Tuesday, August 18, 2015
Time Series Homeopathy
Recently Doug Keenan has again taken up the cudgels against the Met Office with backing from the usual suspects and indeed, the Met Office has taken a decision not to engage further. While the Met Office thinks that some things Keenan says are interesting and very brave, with the greatest respect and they will hear what he says and bear it in mind, well, they choose not to devote further time to dealing with him. A feature of these attacks are Keenan's claims that the best time series analysts tend to be in finance (see comment). This, of course, is something that Richard Tol also firmly believes, giving rise to an hilarious disputation over at ATTP.
Eli has thought about this for a while and concluded that Keenan's claims about time series analysts in the financial industry are a clear reference to chartists, which happen to be the homeopaths of stock pricing. There is no there there, only a few lucky winners when they roll up the sidewalks
7 comments:
Dear Anonymous,
UPDATE: The spambots got clever so the verification is back. Apologies
Some of the regulars here are having trouble telling the anonymice apart. Please add some distinguishing name to your comment such as Mickey, Minnie, Mighty, or Fred.
You can stretch the comment box for more space
The management.
I would suggest that Chartism is actually more like Tarot or palmistry than homeopathy. To be fair (just for the sake of argument) to the finance people, time series analysis is used as part of Black-Scholes options pricing. But, among other things, misuse of that technique produced some rather ugly real world outcomes in 2008.
ReplyDeletehttp://www.theguardian.com/science/2012/feb/12/black-scholes-equation-credit-crunch
I guess Tol and Keenan would argue that the 2008 financial crisis might have been bad, but it would have been much worse if the time series analysis had been done by geoscientists.
ReplyDeleteChartists !
ReplyDeleteBreak out the ouija board and order Lord Melbourne to send in the Guards
I'd argue the best time series analysts are those that know the most about the subject being analysed, and not those with the cleverest statistical techniques.
ReplyDelete+1
ReplyDeleteDid you hear about the suicidal homeopath? He overdosed by taking 1/60th of the recommended dose.
ReplyDeletesnark--
ReplyDeleteROFL!