Naturalness priors and fits to the constrained minimal supersymmetric Standard Model

We examine the effect of a prior that favours low values of fine-tuning on Bayesian multi-dimensional fits of the constrained minimal super-symmetric Standard Model (CMSSM or mSUGRA) to current data. The dark matter relic density, the anomalous magnetic moment of the muon and the branching ratio of b ->) s gamma are all used to constrain the model via a Markov chain Monte Carlo sampler. As a result of the naturalness prior, posterior probability distributions skew towards lighter Higgs and sparticle masses, the effect being most pronounced in the gaugino sector. Interestingly, slepton masses are an exception and skew towards heavier masses. The lightest CP-even Higgs h(0)-pole annihilation mechanism becomes allowed at the 2 sigma level for the latest combination of measurements of m(1) = 172.7 +/- 2.9 GeV, provided we allow for a theoretical error in the prediction of its mass m(h0). m(h0) is constrained to be less than 120 GeV at the 95% C.L. Probing the branching ratio of B -> mu(+)mu(-) to the level of 2 x 10(-8), as might be achieved by the Tevatron experiments, would cover 32% of the probability density, irrespective of which of the two priors is used. (c) 2006 Elsevier B.V. All rights reserved.