## We Love Math, Electoral College Department.

Andrew Sullivan says that this question-and-answer is why he doesn’t do math.

That Sullivan is quantitatively challenged is, of course, no surprise to anyone who reads his blog on a regular basis. But his excuse here is pretty lame.

The problem posed at FiveThirtyEight.com was “How many unique ways are there to acquire at least 270 electoral votes without any excess?”

The solution to that did indeed turn on a sophisticated application of combinatorial methods. According to the analysis by Isabel Lugo (posted at 5:22 p.m. on June 10), there are 51,199,463,116,367 different possible ways to accumulate 270, 271, 272 or 273 electoral votes.

Lugo’s solution does indeed demand both smarts and training, and she received her just due of praise from the comment thread. Certainly, though I can follow, gasping, the reasoning as she explained it, I can’t claim any greater chance of cracking such a problem than Sullivan could — which is to say, none.

But Sullivan’s surrender — “it’s just too hard, and look at the cute, big number” makes me crazy, and illustrates one of the persistant reasons why our discourse is so bad, why, as Brad DeLong keeps asking, we can’t get a better press corps.

That is: there is a difference between ignorance of advanced math (in which I take second place to no one), and an inability or unwillingness to master the basics of quantitative reasoning.

What’s remarkable, is how far you can get with not that much, just a basic disciplined approach to simple concepts — estimation, use of ratios and so on.

And with such simple tools it is possible to get a handle, if not always a precise result, even for such subtle, complex problems as the electoral vote question that so flummoxed Sullivan.

As Lugo pointed out, introducing her analysis — her exact number was anticipated by a much simpler simulation by commenter Brian at 4:43. Even if you don’t follow Brian all the way through the simulation, his exercise begins with a simple piece of arithmetic that gives the first hint of the scale of a likely solution, the fact that with 51 jurisdictions there are 2 to the fifty first power, or 2.25 quadrillion possible win/lose outcomes.

That’s enough to tell you from the start that you are dealing with a big number. The next steps take you further, and show how the simulation produces a plausible argument that the number of outcomes where the electoral vote totals hit the desired range (270-273) is going to come in at just a bit under three percent of that huge total number of outcomes, or right in the range of the 51.2 trillion outcomes that Lugo derived.

And my point is that whether or not you can imagine performing this bit of computer-mediated approximation, even the very first step, one that comes from high school math, is enough to get you into the right neighborhood, the right scale in which any answer will have to land.

It’s a necessary skill for any reporter today, I think, really any citizen. I won’t go here into the same riff I’ve blogged many times before. I’ll outsource instead to my new blog humor BFF xkcd:

**Explore posts in the same categories:**Journalism and its discontents, Mathematics, numbers, political follies, Politics

**Tags:** Andrew Sullivan, Combinatorics, Electoral College, math

June 12, 2008 at 6:06 pm

Yes, but it’s acceptable to say “I never was any good at math” and leave it at that.

More and more I am tempted to respond with “well, I never got the hang of reading” and se how they react. But I’m afraid they might actually believe me and I don’t want to seem stupid.

June 12, 2008 at 9:57 pm

Say “I never got the hang of talking” and they might get it. Might.

June 13, 2008 at 11:41 am

$15.05, what’s the tax? Does that include tip, is the tip % variable depending on how quickly the server can bring the appetizers?

June 16, 2008 at 11:43 am

What I find notable in these discussions is the lack of attention paid to the incredibly frequent presence of members of the Fibonacci sequence among the numbers of state electors. Nearly 40% of the US population lives in a state with either 1 or 1 or 2 or 3 or 5 or 8 or 13 or 21 or 34 or 55 electors.

This oversight is particularly egregious when all the participants have spent the last few days been running to their textbooks to find dynamic programming and recursive solutions to the original problem. What’s the second example (the first being computing factorials) that every such text gives?

Is this the harbinger of a golden age in American politics? Or another mean age?

MK