Fisher by-election live

Monday 9.30pm. The indefatigable Tom Richardson relates that the final margin in favour of Labor after the preference distribution came down to just nine votes: 10,284 to 10,275.

Monday 7.30pm. Tom Richardson tweets that the conclusion of adjudication of informal votes has ended, and that the result after finalisation of the preference distribution will be a 13-vote win to Labor. There would appear to be a good chance of the Liberal Party will seek to have the Court of Disputed Returns revise the ballot paper rulings that went against them.

Monday 2.30pm. The recount that was granted to the Liberals, somewhat surprisingly, is now under way. Nine’s Tom Richardson relates on Twitter that reassessment of informal votes has caused nine formerly informal votes to be admitted, resulting in the Labor margin to narrow from 23 votes to 20, with the “validity of several to be adjudicated”.

Saturday 7.30pm. ECSA has published the preference distribution. Woodyatt needed a 24.75% greater share of preferences than Labor in the three-party preferred count, but could manage only 16.92%. The differentials at the Greens and Democrats exclusions were actually 17.04% in Labor’s favour, compared with 27.26% in Woodyatt’s favour among the other candidates. Slightly less than a quarter of the preferences went to the Liberals.

Saturday 2pm. The ABC reports that Labor has secured its parliamentary majority of 24 seats out of 47, having somewhat outperformed expectations in the preference distribution. Labor’s victory over the Liberals on the two-party count will take effect after its candidate Nat Cook survived the last exclusion ahead of independent Daniel Woodyatt by, according to Tom Richardson on Twitter, 226 votes. It appears the preference distribution process turned up another two votes for the Liberals and one for Labor, such that the latter’s winning margin at the final count is now 23. Hopefully the full preference distribution should be published shortly on the ECSA site. I’ve changed the time stamp on this post to bump it to the top of the page.

Friday 4.30pm. Rechecking ends with Labor 24 votes in the clear, so the Liberals are out. The issue between Labor and Dan Woodyatt will be decided at the preference distribution to commence from 9am tomorrow.

Friday 1pm. It seems the preference distribution will be conducted tomorrow. Well-connected observer Independently Thinking relates in comments that Labor appears pessimistic about its chances of staying ahead of Woodyatt, although clearly there will be very little in it. He also passes on an entertaining account of the Liberals’ confusion on election night by Nine Network reporter Tom Richardson at InDaily.

Friday 12.30pm. The last postals have been counted, of which there were only 30, and they have made no difference at all to Labor’s 21-vote lead. So unless anomalies emerge in the preference distribution or any recount that might be conducted, we can now say that the Liberals have not won the seat. Rather, the issue is whether it’s Labor’s Nat Cook (5501, 26.7%) or independent Dan Woodyatt (4794, 23.3%) who survives at the last exclusion. This is down to the 2861 votes cast for other candidates, how they split between Liberal, Labor and Woodyatt, and whether Woodyatt’s share of that total is 708 votes (24.75%) more than Cook’s. Woodyatt’s observation of the count has reportedly been that a “conservative” projection would bring it down the wire, suggesting he rates his own chances as better than even.

Wednesday 4.30pm. ECSA advises it will knock over whatever postal votes arrive in the available time frame of the next two days on Friday. These votes will be small in number, but given the lateness of their arrival they will almost certainly be coming from overseas, so it should not be assumed they will follow the same pattern as today’s batch which came in on Monday and Tuesday. The provisional votes have already been counted, contrary to what I stated earlier, so Labor is not awaiting upon the small boost that such votes invariably provide it.

Wednesday 11am. I should probably know better than to find anything about this count surprising by now, but that is undeniably where I stand after today’s batch of 194 postals behaved very unlike those that preceded in breaking 113-75 to Labor and putting them 21 votes ahead. As related on Twitter by Haydon Manning, today’s primary votes are 52 each for Labor and Liberal, 47 for Woodyatt, 17 for Golding, nine for the Greens, seven for De Jonge, four for Couch and nothing for Walker. By my reckoning, Labor’s two-party lead is now 10,282 to 10,261. The numbers are yet to be updated to the ECSA site, but it’s all accommodated in the table above (UPDATE: ECSA now updated).

