I was Wrong about Cowan - 78 seats now possible

In my last post, I'd all but assumed Cowan would go to Labor, but in a completely unexpected twist the first absent votes have swung more to the Libs than even the postal votes. My model's now predicting it as more-likely-than-not a Coalition victory, and this is a quite conservative model as it assumes different non-ordinary vote types (e.g. absent and postal votes) will swing together. If this early trend in absent votes isn't just a fluke, it could be a very comfortable LNP victory despite being marked as a probable Labor win by analysts.

That would mean if Herbert and Flynn fall to the LNP as expected (by me, but I could easily be wrong again!), their promising Capricornia results continue on, Forde goes their way as also expected by analysts, and absent votes get them over the line in Cowan - 78 seats is now looking, if not likely, then at least far more feasible than it did last night.

Battle of the Forde

I'd actually managed to stay away from election stuff today but it drew me back with a couple of articles from the ABC, with Antony Green and Barrie Cassidy weighing in on the situation. My prediction of 76 or 77 seats is looking pretty likely, but the thing that interested me is Green and Cassidy both suggesting that the Coalition is "on track" and "too far ahead" in Forde respectively.

Now, Herbert and Flynn look to be pretty much wrapped up for the Coalition, and Capricornia is still close but may well head in that direction after a better result their way in recent counting, giving them a likely 76 seats. That is by no means certain though, as the graph below may help demonstrate (as in previous posts, the Libs win if the red line is above the black horizontal line at the right side of the graph; current progress based on counting is red crosses; simulated possible futures are red wibbly lines; results if non-ordinary votes behave like in 2013 relative to the ordinary votes are in black; and the track needed for the Coalition to win in blue)



Forde, on the other hand, is looking very interesting. The Coalition is probably more likely than not to win but after a relatively strong showing by the ALP in recently counted absent votes (the downwards red kink) there's not a lot in it, and even with a few data points my simulations are still showing a lot of variation that could easily bounce either way in the remaining votes. So I'm surprised at the analysts' confidence, especially given they've been relatively cagey on other results. Maybe there's something I'm missing?



And then there's Hindmarsh, which is expected to go to Labor and is looking good for them, but it's still early days...



All of this together means that while the 77 seat scenario is looking likely, any result between 75 and 78 is still very possible depending on these three seats. And then the only close seat not mentioned is Cowan which is an expected Labor win but anything can still happen...

Postal votes coming in

As the postal votes trickle in for the remaining seats, the situation seems to be getting firmer (see previous analysis here) on top of the 72 locked-in seats for the Coalition - Capricornia remains on a knife edge but the remaining seats are looking more and more like Coalition wins except for Cowan and Hindmarsh (likely to go to the ALP).

I've cobbled together simulations for each seat of 100 potential non-ordinary vote futures (thin red lines below), all of which go to the LNP in Flynn, Gilmore (already called for the Coalition on ABC) and Herbert. In Forde it's less clear but still 90 simulations are predicted the LNP's way; in Capricornia, 54 pick the LNP making it a real toss-up, and Hindmarsh has only 31 to the LNP and Cowan 6. I may explain how I put the simulations together in a later post.

According to this so far (and things can easily change - my assumptions are cobbled-together and by no means rock solid), the Coalition is looking pretty well set for 75 seats, likely for 76 and still very possible for 77 seats.

The 77 seats scenario

Now the postal votes have started coming in, I've been casting around for a way to visualise what's going on. One way I've found useful is, as in my previous post, by using the predicted swings in non-ordinary votes (postals etc) compared to ordinary votes, using the 2013 results as a guide - these swings tend to go to the Coalition. In the plots below, I've put those predictions in black, the relative overall swing needed on top of the postals swing for the Coalition to hit 50-50 in blue, and the data so far in red. I've assumed the postals will come in first (they're all we have so far) and the other vote types afterwards (absent, pre-poll and provisional in that order - this probably isn't accurate!).

In the seat of Forde, my earlier model currently predicts an LNP win by about 400 votes despite them only being ahead by 94 votes after ordinary votes. Though commentators were surprised by a "Coalition surge" as postal votes started coming in, it's actually quite consistent with 2013 results (see the red cross).



