Comment: Don't stop the votes

Sophie Mirabella has quit her frontbench role as she continues to trail Cathy McGowan in the seat of Indi

The race to the Lodge may have been run and won, but the counting continues - and that's where things get really interesting, writes Paul Taylor.

While for most, the election (and the watching of the results) was done and dusted on Saturday night, for political tragics and statistics nerds like myself the results from election night are only the beginning. It's now been nearly a week since election day and with counting still going in every seat a few people (myself included) are keeping an eye on the results in the seats that are still too close to call and tweeting numbers every time there's an update.

The two seats with the most interest on Twitter are Indi - currently being led by Independent Cathy McGowan over the Liberal's Sophie Mirabella, former Shadow Minister for Innovation, Industry, Science and Research - and Fairfax,  where Mining Magnate Clive Palmer is leading the LNP's Ted O'Brien. 

Indi is a fascinating example of how a likeable candidate against an unpopular sitting member can attract attention from across the country. The hashtag #indivotes was among the top trending topics in Australia for a few days and will trend again once it becomes obvious who will win. The interest in Fairfax is obvious: Clive Palmer is famous and is a colourful character. Love him or hate him people will always talk about him.

There are other close seats that aren't attracting as much attention. All are Labor held and all but one look like they will fall to the Coalition: Capricornia, Eden-Monaro, McEwen and Parramatta. 

Thanks to the type of data the AEC publishes we can attempt to "predict" who will win a close seat before all the votes are counted by assuming the two candidate-preffered figure for a certain type of vote will hold, and using maths to work out the vote difference after all those votes are counted.

This method is for those of us who don't have a computer program we can feed the raw data into and have a result calculated for us.  For example, in Capricornia the AEC shows us that of the 4,981 postal votes counted so far 2.85% were informal and of the formal votes the LNP is winning the 2PP 59.66%-40.34%. If we assume this trend holds for the remaining 5,000 or so postal votes and do the maths with the firgures for absentee votes we can come up with a rough estimate of the winning margin.

In the table below I've done just this for the seats I've mentioned. Let's see how close we get to the actual winning margin using this method.

Seat

Current Margin

Predicted Margin

Fairfax

PUP by 1,132

LNP by 239

Parramatta

ALP by 389

LIB by 47

Capricornia

LNP by 624

LNP by 1,241

Eden-Monaro

LIB by 633

LIB by 220

Predicted margin assumes all the postal and absentee votes issued will be returned in time and does not include provisional votes due to big fluctuations in acceptance and rejection rates. 

Indi has not started counting absentee or absentee pre-polls yet. We're going to give them the same 2CP as ordinary votes. The predicted margin may therefore be quite off at the moment but for those of you interested those calculations show Mc Gowan winning by 698. The same applies to McEwen.

These figures will have a slight margin of error due to some votes not being received by the AEC in time, changes in the 2CP for a certain vote type and some votes being rejected instead of added into the count. For updates as the count continues see my twitter feed - @Australiavotes.

Paul Taylor is a self-confessed statistics nerd. This article was first published on his blog here.