Part of the surprise of the Liberals’ late count recovery was a dramatic improvement in their preference share, which was 34.6% on ordinary votes, 43.9% on the first batch of postals and 44.9% on pre-polls. But on this batch of postals, it was 27.4%. On this form, you would suggest that the late-arriving postals are behaving very differently from the early ones, and that the wind should stay at Labor’s back for the rest of the count, particularly if provisionals await to be added. But by this stage, the only thing it would seem prudent to expect is the unexpected.

Tuesday 4pm. The shocks keep rolling in: pre-polls, while slightly less favourable to the Liberals than postals, have gone to them with sufficient strength (2519-2051) to give them a 17-vote lead. Since there will presumably be about 200 postals still to trickle in, the likelihood now is that the Liberals will pull the iron out of the fire, unless a distinct trend in outstanding postals together with provisionals can yet save the day for Labor. Then there’s the fact that Dan Woodyatt’s deficit compared with Labor is down to 3.4% – probably bigger than he can overcome on preferences, but not definitely.

Tuesday 11:30am. Daniel Willis of The Advertiser tweets: “Told early indications from small sample of Fisher pre-polls shows same trend as ordinary ballots cast on Sat. Labor position strengthens.”

Monday afternoon. Things have taken another turn with the first 1217 postal votes flowing very heavily to the Liberals, to the extent of offering them a glimmer of hope. The votes have split 679-504, or 57.4-42.6, and in doing so cut the lead from 626 to 451. If all of the declaration votes were to divide thus, the Liberals would end up winning by around 250 votes. However, it’s all but certain that they will do less well on the pre-polls, of which around 4650 will be counted tomorrow. Of postals, there should be perhaps around 300 more to come in through the rest of the week. Kevin Bonham has been holding on to the idea of Dan Woodyatt still taking the seat after getting ahead of Labor in late counting, but suggests this is conditional on the 5.4% gap on polling booth votes being narrowed to around 3% on late counting. So far he’s pared it back to 4.6%, and if that trend continues it would fall below 2%. In short, a lot of election night prognostications were premature, although the likelihood still remains that Labor will win the seat. Tomorrow’s counting should clarify the situation.

Sunday evening. The recheck indeed confirmed the anomaly in Aberfoyle Park, which together with other rechecking puts Labor’s lead at a formidable 7614 to 6988, or 626 votes (1.2%).

Close of the evening. Multiple reports are circulating to the effect that there is an anomaly in the published numbers detailed above, and that its correction will confirm a stunning result that tips Labor over the line to a parliamentary majority of 24 seats out of 47. The ECSA score has Labor’s lead at 7384-7115, giving them a fragile margin of 269. But it seems there is an anomaly with the Aberfoyle Park High School result, where Labor has received only 48% of minor party and independent preferences compared with a fairly consistent 65%-67% elsewhere. Two stories have emerged on Twitter as to what has gone wrong here: David Washington of InDaily relates via an unnamed source that the two-party result from a booth was entered the wrong way around, and Kaurna MP Chris Picton indicates that 200 votes have been assigned to the wrong pile. In either case, Labor’s preference share from the booth would be well in line with the overall trend. Their lead would be 599 votes (2.1%) on the former scenario, or 669 (2.3%) on the latter. Nine Network reporter Tom Richardson splits the difference by tweeting the word from a “VERY seasoned” Labor source that their lead is in fact at 638.

Whichever it might be, it would take something quite miraculous to reverse the result in late counting. The check vote will be conducted tomorrow, which will presumably get to the bottom of the Aberfoyle Park anomaly. There will follow counting of postals on Monday and pre-polls on Tuesday.

9.00pm. All primary and 2PP booth counts are in, and that’s apparently it for the evening – except there is talk that ECSA has the 2PP numbers in the wrong way around for one booth, for which the likeliest candidate is Aberfoyle Park High School. If that’s the case, Labor’s lead is actually 7549 (52.1%) to 6950 (47.9%), in which case they’re pretty much home and hosed. Will keep you posted on that one. I’ve rejigged the chart to feature the Labor-versus-Liberal 2PP result.

8.37pm. The outstanding booth on 2PP, Aberfoyle Park North, was mid-range in terms of primary vote swings, so presumably won’t make difference to the knife-edge 2PP projection.

8.34pm. All booths now in on the primary vote.

8.18pm. Two more booths reporting on two-party, and Antony’s projection now has Labor 0.8% ahead. A big week of pre-poll and postal counting awaits.

