Lingyis on 15 August 2007
| marc said: This is a case of a statistic with very low power. Many engineers (especially chemical/bio-engineers) will tell you that no statistic is meaningful if it is not followed by its power calculation. Unfortunately... many statisticians don’t even know what "power calculations" are... and certainly none of those stock calculator/predictors will ever even mention them to you. |
a) stock prices (or social sciences in general) aren't repeatable experiments. it's a reasonable critique, but a non-constructive one. plus, there are plenty of things to do even with one set of data--situations do present them over and over again, from months (or years in the above example) down to on the order of seconds. that's how many quantitative-strategy based hedge funds make their living.
b) sure statisticians know that. it's just that it's usually under a different name than "power calculation".
hmm, i wonder if people still use stock calculators.
the Wii is an epidemic.
