Lattice Monte Carlo Data versus Perturbation Theory

Abstract

The differences between lattice Monte Carlo data and perturbation theory are usually associated with the ``bad'' behaviour of the bare lattice coupling g_0 due to the effects of large (and unknown) higher order coefficients in the g_0 perturbative series. In this philosophy a new, renormalised coupling g' is defined with the aim of reducing the higher order coefficients of the perturbative series in g'. An improvement in the agreement between lattice data and this new perturbation series is generally observed. In this paper an alternative scenario is discussed where lattice artifacts are proposed as the cause of the disagreement between lattice data and the g_0-perturbative series. We find that this interpretation provides excellent agreement between lattice data and perturbation theory in g_0 corrected for lattice artifacts. We show that this viewpoint leads typically to an order of magnitude improvement in the agreement between lattice data and perturbation theory, compared to typical g' perturbation expansions. The success of this procedure leads to a determination of Lambda_MSbar^{N_f=0} of 220 +- 20 MeV. Lattice data studied includes quenched values of the string tension, the hadronic scale r_0, the discrete beta function Delta\beta, M_rho, f_pi and the 1P-1S splitting in charmonium. The new 3-loop term of the lattice beta- function has been incorporated in this study. A discussion of the implication of this result for lattice calculations is presented.