As I noted back in October, betting on Australian elections was still a novel idea in 2004. But this year it was big business. At least three major online bookmakers offered odds on results in individual seats, as well as a variety of other combinations, and many commentators looked to them as a guide to voter sentiment.

I’ve been taking advantage of this opportunity, and over the course of the year I’ve staked a little over $4,400 in election bets. The results are now all in (with the exception of McEwen, where I’ve got $35 on Labor), so it’s time to balance the books.

I’m showing a profit of $1,092.25, or a return of about 25%. I’m fairly pleased with that, but I don’t think it shows an exceptional degree of either skill or luck; it’s more a matter of understanding how elections differ from more typical betting markets like horse racing.

A horse race is a random event in a way that an election is not; it’s much easier for things to go suddenly and unpredictably wrong for even the most fancied runner. That’s why many of the individual seat markets showed the sort of odds you’ll never see on a racetrack, with favorites at 50-1 on or even 100-1 on.

But even so, some of the odds-on bets were good value. I staked nearly $2,000 on bets at odds of worse than 2-1 on: the Coalition in Macarthur and Greenway, and the rest Labor (the shortest was 100-9 on for Labor in Corio). That would be a very risky way of trying to make money on the horses, but at the election every one of those bets came good, yielding a profit of $435.40.

Some of the initial prices on Labor-held seats were wildly unrealistic.

There was never the slightest chance of Labor losing seats like Ballarat, Lowe and Rankin, yet you could get odds of 5-1 on (that is, a 20% return) or better early in the piece.

The other place to make money was the marginal seats, where Labor was consistently doing better in both the polls and the overall market odds than the average of its prices in the marginals would suggest. I concluded that it was the individual seat markets that were wrong, and I bet mostly on Labor to win a range of marginal seats at prices from 2-1 on to 3-1 against.

Those bets produced losses of $697.15, easily outweighed by profits of $1,507.35.

The third category of bets was less successful: the long shots, at better than 3-1 against. It turns out they were long shots for a reason; three of them came home (the best being 10-1 for Labor in Dawson), but the $172 profit there was more than offset by losses of $325.35.

That included some I’d thought were good chances, such as Labor in Dunkley and North Sydney, as well as others that were frankly speculative – the longest was the Greens in Brisbane at 66-1. Horses do win at 66-1 from time to time, but candidates don’t.

Of course, all gambling involves risks – then again, so does the sharemarket. My experience mustn’t be taken as encouragement to bet more than you can afford to lose. But if you’re going to have a flutter, you might find that the ballot box has its advantages over the racetrack.