Properties of MT2 in the massless limit

Abstract

Although numerical methods are required to evaluate the stransverse mass, MT2, for general input momenta, non-numerical methods have been proposed for some special clases of input momenta. One special case, considered in this note, is the so-called `massless limit' in which all four daughter objects (comprising one invisible particle and one visible system from each `side' of the event) have zero mass. This note establishes that it is possible to construct a stable and accurate implementation for evaluating MT2 based on an analytic expression valid in that massless limit. Although this implementation is found to have no significant speed improvements over existing evaluation strategies, it leads to an unexpected by-product: namely a secondary variable, that is found to be very similar to MT2 for much of its input-space and yet is much faster to calculate. This is potentially of interest for hardware applications that require very fast estimation of a mass scale (or QCD background discriminant) based on a hypothesis of pair production -- as might be required by a high luminosity trigger for a search for pair production of new massive states undergoing few subsequent decays (eg di-squark or di-slepton production). This is an application to which the contransverse mass MCT has previously been well suited due to its simplicity and ease of evaluation. Though the new variable requires a quadratic root to be found, it (like MCT) does not require iteration to compute, and is found to perform better then MCT in circumstances in which the information from the missing transverse momentum (which the former retains and the latter discards) is both reliable and useful.