4 ms·
Nassim Taleb somehow likes to beat up on normals... We Bayesians have similar notions, but we usually try not to overly bully frequentist methods, the poor thi
by bayesianhorse 13y ago
Nassim Taleb somehow likes to beat up on normals...
We Bayesians have similar notions, but we usually try not to overly bully frequentist methods, the poor things. Also, being familiar with Bayesian methods, a lot of what Taleb is saying sounds vaguely familiar...
- kylebrown 13y agoWell, frequentist methods are taking the blame for the most recent financial crisis, such as assuming a normal distribution when the empirical one is fat-tailed. Perhaps the Bayesian methods will take the blame in the next financial crisis. Such as the error in estimating a non-stationary distribution and quantifying the uncertainty.
- bayesianhorse 13y agoStatistical economists or analysts always said that the normal distribution is a simplification and that this simplification has its own problems. It's just that traders and bankers would fire statisticians who were too vocal about it as wasting their time with unnecessary explanations... Bayesian methods in general can only take the blame if you can prove some other method being more reliable. Also they have another advantage: Bayesian methods are so mind blowing and beautiful, that it is hard to blame them for anything!