Goodhart’s law is an adage named after economist Charles Goodhart, which has been phrased by Marilyn Strathern as: “When a measure becomes a target, it ceases to be a good measure.”[1] One way in which this can occur is individuals trying to anticipate the effect of a policy and then taking actions which alter its outcome.
Goodhart’s Law reflects part of human nature, there is no perfect rules, so there is always people using the defect of rules,. I think there are two advantages to understand Goodhart’s law for a data scientist. First, Goodhart’s law tells that data is not everything, a data scientist may be able to find and analysis data, but the difficult is the people behind data, if the data we collect and analysis is to help making rules, we must be aware of this, and be cautious so as to try to avoid unnecessary deviation; Second, as a data scientist, it is his/her duty to try to find the pattern behind data, that means a data scientist may find the abnormal pattern which may caused by Goodhart’s law and avoid further influence.
简要翻译:
古德哈特定律(Goodhart’s law),是以 Charles Goodhart的名字命名的,这是一个非常有名的定理:当一个政策变成目标,它将不再是一个好的政策。作为前英格兰银行的建议者,提出:当政府试图管理这些金融财产的特别标识时,它们便不再是可信的经济风向标。
作为数据科学家,了解古德哈特定律有两方面的启迪,其一:如果数据用于制定政策,那么务必要非常谨慎,尽可能避免政策引导激发古德哈特定律,其二:通过数据挖掘与分析,可以发现不正常的数据模型,以及早发现古德哈特定律。
其他资料:
古德哈特定律的例子,某地老鼠成患,政府部门提出对抓到老鼠的人予以嘉奖,这导致了很多人私自养殖老鼠,或者放生老鼠以谋取利益,反而导致鼠患更加严重。