by

# Not cray!

K and I are obviously not the only ones doing this prediction caper for the Australian election. Aside from the AFR coverage using our stuff, we enjoy reading Simon Jackman’s stuff in The Guardian. We dont know Simon but he’s got a stellar background – professor at Stanford in both the politics and statistics department. Some of his research uses Bayesian techniques which is stuff K and I have been reading.

On Aug 2, Simon Jackman had an article in the The Guardian about betting markets. He uses effectively the same simulation we do – get implied probabilities from betting markets, runs lots of simulations and get the distribution of outcomes. Based on his 1 August simulation, he predicts the ALP will win 61 seats. This is a few less than our model predicts but in the same ballpark, unlike the national polls which are effectively predicting the ALP winning around 75 seats.

Why does Simon’s predictions differ from ours? From what we can tell, the only difference is that he is using data from both Centrebet and Sportsbet, and then averaging the implied probability. We’ve only got data from Sportbet. It may also be possible that he is not correcting for the longshot bias that we discussed in a previous post. This might mean that in his model, seats that will likely end up in the ALP column are given to ‘other’ candidates. If you look at our predictions before we corrected for the longshot bias, we had thought the ALP would win around 60-62 seats too. Anyway, after reading Simon’s article, it’s good to know me and K are not cray. We’re going to enjoy this song and the rest of our weekend. Hope you do too!

## 2 thoughts on “Not cray!”

1. I am also keeping track of seat betting alongside polling on a weekly basis (eg this week’s example http://kevinbonham.blogspot.com.au/2013/08/poll-roundup-and-seat-betting-watch.html). However I am not running any Monte Carlos or similar, just tracking the basic measures of the numbers of seats in which each party is favourite and the number of “close” seats (which I define as both parties under \$3 on either Sportsbet or Sportingbet). The aim of the exercise is not to use betting to predict outcomes, but to test how seat betting fares as a predictor in the current and unusual circumstances. In particular, are betting markets accurately projecting where the polls will go? So far, after several weeks of the seat markets tending to move towards the polls, this was the first week in which the polls have moved towards the markets.

Checking Simon Jackman’s assessment of the seat of Brisbane it did indeed appear that he was not adjusting for longshot bias (indeed I was able to exactly replicate his Brisbane probabilities on that assumption). I think this will be the primary source of the difference between his expected total and yours.

A question I would however ask about your Monte Carlos: are you treating the different seat outcomes as independent of each other? Because they’re not, and the (smart) punters know this, and we know they know this because the first Australian bookies to offer “multis” on electorate odds several years ago got stung by the punters and lost money. (It was all rather funny actually; I forget who the victim was.)

If Labor is at \$3 in one seat and \$4 in an adjacent seat, and it wins the \$4 seat, it’s quite likely this is because a regional or national swing has occurred, greatly increasing the chance of carrying the \$3 seat as well. This shouldn’t affect the mean or median outcomes significantly but it has a larger impact on the probability spread and a massive impact on the underdog’s chance of overall national victory.

• Leng

hi kevin. thanks for leaving a comment on our blog! we’ve enjoyed reading the stuff you’ve come up with. re your comment, what good timing! we’ve been working on tweaking our model to deal with the independence assumption for a while. we’ve just posted the results with the new model. thanks for your comment.