Chew (CDP) 1170 6.9%
Tinley (ALP) 9123 53.8% 1.7% 60.0%
Harper (GRN) 5177 30.5% 13.8% 40.0%
Georgatos (IND) 1478 8.7%
TOTAL 16948
COUNTED: 76.8%
BOOTHS (OF 12): 12

Sunday. I’ve identified nine previous by-elections which were contested by Labor and the Greens but not the Liberals – four federal in which Labor was in opposition, and five state in which they were in government. The average result overall was a 0.11 per cent primary vote swing against Labor. In the state by-elections it was 0.9 per cent against Labor; in the federal by-elections it was 0.88 per cent to Labor. Six of the results were worse for Labor than Willagee, while only three were better.

However, I’d argue that one of these by-elections didn’t fit the mould. That was the Isaacs federal by-election in 2000, caused by the death of Greg Wilton. All the others were like Willagee in that they were caused by the voluntary mid-term departure of the sitting member, which demonstrably leads voters to be unhappy with the party concerned. Sure enough, Isaacs was Labor’s best result out of the 10, with their primary vote increasing 8.1 per cent. Take that out of the equation and Labor on average suffered a 1.14 per cent swing overall, or 1.53 per cent against when limited to the federal by-elections conducted while they were in opposition. Only two of eight results were better for Labor than Willagee, against six worse.

8.19pm. It’s in. With the notional two-party result with all booths counted plus (I presume) the evening’s supply of postals and pre-polls, the WAEC’s notional 2CP result is 60.53-39.47. My rough yardstick for the evening had been 50-30-10-10, so Labor can feel pleased, particularly with the psychologically important achievement of having improved their primary vote. My expectation of a Greens primary vote of 30 per cent factored in that it had been an unhappy campaign for them, but their 30.6 per cent is nonetheless a reasonable result that again demonstrates Liberal voters’ willingness to vote tactically. I’m told the Gerry Georgatos campaign had a fairly low profile at polling booths, which probably helps explain the surprising fact that the Greens got more preferences than Labor despite both minor candidates directing against them on HTV cards. Much of his support would have come from Liberals parking their vote with the only available independent and following their normal habit of putting Labor last, and he equally has a support base among natural Greens sympathisers.

8.04pm. I gather we’re still awaiting one booth’s notional 2CP count, and that will be it for the evening.

7.52pm. To clarify, the WAEC’s “2CP count” obviously refers to the full distribution of preferences, which will not be conducted until all the votes are in. They would do better to call it that.

7.45pm. Final two booths, Coolbellup Primary School and Southern Districts Senior Citizens Centre, have reported, respectively giving Labor a relatively poor and relatively good result.

7.43pm. Looks like the Greens are doing quite a bit better on preferences than either I or Antony Green had estimated. With real figures to play with, my 2PP figure for Labor has gone down from 63.0 per cent to 60.4 per cent.

7.37pm. The WAEC have outsmarted me. I had been hitting refresh on their “two-candidate preferred” page and coming up with nothing. It turns out they have a separate page called “notional distribution of preferences”. I will be interested to learn what the distinction between these two concepts is. No polling booth breakdown is offered.

7.33pm. The Greens might have spoken too soon in claiming victory there – 306 to 297 in favour of Tinley. Nonetheless, it’s given them their biggest primary vote swing of 23.0 per cent. Continuing the trend of Fremantle, evidently a lot of Liberals are happy to thumb their nose at Labor by parking their votes with the Greens.

7.32pm. Anglican Church of the Holy Cross Hall in Melville added.

7.25pm. Labor primary vote up slightly at Samson Primary School; just keeping their nose in front in the race to improve on their 2008 primary vote.

7.24pm. Samson Primary School added.

7.19pm. Greens Twitter feed reports they have won the Anglican Church of the Holy Cross Hall, which presumably means on the primary vote. This is Labor’s weakest and the Liberals’ strongest booth in the electorate. The other strong Liberal booth is the just reported Melville Senior High School – with a lot of slack to be taken up here there were solid primary vote swings for both Labor and Greens.

7.18pm. Melville Senior High School and Southwell Primary School added.

7.15pm. Labor vote down 5.1 per cent in Hilton, their weakest result yet.

7.13pm. Hilton Primary School and 933 pre-polls added.

7.08pm. East Hamilton Hill the first polling booth to give the Greens a single figure primary vote swing – however, this was a particularly poor booth for the Liberals, so there was less slack to be taken up. Labor down 2.3 per cent; touch and go whether they’ll break even.

7.07pm. East Hamilton Hill Primary School added. Still no real 2PP results.

7.02pm. Southern Districts Senior Citizens Centre (Drive-in) sees a lot more business this time around – 126 votes compared to 18.

6.59pm. Labor down slightly in both, Greens up 13.0 per cent and 17.9 per cent.

6.58pm. Caralee Community School and North Lake Senior Campus added.

6.52pm. Phoenix Primary School in – Labor down 3.6 per cent, Greens up 11.7 per cent.

6.48pm. All three sources have Labor up on the primary vote, though by wildly varying amounts. Same goes for the Greens – if their increase in Palmyra from 16.4 per cent 35.8 per cent is typical of the polling booth results, they will make up a lot of ground from the present scoreline.

6.47pm. 1658 postal votes added (this won’t be all of them).

6.46pm. Palmyra Primary School booth added.

6.35pm. A note of explanation. Vote numbers and the percentage figures to their right are raw votes. The swing and 2PP figures are derived through booth matching, so Tinley’s primary vote from special institutions and hospitals is 22.4 per cent higher than Carpenter’s was. The 2PP figure is based on a guesstimate that Liberal preferences would have gone 80-20 in favour of the Greens in 2008, and that this time CDP preferences will go 70-30 to Labor and Georgatos’s will split 50-50. On that basis, there has been a 25.2 per cent swing to Labor from special institutions and hospitals in Labor-versus-Greens terms. The CDP and Georgatos preference splits I’m using will be superseded by real preference splits when notional two-party figures start to come in.

6.31pm. Special institutions and hospitals are in. Good enough for me – I’m calling it for Tinley.

6.10pm. Welcome to the Poll Bludger’s live coverage of the Willagee by-election count. First figures should be through in about 15 minutes or so.