by Ian Pollock

*any*coherent safety decisions without using cost/benefit analysis.

The Calibration Game (great for queues)

PredictionBook (group calibration practice)
and a bonus, Scott Siskind on prediction markets and futarchy

Rationally Speaking is a blog maintained by Prof. Massimo Pigliucci, a philosopher at the City University of New York. The blog reflects the Enlightenment figure Marquis de Condorcet's idea of what a public intellectual (yes, we know, that's such a bad word) ought to be: someone who devotes himself to "the tracking down of prejudices in the hiding places where priests, the schools, the government, and all long-established institutions had gathered and protected them." You're welcome. Please notice that the contents of this blog can be reprinted under the standard Creative Commons license.

by Ian Pollock

The Calibration Game (great for queues)

PredictionBook (group calibration practice)
and a bonus, Scott Siskind on prediction markets and futarchy

Subscribe to:
Post Comments (Atom)

If you like Barbara Connors and “sunk cost”, check out Dan Ellsberg’s book, “Risk, Ambiguity, and Decision”, or read (for free) the article it was based upon: “Risk, Ambiguity, and the Savage Axioms”.

ReplyDeleteAvailable at: http://www.nssl.noaa.gov/users/brooks/public_html/feda/papers/ellsberg1961savage.pdf

Great, thanks!

DeleteI think I will read that book. The article is very interesting, although I confess my intuitions are totally indifferent between bets in these cases of ambiguity.

DeleteI think what's intuitively throwing us off is that an "ambiguous" bet carries with it a high value for further information. So if you offered me a bet at say 1.2:1 on

(a) a coin flip coming up heads; or

(b) the democrats winning in 2016 (to which I assign about 1:1 odds)

my response would not be a *preference* for the coin flip, just a request for an hour on google to investigate recent polling etc.

If you refuse to give me an hour on google, then I have no preference between the two bets. (Well, except that I can collect earlier on the coin flip. Nevermind that factor.)

Small problem with the circumcision article. Although it is well and good to question the methodology of this kind of study, I have an issue with the following statement by Earp:

ReplyDelete"That’s right: 60% is the relative reduction in infection rates, comparing two vanishingly small percentages: a clever bit of arithmetic that generates a big-seeming number, yet one which wildly misrepresents the results of the study. The absolute decrease in HIV infection between the treatment and control groups in these experiments was a mere 1.31%, which can hardly be considered clinically significant, especially given the numerous confounds that the studies failed to rule out."

Ahem! If I say the rate of something is decreased by 60% due to some measure, it means this: The rate after the measure is applied is 40% lower than the rate without the measure. That is all.

For example, suppose I read a study that shows that using seat belts decreases the rate of fatalities from traffic accidents by 60%. I am smart enough to know that the rate wasn't (X + 60)% before the measure, and is X% after the measure. If there are 1000 deaths for every million accidents without seat belts, then this means there would be 400 deaths for every million accidents with seat belts. This isn't some "clever bit of arithmetic", it's just common sense. I just wish Earp could see that. I'm sure he'd be happy to point out that wearing seat belts only decreases the risk of death from traffic accidents by .06% (assuming my hypothetical numbers).

Correcting myself: "The rate after the measure is applied is 40% lower than the rate without the measure" should read: "The rate after the measure is applied is 40% of the rate without the measure."

Delete