The global fraction on ground, that you find in Bifacial system > Unlimited Trackers 2D model, is calculated for clear sky conditions. As such, it is not applicable to your simulation results (unless you have a clear sky meteo file). Your intuition about how the global incident on ground is calculated is correct. However, PVsyst does the calculation hour by hour and differentiates the beam and diffuse irradiance components. The beam and diffuse components have different "fractions on ground", that will generally vary every hour with the sun / tracker positions. Therefore it is difficult (probably even impossible) to calculate it from the current outputs of PVsyst, in particular from the monthly results.
Indeed this fix has been (long) forgotten, thank you for the reminder. It should be implemented in version 7.2.10.