We develop techniques to determine the mass scale of invisible particles pair-produced at hadron colliders. We employ the constrained mass variable m2C, which provides an event-by-event lower-bound to the mass scale given a mass difference. We complement this variable with a new variable m2C,UB which provides an additional upper bound to the mass scale, and demonstrate its utility with a realistic case study of a supersymmetry model. These variables together effectively quantify the 'kink' in the function max mT 2 which has been proposed as a mass-determination technique for collider-produced dark matter. An important advantage of the m2C method is that it does not rely simply on the position at the endpoint, but it uses the additional information contained in events which lie far from the endpoint. We found the mass by comparing the HERWIG generated m2C distribution to ideal distributions for different masses. We find that for the case studied, with 100 fb −1 of integrated luminosity (about 400 signal events), the invisible particle's mass can be measured to a precision of 4.1 GeV. We conclude that this technique's precision and accuracy is as good as, if not better than, the best known techniques for invisible-particle mass-determination at hadron colliders. * a.barr@physics.ox.ac.uk