How internal reward gain (αE) and threat error gain (αN) warp perceived competence across 40 identical trials (28 Wins, 12 Losses).