In this paper, we present a set of reasons for the incorrect results presented by algorithms for estimating truck (or bus) factors, which are usually based on the analysis of commits. We report that commits do not explain the TF of 10 systems in a dataset of 33 systems with well-known TF results.