jet reconstruction and jet algorithms

When a single quark or gluon is knocked out of a proton in a collision, you never see it alone. The strong force forbids a lone quark from travelling freely, so within a fraction of a moment it sprays into a tightly focused shower of ordinary particles — pions, kaons, protons — all flying roughly in the same direction, like a burst from a shotgun. This narrow spray is called a jet, and it is the closest thing a detector ever gets to seeing the original quark or gluon. Jet reconstruction is the procedure for gathering all those scattered particles back into a single object that stands in for the parton that made them.

A jet algorithm is the recipe that decides which of the dozens of measured particles belong together in one jet. The most widely used today is the anti-kt algorithm: it works by repeatedly merging the particles that are closest together in angle, building up nice round jets around the hardest particles. Once a jet is reconstructed, you add up the energies and momenta of its members to estimate the energy and direction of the original quark or gluon. Getting the jet energy scale right — knowing how a measured jet energy maps back to the true parton energy — is one of the most painstaking calibration tasks in the whole experiment.

Jets are the everyday currency of a hadron collider: most high-energy collisions produce them, and reconstructing them well is essential for measuring almost anything, from the top quark to searches for new particles decaying to quarks. The honest subtleties are real. A jet is not the quark itself but a fuzzy, algorithm-dependent reconstruction of it, so two analyses using different algorithms or radius parameters can count jets differently. Pile-up dumps extra particles into jets, the jet energy scale carries a sizeable systematic uncertainty, and there is no perfectly unambiguous way to say where one jet ends and another begins.