82,891
edits
Metric: Difference between revisions
no edit summary
Amwelladmin (talk | contribs) No edit summary Tags: Mobile edit Mobile web edit |
Amwelladmin (talk | contribs) No edit summary Tags: Mobile edit Mobile web edit |
||
| Line 24: | Line 24: | ||
*'''Regressive''': using a single metric as a proxy to measure “[[multivariate]]” phenomena that are driven by several factors. Here [[Simpson’s paradox]] is not your friend. Much “[[social justice]]” — which we define as the wishful, if not wilful, tendency to boil complex socioeconomic phenomena down to simplistic moral propositions that even a dull fifth-former could understand, and only a dull fifth-former would fall for — stumbles into this trap. | *'''Regressive''': using a single metric as a proxy to measure “[[multivariate]]” phenomena that are driven by several factors. Here [[Simpson’s paradox]] is not your friend. Much “[[social justice]]” — which we define as the wishful, if not wilful, tendency to boil complex socioeconomic phenomena down to simplistic moral propositions that even a dull fifth-former could understand, and only a dull fifth-former would fall for — stumbles into this trap. | ||
*'''Extremal''': Where a given metric is useful ''within a range'' — such a range generally corresponding to “normalcy”: “peacetime”, “normal operating conditions”, “[[business as usual]]” and similar platitudes — but which breaks down, fails, or even reverses itself in extremes or unusual cases beyond that range. Paging Messrs [[Black Scholes option pricing model|Black and Scholes]]. You know, using [[normal distribution]]s of independent events to model dependent events, like human behaviour of the market. These are especially fraught because the 80% of the time these metrics work, and they work fabulously, is exactly the range over which ''it doesn’t matter whether they work or not''. When things ''aren’t'' blowing up. The use case for the metric in the first place was | *'''Extremal''': Where a given metric is useful ''within a range'' — such a range generally corresponding to “normalcy”: “peacetime”, “normal operating conditions”, “[[business as usual]]” and similar platitudes — but which breaks down, fails, or even reverses itself in extremes or unusual cases beyond that range. Paging Messrs [[Black Scholes option pricing model|Black and Scholes]]. You know, using [[normal distribution]]s of independent events to model dependent events, like human behaviour of the market. These are especially fraught because the 80% of the time these metrics work, and they work fabulously, is exactly the range over which ''it doesn’t matter whether they work or not''. When things ''aren’t'' blowing up. The use case for the metric in the first place, remember, was to warn you about risk events. A “heightened risk” metric that you can only rely on when there isn’t a heightened risk is a ''waste of cheese''. | ||
*'''Causal''': the old [[correlation]] is not the same as [[causation]] chestnut. | *'''Causal''': the old [[correlation]] is not the same as [[causation]] chestnut. It may be true that people often buy ice-cream when they are wearing sunglasses, but handing out complimentary sunglasses will not improve ice-cream sales. | ||
*'''Adversarial''': Substitute targeting the desired outcome — a senior tranche in a portfolio of mortgages that will not default in any circumstances — with one that is rated AAA. Hello, [[global financial crisis]]. | *'''Adversarial''': Substitute targeting the desired outcome — a senior tranche in a portfolio of mortgages that will not default in any circumstances — with one that is rated AAA. Hello, [[global financial crisis]]. | ||