Scientists compare PVsyst, Comsol software in assessments of bifacial PV

The blurb under the title of this PV Magazine article is misleading:

"They found that PVsyst outperformed Comsol in calculating solar production in nearly all cases."

What does "outperformed" mean? PVsyst is more accurate? No, it seems to be a misinterpretation of this line:

"With the same simulation method between softwares, PVSyst gives us a higher value for solar production in almost all cases."

In fact, the SEJ article (fortunately open access) does not compare to any measurements and therefore can only report on differences, but not on which simulation is more accurate. Anyway, simulation accuracy is not just a question of the software itself but also of the inputs to the software.