Join 3,553 readers in helping fund MetaFilter (Hide)


Looking for a type of self-test on ballpark estimates
June 6, 2014 5:16 AM   Subscribe

A number of years ago (I'm going to say with 90% confidence it was between one and ten years ago) I came across a test that asked you to provide a range for a number of values that you could make a somewhat educated guess at but probably wouldn't know the answer to off the top of your head (so, maybe, the land area of Maui, or the number of words in A Tale of Two Cities). The goal was not to pinpoint the values, but to provide a range that contained the correct answer with 90% certainty, so, the best score was 9 out of 10 questions correct. A score of 10/10 was as bad as 8/10 -- the aim was to get 9, so that your ranges weren't ridiculously broad.

The only specific question I remember from the test was the weight of an adult (African I think) elephant. I remember this one because I guessed a range that I thought was reasonable and maybe a bit broad (like say 2000-8000 pounds) and the correct answer was way more than that (I'm apparently not alone here).

Any leads? It doesn't have to have the same questions if there are variations of this test around.
posted by payoto to Science & Nature (7 answers total) 12 users marked this as a favorite
 
A friend of mine sent me a version of this, but it was a plaintext email, not any kind of web quiz. I suspect it came from Less Wrong, or a Less Wrong meetup group.
posted by grobstein at 5:51 AM on June 6


Here are the questions my friend sent me. They include the African elephant weight question, so I'm pretty sure this is what you are looking for. The questions (and the numerical answers, not included) come from a boxed game called Wits & Wagers. But this use of them comes from a calibration exercise at a Less Wrong meetup.

1. Weight of liberty bell in pounds

2. Year when 8000+ people moved to California looking for gold

3.Year Orwell first publishes 1984

4. Year Sally Ride goes into space as first US woman

5. Barry Bonds 2007 Salary (benchmark was 22,000 in 1988)

6. In this year a woman convinced leading doctors she had given birth to 18 rabbits

7. Number of accredited med schools in the US

8. Fastest wind speed in mph measured inside a tornado

9. What % of animal species have backbones?

10. How many monologues did Johnnie Carson give on the tonight show?

11. How old was Kurt Cobain when Nevermid was released?

12. How many short stories did Edgar Alan Poe publish?

13. How old was Anna Nicole Smith at her time of death?

14. What is the average weight in lbs of an adult African elephant

15. What % of the world population lives in North America (again as of 2007)

16. Number of unprovoked shark attacks in 2006?

17. How many inched wide was the largest hail stone recorded in the US?

18. How many Rhode islands would fit in the area of Alaska?

19. How many theatrically released James Bond movies starred Sean Connery

20. What year was the first Female prime minister elected?

21. What year were post its invented

22. How many Grammys did Barbra Streisand win?

23. What % of apples grown in the US come from Washington?
posted by grobstein at 5:58 AM on June 6


It's called calibration, and lesswrong.com is a good place to search for this. 1, 2.
posted by dhoe at 6:01 AM on June 6 [1 favorite]


Yeah, you could play this game a number of ways, but the way it worked was: give a range of 90% confidence; give a range of 50% confidence; tot up how often the true value falls into each range; gloat.
posted by grobstein at 6:08 AM on June 6


Thanks, grobstein. I don't recognize all of those, and I remember the test having exactly 10 questions, but as I said I didn't need the exact test anyway. This is pretty close (and Wits & Wagers is a great game).
posted by payoto at 6:59 AM on June 6


There's also a book "Super Crunchers: Why Thinking-By-Numbers is the New Way To Be Smart" by Ian Ayres that discusses number crunching and confidence intervals, etc.
posted by santry at 1:28 PM on June 6


This sort of problem is usually known as a Fermi Question, after the physicist. The original question was "How many piano tuners are there in Chicago?". These are really common in certain engineering or business interviews to give the company a way to try to see how you think. There is a website called fermiquestions.com that has a lot of them, but I haven't used it and can't vouch for it.
posted by tau_ceti at 4:25 PM on June 6


« Older I now have two iPhones - my pe...   |  I am by nature a restless, fid... Newer »

You are not logged in, either login or create an account to post comments