How much would it cost to melt all the snow in Buffalo?

February 17, 2010 2:55 PM

A Diary of Numbers: Physicist Aaron Santos, author of How Many Licks (Or, How to Estimate Damn Near Anything), uses science and math to estimate answers to life's greatest questions, including: How likely is it that we inhale a molecule breathed by Carl Sagan?; and What costs more: the Silverdome or the number of peanuts needed to build a life-size replica of the Silverdome?

And a question from author Christopher Moore: If your body could be reduced to its pure sugar content, how long would it take 1000 hummingbirds to drink you as nectar?

Dr. Santos discussing the thought behind the books and some questions (including how many people are having orgasms at this very moment?) on the Podcast Skeptically Speaking.

posted by sarahnade (25 comments total)

I love this game!

How many different frames of video are in the entire series of Dr. Who?
posted by Bathtub Bobsled at 3:11 PM on February 17, 2010

I don't think this can even brush the surface of how much wood a woodchuck would chuck if ia woodchuck could, in fact, chuck wood. And the answer is not "as much as a woodchuck could".

*ZOT*
posted by kmz at 3:23 PM on February 17, 2010

I'm breathing Saganized air! My day just got better.
posted by Roman Graves at 3:25 PM on February 17, 2010

One cubic meter of peanuts is worth at least three chairs. Chairs have a lot of empty space in them.

When the basic assumption underlying your whole estimate is that far off, your estimate isn't worth peanuts.
posted by darksasami at 3:29 PM on February 17, 2010

I love this kind of stuff. Site bookmarked. Thanks for the post.
posted by Splunge at 3:34 PM on February 17, 2010

The bad news: That Sagan molecule you breathed was sneezed by Hitler.
posted by weapons-grade pandemonium at 3:42 PM on February 17, 2010 [5 favorites]

There are some big assumptions about atmospheric mixing built into the Carl Sagan's breath thing, and anyway, surely it'd be better to substitute Fermi for Sagan in this case.
posted by overyield at 3:48 PM on February 17, 2010

That site needs a professional white background. The links just look like blurs and they hurt my eyes. Waaa!
posted by ekroh at 3:51 PM on February 17, 2010

According to homeopathic principles, by now I should be Carl Sagan.
posted by joshwa at 3:52 PM on February 17, 2010 [2 favorites]

One cubic meter of peanuts is worth at least three chairs. Chairs have a lot of empty space in them.

When the basic assumption underlying your whole estimate is that far off, your estimate isn't worth peanuts.

He said "square meter". Presumably, the chair would be constructed from one-peanut-thick sheets of peanuts. A square meter of peanut planking seems about right.

That said, 4 kg for a chair seems a bit light. I'm thinking you'd need at least double-ply, and maybe triple-ply peanut planking to build a chair that would support a person. Order of magnitude, though...
posted by mr_roboto at 4:13 PM on February 17, 2010

I don't think this can even brush the surface of how much wood a woodchuck would chuck if ia woodchuck could, in fact, chuck wood.

According to Wolfram Alpha, 361.9237001 cubic centimeters per day.
posted by Paragon at 4:14 PM on February 17, 2010 [2 favorites]

D'oh! Ok, there I go again, trying to add an unnecessary dimension to everything.
posted by darksasami at 4:16 PM on February 17, 2010

That said, 4 kg for a chair seems a bit light. I'm thinking you'd need at least double-ply, and maybe triple-ply peanut planking to build a chair that would support a person.

Yeah, but the question wasn't about structural integrity, it was about the volume of peanuts needed to build a replica.
posted by dirigibleman at 4:28 PM on February 17, 2010

Also, why is he using retail peanut prices? At the kind of quantities we're talking about, he should be doing commodities trading.
posted by Horace Rumpole at 4:30 PM on February 17, 2010

The bad news: That Sagan molecule you breathed was sneezed by Hitler.

