Comments on: I don’t even know how wrong I am!

By: Dan

Dan — Sat, 12 Jun 2010 10:37:26 +0000

Hey JD, It’s Dan from the “R Flashmob”. Funny to come across this, I work with Sharad and Dan at Yahoo, and we’re building a game based on these confidence intervals. And I know you for completely other reasons. Small world!

By: Michael P. Manti

Michael P. Manti — Tue, 13 Apr 2010 00:41:47 +0000

I scored only 5 out of 10. Interestingly, my near misses were on the answers I thought I had some knowledge, and so I narrowed my interval. If only I’d been less confident!

BTW, I always thought that Rummy took undeserved flack for that statement, which is perfectly sensible. His real problem was that he overestimated how much he knew he didn’t know and underestimated how much he didn’t know he didn’t know. Or something like that.

By: Daniel Reeves

Daniel Reeves — Mon, 12 Apr 2010 20:52:28 +0000

Hey JD, you were one of only 12 to overshoot on the weight of an empty 747, compared to 148 people with the opposite problem (too low). Only 58 people had good intervals. (This is as of when we posted the follow-up piece with the answers.) Thanks so much for the kind words about our blog!

By: Will Buthod

Will Buthod — Mon, 12 Apr 2010 16:30:39 +0000

OK JD, I’m going to be pondering this listlessly for days. thanks.

I’m feeling just snarky enough to wonder if, should this same test be administered to groups of people instead of individuals, the % of PhD’s has a similar effect. We’ll bill it as a death match between Surowiecki and Taleb, and let Gladwell referee.

By: Harlan

Harlan — Mon, 12 Apr 2010 15:48:52 +0000

Interesting. I also only got 4 right!

Did you see this? http://www.johndcook.com/blog/2010/01/31/parameters-from-percentiles/ A common task when building Bayesian and/or Expert systems is to pull statistical knowledge out of people’s heads. Cook is talking about some simple ways to compute distribution parameters given percentiles.