Thanks to their absence of play, absence of contact friction and possible monolithic fabrication, flexure pivots offer advantages over traditional bearings in small-scale, high accuracy applications and environments where lubrication and wear debris are proscribed. However, they typically present a parasitic center shift that deteriorates their rotational guidance accuracy. Existing solutions addressing this issue have the drawbacks of reducing angular stroke, prohibiting planar design, or introducing overconstraints or underconstraints. This article presents a new triple crossed flexure pivot we have named TRIVOT whose kinematics theoretically nullify its parasitic center shift without overconstraints nor internal mobility. In the physical implementation, the center shift is non-zero but we show using the finite element method (FEM) that it is reduced by one order of magnitude in comparison to the widely used crossed flexure pivot (CFP). This allows to choose a crossing ratio of the flexures that either maximizes the angular stroke limit for given flexures or results in a compact planar design with the possibility of a remote center of compliance (RCC). Based on a pseudo-rigid-body model (PRBM), formulas for the rotational stiffness and angular stroke limit of the TRIVOT are derived, which are then validated by FEM. Finally, we show that a high support stiffness can be achieved based on a preliminary study for a mechanical watch time base application. We expect this new pivot to become a competitive alternative to the standard CFP for applications where high accuracy and compactness are required.