8.06pm. There’s now a sixth booth in on the Liberal-versus-Labor 2CP, and whichever one it’s been, it’s a bad result for Labor, such that Antony now has the Liberals nudging into a 0.3% lead on his projection. But it looks to me like Reynella East was a particularly good result for Labor in a very large booth, so my feeling is that that will put them back in front when it reports.

8.01pm. Reynella East has now reported, and it’s left that last entry of mine looking pretty good. Woodyatt now well behind Labor, who would probably be getting a bit excited around about now.

7.57pm. The two booths outstanding are Aberfoyle Park North, which was a good booth for Such (45.9%), and Reynella East, a bad one (30.0%). But the Woodcroft booth, which was Such’s worst (22.2%), isn’t in use this time, and the nearest booth is Reynella East, so expect that booth to hit pretty Woodyatt hard.

7.53pm. And sure enough, the Happy Valley West booth puts Labor ahead of Woodyatt on the primary vote.

7.48pm. Interestinger and interestinger. A big fly in Woodyatt’s ointment is that the outstanding booths were far Such’s weakest part of the electorate, and if that flows through to him he has little chance of finishing ahead of Labor. And according to Antony’s projection, Labor now have a 1.3% lead on the Liberal-versus-Labor two-party count with five booths counted. So it appears a gravely underestimated their chances a few posts ago.

7.38pm. Aberfoyle Park booth added. With each booth my projection continues to look better for Woodyatt, who has also inched further ahead of Labor on the raw primary vote. I’ve just fixed a bug in my primary vote percentages.

7.33pm. I should observe that my primary vote swings are based on booth matching, so this looks a rather poor result for the Liberals, who are hardly making any headway in Such’s absence, whereas Labor are up around 10%.

7.30pm. Antony projecting 2.2% Liberal lead on the Liberal-versus-Labor count, based on three booths. A Labor win is thus not impossible if they indeed finish second, but the odds appear against them on both counts.

7.13pm. Aberfoyle Park Central booth pushes projected Woodyatt lead out to 4.5%, but he’s only just clear of Labor in second place.

7.10pm. Aberfoyle Park South puts Woodyatt’s nose in front. But I must again stress that this is based on a highly speculative preference allocation.

7.09pm. According to my calculations, Clarendon implies a Liberal win probability of 68%, but in Cherry Gardens it’s 26%.

7.07pm. I’ve cleaned a bug that was causing my 2PP projection to be based entirely on the Clarendon result. Revised verdict: very interesting.

7.05pm. Raw results suggest Woodyatt should finish ahead of Labor. Using the Such-versus-Liberal preferences from March, I’ve got an 11.9% swing to Liberal compared with a 9.6% margin from Clarendon, but a 6.4% swing in Cherry Gardens.

7.00pm. Mixed signals coming through. Labor MP Michael Atkinson pessimistic, but Daniel Wills of The Advertiser apparently hearing Woodyatt looking good. Both booths in so far are very small, so the the later results might change the picture.

6.57pm. Second booth in, Cherry Gardens, and it looks like this was the one Antony was referring to. So we’ve now got two unexpectedly good results in from the Liberals, who I’m projecting to win quite comfortably.

6.53pm. The Clarendon booth primary votes are in – I’ve been a bit slow passing it on because I’ve had bugs to iron out. Antony reckons the Libs vote is up 6.8% but my output says higher, so I might have to look at this.

6.40pm. There’s quite a bit coming through on Twitter about how the count is progressing. My live coverage will be strictly concerned with published results, but you’ll find the diligent PB community relating the Twitter info in the comments thread.

6.38pm. A further point of explanation: the swing figures shown for Woodyatt will be compared with the result for Such at the election. Sorry if this seems to suggest that I’m buying into Woodyatt’s campaign pitch, but it does seem the most instructive way of going about it.

6.30pm. ECSA will be conducting a Liberal-versus-Labor two-party count. Here however you will find something different – a Liberal-versus-Woodyatt count based on the assumptions that preferences will split between the two in the same way as they did between Liberal and Bob Such in each individual polling booth at the elections.

6pm. Polls have closed in South Australia’s Fisher by-election, which you can read all about in the post below. This being a suburban seat with large booths, I’m guessing we won’t see any numbers for about an hour or so.

Author: William Bowe

William Bowe is a Perth-based election analyst and occasional teacher of political science. His blog, The Poll Bludger, has existed in one form or another since 2004, and is one of the most heavily trafficked websites on Australian politics. View all posts by William Bowe