Dev Tools · 1h ago
Why Scoring Systems Fail: Lessons from Wine and Credit Ratings
The article explores how scoring systems like Robert Parker's 100-point wine scale and credit ratings compress into narrow bands, losing their ability to differentiate. Parker's scale effectively became 85–98, while Moody's AAA rating masked risk in subprime CDOs. The pattern recurs in any domain where scores drive economic outcomes, leading to inflated trust and systemic failures.
Meridian48 take
A cautionary tale for tech metrics: any score that becomes a target ceases to be a good measure, whether for code quality, user engagement, or AI safety.
scoring-systemsquality-metrics