Calibrate your self-assessments

Extends calibration from probability estimates to self-assessment. Scott's anchoring story: in Ireland he walked out of every exam certain he'd failed, then passed with honors — so he learned to override the feeling (calibrating a consistently-too-low self-estimate). Other miscalibrated self-areas: embarrassment (he literally cannot recall a single classmate's wrong answer, so no one remembers his), interestingness, and flirting (his hunch that a woman likes him is 'almost always wrong'). Then the research: self-rated attractiveness correlates only ~r=.2 with others' ratings but ~.5-.6 with self-esteem; self-rated likeability has ~zero correlation with actual likeability; both over- (94% of professors above average) and under-estimation (BDD) are common. The practical/therapeutic upshot: self-esteem is a poorly-calibrated set point, not a measurement of reality, so replace inside-view self-judgments with outside-view/empirical ones (hotornot beats your gut), especially when your brain insists you're worthless.

Quality 75 · Excellent. Strong. A clear, well-evidenced, and genuinely useful piece — the calibration-of-self-worth framing plus the attractiveness/likeability self-correlation research gives it real practical and even therapeutic value. 75.

Claude’s paradigm shift 46 · Moderate. Moderate. Calibration is established LW canon; applying it to self-assessment/self-esteem with supporting research is a useful fresh synthesis. 46.

Real-world impact 2 · Minor. A clear, well-evidenced piece extending calibration from probabilities to self-assessment (override the consistently-too-low self-estimate), with real practical and even therapeutic value. Conceptual/practical influence within rationalist discourse, no material change — low RWI.