datadataeverywhere comments on Humans are not automatically strategic - Less Wrong

153 Post author: AnnaSalamon 08 September 2010 07:02AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (266)

Sort By: Controversial

You are viewing a single comment's thread. Show more comments above.

Comment author: datadataeverywhere 09 September 2010 06:02:58PM *  10 points [-]
P(micronutrient deficiency) = 0.2

I would go as high as 0.3 if you extend to third world countries, but suspect it's lower among people like ChronoDAS who can afford a variety of food. Either way, it's good enough.

P(micronutrient deficiency|no multivitamin) = 0.8

The law of conditional probability indicates that you think that a minimum of 75% of the population takes a multivitamin. I think this is way too high, especially for a population that has a 20% micronutrient deficiency rate.

P(micronutrient deficiency|no multivitamin & depressed) = 0.95

So the rate of depression among those with micronutrient deficiencies (and who don't take their vitamins) is about 119% that of the general population? I can buy that, but if it's that low, then why are you so sure that a micronutrient deficiency is "greatly contributing" to his depression?

I agree that there's no harm in having CronoDAS gather data or experiment a little, since sulbutiamine seems to have very few negative side effects with recommended doses.

My main reason for brining it up is that I see some very high probabilities tossed about on Less Wrong, and it bothers me when I feel like they're assigning numbers that they can't justify. I'm still skeptical about your 95% confidence, but it's nice to see a break down.

Would you be willing to take a bet at a 2/3rds payoff that CronoDAS has no thiamine deficiency? How about a 1/19th payoff that taking a daily multivitamin wouldn't significantly alter how he feels?

[EDIT: Revised payoffs in bet to reflect professed certainty]