TRANSCENDING THE LEAST SQUARES

However, physicists in-volved with the ever more delicate and expensive experi-ments — and with the computing power at their disposal exponentiating by Moore’s law — are increasingly impa-tient with a sub-optimality of the venerable method in the non-linear, non-gaussian situations, beyond the limits of the Gauss-Markov theorem (see e.g. sec. 7.4.4. in [6]), that they have to deal with. Two examples are: (1) the neutrino mass measurement experiment [10] observing Poisson-distributed events, that has been providing con-text and a test bed for the methods described in the present Letter, and that will be referred to for the purposes of il-lustration; (2) the very high precision muon anomalous magnetic moment measurements where a non-trivial spe-cial case of the method we advocate was discovered [12] and put to use [18]. The concept of counting