Funny things polls. Not always with explication, they can be all over the place, stretching the pollsters' credibility. But do gambling odds offer a more reliable guide?

Don't bother entertaining this idea at a pollsters' convention. Pollsters go all out to find representative samples for their data and to statistically validate their methods. They are understandably pissed-off at people who don't like the results they produce. All they do is take a picture - it's not their fault if the data displease. Now they have to hear, heresy, that betting spreads might be more accurate?

JM Keynes (rhymes with "brains" his father used to tell him) compared picking winning investments to a beauty contest - you don't pick the girl you think most attractive. You go for the one you think the others will vote for. That's what people do when their money is involved and it's reasonable to argue that odds thus reflect a candidate's "value" accurately.

It's a cute idea. But what happens when the poll is seriously out of whack with the spread?

Consider the most recent poll for Maryland, by Survey USA, which reveals a shocking deterioration of support for Kerry. The president is now tied with his challenger in a state that looked done and dusted for the democrats. On the face of it, a hard to believe change of trend, reflected by the spread for Maryland - Kerry to win 86-90.

Now check out the Maryland Department of Labor. Maryland is the number one ranked state for job-creation over the last year. That factor will have the greatest influence on the election result by the reckoning of the model used by this website.

Maybe it's me, but the possibility that the spread is lagging an obscure state dynamic deserves examination. Risk is low - it would be hard for the spread to punish a speculator from where it sits now.

Ah, did I mention that Nader just got on the Maryland ballot?

Bookmark and Share


Related Posts with Thumbnails