The weaker students that marked higher (read inflated) ended up doing better in the Assessment Portion than the accurately grading students who ended up being penalized for being lower (but more accurate)
Yes, that may happen. As is described in the documentation, the Workshop module itself can't really decide which assessment is "accurate" so it generally relies on vox populi (with all the consequences).
As a teacher you have several ways to minimise the impact of these weak assessments - such as providing your own assessments, or increasing weight of assessments by accurately grading students.