That's OK. Hitler's sneeze was farted by Bach.

whoa
posted by jquinby at 5:23 PM on February 17, 2010

I remember reading somewhere sometime long ago that in any given litre of water that you consume, you are likely to be drinking at least one molecule of a victim of the Titanic sinking. Shortly thereafter, I discovered beer.
posted by Mike D at 5:53 PM on February 17, 2010

Good idea! There's nothing better to mask the flavor of Titanic victims than yeast shit.
posted by Malor at 6:32 PM on February 17, 2010 [2 favorites]

This is great, reminds me of a biophysics prof. I had who, sensing that the class was not up to snuff on statistics and probability, taught us a couple semesters' worth in a week using real-worldian examples (eg, calculating the odds of winning the lottery). Then he gave us loony problems like developing an imaging device that shoots pumpkin-sized particles and figuring out what size object it could resolve (elephant-size, iirc).
posted by Mister_A at 7:02 PM on February 17, 2010 [1 favorite]

These things tend to come in two varieties -- one is the basically frivolous jellybeans-in-a-jar style estimations, how many pancakes are in a stack to the moon, etc., which sort of help develop general numeracy but always have the faint whiff of mental wankery. The second is those that actually help develop physical intuition about the world -- how much power is your body radiating right now; how high could the energy in a D cell lift you; etc.

I always thought Fermi problems were unfairly pigeonholed because everyone always remembers the silly one (how many piano tuners are there in Chicago) and forgets that Fermi used this style of estimation to do real work. The classic case here is his correct, and fast, estimation of the Trinity bomb yield based on a handful of dropped confetti.

A friend wrote his PhD thesis (link to PDF at bottom of page) on the second type and how you can use this kind of thinking to develop intuition. In this case one of the best example is "How much power does it take to ride a bike?" -- it turns out there's almost a dozen different ways to tackle the problem (aerodynamics, mechanics, food intake, physiology, etc.), so getting to an answer is much less a matter of the facts you know -- virtually everyone knows enough to solve the problem -- than it is a matter of knowing how to systematically apply whatever knowledge you already have, a skill that's much harder to acquire and one we're tragically bad at teaching in schools.
posted by range at 7:07 PM on February 17, 2010 [2 favorites]

it is a matter of knowing how to systematically apply whatever knowledge you already have, a skill that's much harder to acquire and one we're tragically bad at teaching in schools.

In all fairness, it's a difficult skill to teach, especially in a classroom setting.
posted by invitapriore at 8:06 PM on February 17, 2010

In all fairness, it's a difficult skill to teach, especially in a classroom setting.

Oh, for sure -- I work in a group that's actively working on (new) ways of doing this, and we're still hunting. We just did this sort of estimation exercise with 150 20-year-old engineering students, and when they had a rough time with it we turned the tables and offered to play "stump the chump" with us working out things they proposed live. First question? How many piano tuners etc. etc. The jellybeans-in-a-jar questions are fun but they seem to put these skills in the "party trick" bin¹ in people's heads, rather than the "useful tool" bin, and so they don't reach for them when they have an actual problem to solve.

Partly I think we need to teach that Knuth wasn't really joking when he wrote that a technique is a trick you use twice.

¹I say this as a proud contributor to Geek Party Tricks...
posted by range at 8:53 PM on February 17, 2010

Despite the title of his book, I can't seem to find the answer to the Lil' Kim song anywhere.
posted by Joakim Ziegler at 9:26 PM on February 17, 2010

range, this might be something you've already been over a million times, but what about taking the lab format a step further? The labs that I did for my computer science classes were often guided to the point that the strategies and processes necessary for solving the problem were completely obvious if you paid attention during class. This is me pulling something out of my ass, but what about making labs more difficult and more free-form? I imagine the grading process would have to be more lenient, but putting more obstacles in between the progression from learning about the tools to using them effectively in novel situations seems like a way to encourage that sort of thinking.
posted by invitapriore at 10:07 PM on February 17, 2010

That Skeptically Speaking podcast is excellent. Just in case people skipped past the link.

(Though their feed has horrifically broken formatting)
posted by srboisvert at 2:13 AM on February 18, 2010

invitapriore, that's really the important part -- in school we get trained that every problem has a specific solution method, and even in "lab" we usually know that we're there to apply Solution Method X. One of the things we try to do is introduce real-world problems with real-world complications, with less guidance about solution method and more guidance about process. You're completely correct that this puts more of a burden on the assessment side, but mostly it forces the teacher to think more critically about what they're really trying to teach -- sometimes we fall into the trap of grading based on things that are easy to measure, rather than what we're actually most interested in the students learning. Usually, if you can clarify what you really want the students to be able to do, assessment becomes easier, even in "messier" projects.
posted by range at 9:17 AM on February 18, 2010