I've checked a bunch of other seats that my model predicted would be close (Batman, Chisholm, Cowan, Flynn, Gilmore, Herbert and Hindmarsh) and most of the other results have been relatively consistent with 2013 patterns as well (if not better for the Coalition), currently confirming my 76 seats prediction - except for one. Herbert seems to be looking particularly good for the Coalition, well above where they need to be to make up the difference (see graph below - they're well above the predicted trend, and even well above the needed blue line).

I can't stress enough that it's early days, other seats with postals yet to count could well swing the other way to the ALP, and the number of remaining votes could be well out, but if current trends continue, they could pick up an extra seat to make 77. Interesting times ahead...

Update #1: Things are looking worse than expected for the Coalition in Capricornia - it looks too close to even guess one way or the other. As it stands currently, on top of the ABC's current 72 to Coalition, 66 to ALP:

Probable LNP: Flynn, Forde, Gilmore, Herbert
Probable ALP: Cowan, Hindmarsh
Complete guess: Capricornia



For interest, this is why I think Flynn is a probable LNP despite it having been in the ABC election ALP column until recently:

Election results madness

With help from the AEC scraping code by Mick McCarthy here, I've tried to predict the election results once the "non-ordinary" votes (early, pre-poll, postal and provisional) have been counted. These tend to go more for the Liberal party, so seats that Labor is currently ahead in might be lost. 

I’ve assumed that:

• all ordinary votes are in (only non-ordinary votes left), 
• the proportion of formal votes remains the same as in 2013 for all electorates, 
• the number of non-ordinary votes all increase proportionally with each other, 
• and the changes in 2 candidate preferred votes between voting types (I call this "bias") in each electorate stay the same between 2013 and 2016 (e.g. if the Libs get a 2% bounce in postals in Denison in 2013 compared to their ordinary vote result, that same proportion holds in 2016).

It turns out that William Bowe of the Poll Bludger, my favourite election analysis page, pipped me to the post with his analysis here, but I thought I'd run mine anyway. And it gave similar results, but importantly different enough to get the Coalition over the line. Here are the number of votes that the LNP (Liberal/Nationals party, well, Coalition) are expected to be ahead or behind in the closest seats:

Capricornia LNP +547

Chisolm LNP +1230

Cowan LNP -580

Flynn LNP + 1469

Forde LNP +86

Gilmore LNP +713

Herbert LNP -315

Hindmarsh LNP -786

Melbourne Ports LNP -1224 (I didn't do 3CP analysis though...)

Petrie LNP +1447

The Poll Bludger's analysis has them behind in Forde by 18 votes - my extra seat gives them 76 seats, a majority in their own right, whereas his 75 seats is not enough.

I then had a play with adding random variation, relaxing the assumptions of the proportion of formal votes staying the same and the amount of bias staying the same. I assumed that the overall variance between electorates for these properties stayed the same, but that a little (10%), half (50%) or all (100%) of the variance was due to random variation and not the specific effect of being in one electorate or another. The less important electorates became, the less likely a Coalition victory - for a little, 66% of simulations resulted in a win; for half, 34% and for full random variation, only 20% of cases.

Edit: Using 2010 results instead give us similar results, though slightly worse for the Coalition - 76 seats for the Coalition without variation, and 53%, 20% and 16% in the three scenarios described respectively.

Edit #2: Using 2007 results is more difficult because of seat redistributions, but we can do it if we assume that Liberals and Nationals will experience the same swings in Capricornia and Flynn. The model also gives 76 seats for the Coalition without variation, and 91%, 49% and 27% for the scenarios.

The model also pointed to some seats that could potentially be very close (winning 2PP < 50.2%) in these situations, so watch for these potentially coming into play if things get even more interesting:

likely:

Batman (VIC)       

Longman (QLD)

possible:

Banks (NSW)

Dickson (QLD)

Dunkley (VIC)

Griffith (QLD)

La Trobe (VIC)

Lindsay (NSW)

Robertson (NSW)