Sunday, January 22, 2012

Not an indicator

Probably the only thing that irritates me more than people having zero understanding of history is when people completely misunderstand history and read into it all sorts of nonsense, almost as if history itself is some sort of magic.

We’re seeing this right now with the South Carolina Republican Primary in which people are saying that because a pattern exists, it therefore predicts what always will happen. This isn’t about the politics of that election or the campaign generally, but the absurd declaration that in the years since Reagan won it, no Republican has won the Republican presidential nomination without first winning the South Carolina primary, and so, Mitt Romney is in trouble. What a load of codswallop.

To be sure, Romney has many problems—his elitism and the fact he’s a Mormon are chief among them—but losing South Carolina is not one of his problems. What’s happened in South Carolina since 1980 is a coincidence, possibly somewhat interesting, but nothing more than that.

For the meme to be true, it would mean that all Republicans in all the other states would have to say, “gosh, Newt won South Carolina, so we have to vote for him now.” That’s just absurd. Or, it would require some sort of magic spell, because nothing else could bring about something as absurd as having Republicans in more mainstream states vote according to the whims of South Carolina Republicans.

A meme we heard earlier was that in recent elections, the winner of the Iowa Republican Caucuses didn’t often go on to win the Republican nomination, a meme the news media quickly dropped when their anointed frontrunner, Romney, won. I suppose that now that the final votes have given the victory to Santorum it means that they’ll quote it again. I saw an AP story that said, “this the first time in history a different candidate has captured a win in each of the first three presidential nominating contests,” as if that is in itself a shocking fact, as if history’s magic spell is broken.

This sort of thing isn’t unique to politics. I’ve often seen sports commentators say equally absurd things, like, “Over the past 24 years, Team A has never won a game against Team B at Big Huge Stadium”. They forget to mention that the scenario they set up covers only 4 games over that time and the fact Team A lost all four isn’t weird at all.

I know that people can come back at me with longer streaks, ones that seem to defy probability, but human behaviour doesn’t always fit neatly within the laws of probability—which, by the way, also include the possibility of the improbable happening, like long losing streaks or historic “patterns” in election wins and losses.

Superstition can be fun—I always throw a pinch of spilled salt over my shoulder (in my case, to feel a connection with my superstitious ancestors). But superstition is no way to pick someone to vote for or to “predict” future voting behaviour. It is, no matter how long the pattern goes on, nothing more than irrational superstition.

Romney may not ultimately be the Republican nominee, and Newt might not be, either. But when whoever wins that nomination steps up to the podium of their convention to make his acceptance speech, I can absolutely guarantee his doing that will have nothing to do with who won the South Carolina Republican primary over the past three decades. Yes, the past is prologue, but it’s absolutely not an indicator.


Roger Owen Green said...

I was having the very same conversation about the "predictive" value of anything. Take the stock market - the meme is that, over the long haul, you'll end up ahead. But that's based on the past, not the unknowable future.

Arthur (AmeriNZ) said...

That's a really good example—and one I hadn't thought of before, though I should have.

NZ used to have a lot of finance companies selling what they called "first ranking debentures", which were basically shares that were second only to creditors if the company collapsed. They used to advertise their fantastic rates of return, many times higher than anything else available, and they always included a disclaimer somewhere with words to the effect of, "past performance is not an indication of future returns."

As it happens, virtually all of those finance companies collapsed, taking the life savings of plenty of ordinary Kiwis who saw the very high returns and didn't hear—or chose to ignore—the disclaimers. Some were, no doubt, just plain greedy, but others failed to grasp that what happened in the past is just that, and the future is always unknown.

I think we should all be more sceptical of claims of future events that are based on history. You reminded me of an example I'd forgotten about.