informal poll among the participating institutions allows a more restricted list to be formed. It is this latter list which is submitted to the vote of all members of the collaboration.

The final candidates - this year there were four - present themselves in a "town meeting", when they state what their priorities are, what they want to do as a Spokesperson, why they accepted the candidacy, those sorts of things. After that, the poll opens.

Every collaborator (be it a grad student, a researcher, a Nobel prize winner if there was one) can submit to the Election Committee a list of the candidates in order of preference. After the poll is closed, a suitably complicated algorithm selects the winner based on the number of times he was first, second, third, or n-th in the list. Don't ask me the details, I refused to learn them - it should not matter much.

The whole thing seems pretty much a rather democratic process. But there is a catch: the results are not divulgated! The election committee, which is formed by three members nominated by the executive board, just announces the winner, that is all.

I fought several battles in CDF. Some I won, and some I lost. They all took a toll on me, both in terms of time they took to be fought, and in terms of friends and enemies I made in the process. I was tempted to fight this one too, but have just decided that my time is not worth the effort this once: convincing my fellow collaborators that the flaw mentioned above is indeed enormous. What appears as a democratic process cannot in fact be called as such.

Leaving aside any suspicion for the election committee itself, or whomever vigilates over it, since the suspect is beyond the reach of my mean thinking, there however remain quite a few objections to not making public the outcome.

1. I voted. I want to know how many votes my favourite candidate got. He convinced me to vote for him, but I need to know how strong he really is. Next time he can run again, and I need that input to form a new decision, or to help him in his campaign.
2. My favorite candidate was not elected. Was the winner elected by a narrow margin, by a large margin ? I need to know how my colleagues voted. I might find out that my ideas on the candidates is not shared by my colleagues: it is something important to know. Hey, they might be right!
3. How many votes were submitted ? This crucially important datum is not broadcast any more than the others. We are a collaboration of about 700 scientists. Did 600 vote ? 200 ? I do not even get to know that! Not even if I did not vote myself.
4. Did a candidate get one vote ? That would be pretty mean to divulgate, but hey, wasn't it mean to elect G.W.Bush in 2000 by such a ridiculously narrow margin that the real outcome will never be known in earnest ? Democracy is not for the faint of heart. We need to know if a candidate got only the vote of his parents. He would spare us his boring town meeting speech next time.

I really believe that the choice of not making the detailed results public was made in good faith, albeit unwittingly. But I am a little ashamed of the failed democracy in my experiment. As I said, I pretty much would like to fight for it, but, as a colleague (one who ran for co-Spokesperson this year, and was not elected) taught me once, "you have to pick your battles". That's right. I cannot fight them all... Not if I am a researcher rather than a politician.

CDF: an experiment in democracy. Put together ~1000 intelligent human beings, let them elect their leaders and create their own bylaws. Will they evolve toward a happy island of people with flowers in their head, or a jungle of backstabbing lobbyists ? I am a democrat, but sometimes I envy collaborations where there is a undisputable leader, who will take all decisions, for good or for bad. In a few cases, they end up with a great person to lead them!

I am proud to be in CDF, a beautiful, very successful experiment, full of amazing people, endowed with great knowledge and intelligence - hehm, we do have some average Joes ;-). That is why I ache for these small flaws. But I am no idealist, and I know other experiments have theirs....