We investigate proton mass effects in the handbag approach to wide-angle Compton scattering. We find that theoretical uncertainties due to the proton mass are significant for photon energies presently studied at Jefferson Lab. With the proposed energy upgrade such uncertainties will be clearly reduced.PACS numbers: 13.60.FzIn Refs. [1][2][3] we have investigated the handbag approach to wide-angle Compton scattering off protons, γp → γp. Analogous results have been obtained in Ref. [4]. In the handbag approach the Compton amplitude is given by a hard scattering γq → γq at parton level multiplied by soft Compton form factors describing the emission and reabsorption of the quark by the proton. The kinematical requirement for the applicability of this approach is that the Mandelstam variables s, −t and −u are large compared to a typical hadronic scale of order Λ 2 = 1 GeV 2 . This implies s, −t, −u ≫ m 2 , where m is the proton mass. At Jefferson Lab (JLAB) there are ongoing experiments to measure the Compton cross section and certain spin transfer parameters [5]. Presently available beam energies are however not very high. In this note we investigate, as an example, the role of a non-negligible target mass in the handbag approach. An important issue in this context is the way to relate the dynamical variables of the approach to the external kinematics, determined by the experimental conditions. This relation is not unambiguous, which is one of the sources of theoretical uncertainties in the handbag approach. We study three different approximations and take the differences in their predictions as a measure of the theoretical uncertainty, which should be taken into account in attempts to extract the Compton form factors from experimental data.The external kinematics is determined by the beam energy E γ L in the laboratory and by the scattering angle θ in the center of mass frame. These quantities fix the external Mandelstam variables byThese variables should not be changed or approximated in a theoretical calculation. Keeping this in mind we suggest a separate treatment of the kinematical factors from phase space and flux and of the scattering amplitude, which contains the dynamics. In the handbag approach the Compton cross section then reads [1, 3]to leading O(α s ), where R V , R A and R T are the Compton form factors. The one-loop corrections to the hard scattering have been evaluated in [3]. They were found to be small in the backward hemisphere and increased up to about 30% for cos θ = 0.6. The Mandelstam variableŝ s,t,û refer to the partonic subprocess γq → γq. They coincide with the external variables s, t, u up to corrections of order Λ 2 /s. To calculate these consistently is beyond the accuracy of the approach in its present form. In particular, different choices forŝ,t,û lead to different results for the cross section at finite s.We investigate the numerical effects of this ambiguity in three different scenarios. For the beam energy we take E γ L = 4.3 GeV, corresponding to s = 8.97 GeV 2 , where there wil...