UPDATE: if you came here to learn more details about the rumored Higgs signal, which media around the world are discussing and which Fermilab Today just dismiss-tweeted, please visit this other more recent post for more details. Below is the original post which apparently originated a lot of buzz.

And for once, I feel totally free to speculate without the fear of being crucified. If you have followed my past blog adventures for long enough, you know that in at least a couple of occasions my posts have created some friction.

Blogging can mean walking on a rope for particle physicists involved in large collaborations - the ways of the internet are infinite, really: you never know where trouble may come from! The chance to piss someone off forces bloggers to avoid making names even when they discuss humorous incidents; the internal rules of the experiments they participate in make bloggers wary of even discussing stuff that is approved for public distribution. A daily application of self-censoring review procedures before hitting the "submit" button must be enforced.

But not this time. I am sure of one thing: I know nothing at all, so I can certainly talk about it without violating any rule! It so happens that I have heard voices about a possible new "three-sigma" Higgs effect, and I do not even know which experiment this comes from! Surely, no single experiment can get mad at me this time if I tell you what it is about, right ?

...Right. Well, I am not totally sure, but I am willing to declare that I have the right to express myself here, to some extent at least! So let me spill my guts. They are almost empty anyways...

The Rumor

It reached my ear, from two different, possibly independent sources, that an experiment at the Tevatron is about to release some evidence of a light Higgs boson signal. Some say a three-sigma effect, others do not make explicit claims but talk of a unexpected result. That the result comes from the Tevatron is for sure, since the LHC experiments do not have nearly enough data yet to search for that elusive particle, and other particle physics experiments in the world have not nearly enough energy to produce it. However, I am unable to understand whether the rumor comes from CDF or from D0.

Lest you jump at conclusions too early, I need to explain something more: despite being a CDF author, I unfortunately do not follow actively the works of the Higgs Discovery Working Group within CDF, so a Higgs excess in CDF data could well have escaped me. In principle, if I now took on digging hard enough in the internal pages of the CDF experiment I might be able to find out if this signal is coming from there, and maybe learn more about it. But there are at least a dozen analyses to dig into! Too much work - while wild speculation is more fun!

Reasoning On It

So let us take a look at the latest Higgs boson limits, released jointly by CDF and D0 last November. The dozens of analyses combined for the global limit were based on a dataset amounting to anything between 2 and 5.4 inverse femtobarns of proton-antiproton collisions, while right now the experiments have probably in their hands over 50% more processed and analyzable data.

The graph I choose to make a point is actually not one describing the limit on the Higgs boson cross section as a function of Higgs mass. Rather, let me pick the one showing, as a function of mass, a quantity that describes more clearly whether the data are background-like or signal-plus-background-like. The hatched black and red curves in the figure below show the value of the statistical estimator LLR (not going to explain you here what it is, but ask for it in the comments thread if you are interested) that the experiments would have globally observed, on average, if the higgs were there (red) or not (black). The farther the two curves are, the more sensitive the experiments are to a Higgs signal.

Also note the green and yellow bands, drawn around the expected background curve: they denote the typical extent of one- and two-sigma fluctuations expected in the data. In other words, if the Higgs is NOT at 130 GeV, say, then the LLR is expected to be on average equal to 1, but 68% of the time we may expect to find it anywhere between -1 (lower edge of the green band at 130 GeV) and +2.6 (upper edge). This is the so-called "one-sigma" band.

Now, look at the full black line. This shows the actual LLR value of the data, after the complicated analyses that sought the Higgs decay in dozens of different possible final states is processed. You notice several things.

The first thing to note is that the curve stands more in the "signal-plus-background" region for masses below 145 GeV, then going up and following the "background-only" curve for higher values.

The second thing to note is that while at 165 GeV the two LLR expectation curves are quite far apart (meaning that a Higgs boson might have produced a 3-sigma excess there, quite easily), at 120-140 GeV the curve of signal-plus-background stays on the border of the green band: the _expected sensitivity is there at most a one-sigma effect. In other words, a Higgs boson at 130 GeV would on average produce a 1-sigma deviation from the background-only curve, in the Tevatron data analyzed until November 2009. On average, though! The actual observed data, if it contained a Higgs boson, could produce larger signals, if the experiments got lucky.

The third thing to note is that the black curve in the low-mass region stands even lower than the red hatched curve! That means that the data there is definitely more signal-like than background-only-like. But is this a significant observation ? Well, no: the curve is well-contained within the yellow band. A less-than-two-sigma effect.

So, that was the situation last November. What should we expect now ? Could the black curve fall further down, hinting at a Higgs boson in the 115-140 GeV range ?

It could. In my opinion, a further fluctuation of the data, and the addition of 50% more of it, could bring the black curve out of the yellow band, toward a three-sigma signal-like effect. Is this what the rumors are about ? I do not know, but one thing is sure: we will know soon... If you are coming to Paris for ICHEP, you are among the lucky ones who will get the information first-hand from the analyzers.

An Appendix: Why Rumor Mongering ?

Why am I doing this ? I know several "serious" physicists and colleagues who have questioned this care-free attitude of mine in the past. What good does it do to shout "Higgs" every second week ?

It does a lot of good to particle physics, in my very humble, but not quite uninformed, opinion. I have made this point other times, and will not repeat it here. Suffices to say that, in a nutshell, keeping particle physics in the press with hints of possible discoveries that later die out is more important than speaking loud and clear once in ten years, when a groundbreaking discovery is actually really made, and keeping silent the rest of the time.

And there is another reason why I find this kind of rumor-mongering entertaining: maybe some informed soul out there might comment anonymously and share some more gossip about the matter with us... ;-)