The pattern is a mismatch between record and belief. Results are fine, the process is followed, and the trader still experiences their performance as luck about to be exposed.
In markets it is sustained by genuine randomness. You can never fully prove that a good year was skill, which gives the doubt permanent material. The behavioural cost is undersizing, skipped setups, and quitting a working method after a normal drawdown.
Calibration data is the useful counterweight: execution scores, adherence rates, and expectancy over a real sample turn a feeling into a measurable question. See confidence-calibration and undertrading.
Related: confidence-calibration, undertrading, scared-money, luck-vs-skill