Interesting Number Sets Constantly Coming Out for the Next Six Weeks, and How to Use ‘Em

Ralph Fielding is the guy who took the ever-lovin’ math right outta me. Ralph was my math teacher in 1970-something, when I was a freshman in high school. It might have been Geometry — I don’t remember the exact topic.

What I *do* remember is that Ralph had an odd mouth and was a bit of an odd character, and that, when I was done with his class, I was done with math. Done done. I never took another math class in high school or college.

That’s too bad because decades later I came to know that there is something I very much love about numbers (I’m pretty sure it’s their “this is absolutely true” nature) — decades to get back in touch with my inner math-head, to enjoy seeing what groups of numbers have to say about the way the universe is built, at the very foundational level.

*  *  *

Right now we as a country are doing our once-every-fourth-year, new-numbers-coming-out-all-the-time, run-up to what most folks believe is a very important, and ultimately very numeric event — a nationwide vote for President.

For those who think that the Billy Bean brand of baseball is all about how baseball generates tons of numbers every day, from which much power and mastery can be derived if well understood, gosh a’mighty have I got a number set for you!

*  *  *

People who know me as a math-head and presidential election junkie (why all the drug references in this realm?) are asking me these days what’s going on with the election. For them I have a few sites to suggest, like ah so:

The grand-daddy, grand-mammy and rampaging-offspring-hoard of elections number-crunching sites is (caution: New York Times, semi-permeable paywall membrane awaits you when clicking on that link). Nate Silver, its founder, is a Billy Bean sort of number-cruncher who swings both ways — as a sports number-cruncher, and as a polls number-cruncher.

Nate compiles lots of polls, in real time, and has a massive piece of software that he designed which figures out what the polls, when smartly aggregated have to say, and then outputs some predictions. Here smartly aggregated means accounting for, e.g., each pollster’s biases and past performance, the correlations among all the states’ voting patterns, the effects of economic metrics announcements such as monthly jobs numbers, etc., etc., etc.),

Today,for instance, Nate has Obama with a 77.7% likelihood of winning, and forecasts that Obama will receive 309.3 electoral votes (270 are needed, out of 538, with this latter number accounting for the name of Nate’s site).

As time passes, and the election gets closer, predictions necessarily gain accuracy (clients of mine hopefully recognize this as a Reverse Cone of Possibilities). So whatever Nate says will happen the day before the election usually does happen the following day. Four years ago his final predictions got the electoral vote count correct with one exception (he had Indiana going for McCain) and all Senate elections correct.


That’s the first place I look at each day. The second, which is great for an up-to-date list of current polls, is at Real Clear Politics. This is a mostly conservative-leaning site (see the articles on the front page — I am pretty sure they would agree with that characterization) which has lots of intriguing subsites (e.g. RealClearScience, etc.) which appear to be very well done. They also have some great state-by-state tools, such as a handy-dandy electoral map tool.

For a simple, fresh-each-day, few-minute-long read, is a great place to go. It currently sees Obama getting 328 electoral votes.

*  *  *

More Sites.

Those are the three sites my math-head/political-junkie self wants to look at each day for the next six weeks. These are also good to look in on from time to time:

  • Intrade (where people bet real money on, among other things, the outcome of the presidential election — based in Ireland because you can’t build that here) (today it shows Obama with a 72.5% chance of winning)
  • Princeton Election Consortium (using a “meta-margin” analysis — a margin of the margin analysis) (today it shows Obama with a roughly 5% meta-margin)
  • Votamatic (don’t let its vegematic name throw you off — there is lots of good info to be found here) (today it shows Obama getting 341 electoral votes)
  • RAND American Life Panel (uber-L.A.-think-tank/business consultancy using unusual longitudinal approach to predictions) (today it shows considerably more than a 95% likelihood that Obama will win)

So, gang, if you want to get in touch with your inner math-head, while also smartening-up about a very important, very real political phenomenon, these are great places to start.

And maybe, just maybe, if you dig a little deeper in there and read about methodologies, you’ll catch a glimpse of some of the marvel of the universe that is math — which, here, allows us to measure huge, complex phenomena, within a fairly tight arc of accuracy, by smartly aggregating some handfuls of thousands of observations and then knowing, from just those few parts, something very real about the whole — though something at least a tad short of a perfect representation of the whole.


802 words


Leave a Comment