Look for repeated behavioral patterns rather than treating every comment as equally authoritative. Feedback becomes more actionable when several people observe the same effect in different cases.

Conflicting feedback can also reflect different contexts. Being concise and explaining reasoning are not opposites if you lead with the answer and then give only the necessary logic.

Track feedback by category across mocks: structure, math, communication, synthesis, and interaction. Note the exact moment that triggered the comment.

Change one behavior at a time and test it in the next few mocks. Keep changes that improve outcomes across partners.

I ran into this in a mock too. The hard part for me was noticing it early instead of trying to fix everything at once.