# The amazing singing banana – James Grime – does it again: What’s the probability you live in an odd numbered house?

Posted by: Gary Ernest Davis on: August 5, 2011

If only the accomplished, cialis clever James Simons (see previous video) were this entertaining.

James Grime stands a serious chance of turning people onto mathematics.

### 2 Responses to "The amazing singing banana – James Grime – does it again: What’s the probability you live in an odd numbered house?"

How are houses numbered where you live? In the US, odd number houses are on one side of the street and even number houses are on the other street. In Philadelphia, houses on the North side of East-West streets, and on the East side of North-South streets, are odd numbered. Not every street has the same number of houses on the even and odd sides. I wonder if the probability would be different if the Facebook poll reached mostly Americans? Mostly Philadelphians?

Well done, very entertaining vid, and very interesting topic. Two comments:

(1) You say, “What’s the probability that you live in an odd-numbered house?” Actually, that probability for me is zero!

(2) I like your analysis of a street having either an odd or even number of houses. However, the illustration, with a small number of houses, seems to equate “street” with “block”. And of course, looking at house numbers block by block has to allow for lots of variation, such as blocks that begin with even numbers.

I think a slightly more precise way of saying it would be to say that house numbers tend to be doled out in strips, e.g., 501-599 or 2801-2860, and that these strips of numbers usually start with an odd number, something ending in 1, and then end with either an odd or even number.

The second set of poll numbers you provide is: odd = 211,803, even = 209,964, for a difference of 1839, which is roughly 1/229 of the total of 421,767. If we assume that half the strips end in odd numbers, and half end in even numbers, then we would need an average strip length of about 115 to add those additional 1839 odd-numbered houses. I would go on to hypothesize that the true average strip length is less than 100 – just based on observations of how houses are numbered – and that the difference between the true value, and 115 – i.e., why we would need *more* strips to get our extra odd-numbered houses – is accounted for by the fact that more strips end with an even number than with an odd, by virtue of the fact that houses are most often situated in pairs facing each other across the street.