Har qanday kuzatilgan statistik muntazamlik nazorat qilish maqsadida unga bosim o'tkazilgach, yiqilib ketadi.
Charles Goodhart
Shuningdek, odatda quyidagilarga murojaat qilinadi:
O'lchov nishonga aylanganda, u yaxshi o'lchov bo'lishni to'xtatadi.
Marilyn Strathern
Qonunda aytilishicha, o'lchovga asoslangan optimallashtirish o'lchov natijasining o'zini devalvatsiyaga olib kelishi mumkin. Jarayonga ko'r-ko'rona qo'llaniladigan haddan tashqari tanlangan chora-tadbirlar to'plami (KPI) buzilgan ta'sirga olib keladi. Odamlar o'z harakatlarining yaxlit natijalariga e'tibor berish o'rniga, muayyan ko'rsatkichlarni qondirish uchun tizimni "o'yin" qilish orqali mahalliy optimallashtirishga moyildirlar.
Haqiqiy dunyo misollari:
- Tasdiqlashsiz testlar, metrik maqsad yaxshi sinovdan o'tgan dasturiy ta'minotni yaratish bo'lishiga qaramay, kodni qoplash kutilmasini qondiradi.
- Developer performance score indicated by the number of lines committed leads to unjustifiably bloated codebase.
Shuningdek qarang:
The Goodhart's Law on Wikipedia
Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes.
Charles Goodhart
Also commonly referenced as:
When a measure becomes a target, it ceases to be a good measure.
Marilyn Strathern
The law states that the measure-driven optimizations could lead to devaluation of the measurement outcome itself. Overly selective set of measures (KPIs) blindly applied to a process results in distorted effect. People tend to optimize locally by "gaming" the system in order to satisfy particular metrics instead of paying attention to holistic outcome of their actions.
Real-world examples:
- Assert-free tests satisfy the code coverage expectation, despite the fact that the metric intent was to create well-tested software.
- Developer performance score indicated by the number of lines committed leads to unjustifiably bloated codebase.
See also:
O'zbekcha
English