2021 Tasmanian Postcount: Clark

(Link to main 2021 page including link to other postcount threads)

CLARK (2018 Result 2 Liberal 2 Labor 1 Green, when election called 1 Liberal 1 Labor 1 Green 2 IND, when election held 2 Liberal 1 Labor 1 Green 1 IND)

SEATS WON 1 Liberal 1 Labor 1 Green 1 IND.

CALLED WINNERS: Elise Archer (Lib), Cassy O'Connor (Green), Ella Haddad (Labor).

PARTY CONTEST: Liberal vs IND, Liberal well ahead, INDs will gain, both INDs need to overtake Liberals for them to lose.  Liberals strongly expected to win.

WITHIN-PARTY CONTEST: Behrakis (LIB) vs Ogilvie (LIB) for expected

 Liberal 2nd seat. Ogilvie will win.  

IND vs IND CONTEST: Johnston vs Hickey, Johnston ahead and expected to win.

UPDATE: Clark called as wins for Johnston and Ogilvie after exclusion of Davis.  

Introduction To Clark

Warning: The Clark postcount is very complex!  This article is rated Wonk Factor 4/5.

I am starting with Clark in my postcount threads because Clark is by far the messiest and is also the one that decides whether the Liberals win a majority or not, and hence the fate of some chap from up north as the Premier.  Clark has delivered the weirdness that was expected of it.  Before the election it looked like it could be a five-way contest for three seats.  The Greens have romped in with a terrific vote, given the competition from indies, and Labor has crashed and burned with a dire result (barely beating the Greens).  That answers two of the questions, but there is still a very complex contest between the Liberals and two independents for the final two seats.  We don't even know yet which Liberal it is!  This contest might become clearer in primary counting over the next day or two, or it (and the Liberals' majority status) could stay unclear for two weeks and gradually unravel in the distribution of preferences.  (How exciting that would be!) 

Anyway, history has been made.  For the first time since the House was reduced to 25 seats an indie has been elected at the ballot box.  We just don't know who yet (and there's an outside chance it could be two).  

With 70.2% counted, there is one booth remaining that was expected for tonight and that is the Glenorchy prepoll.  The Liberal ticket has 1.86 quotas, Labor 1.30 quotas, the Greens 1.26, Kristie Johnston 0.67 quotas and Sue Hickey 0.58 quotas. The rest (0.33 quotas) is scattered between ungrouped independents, Shooters, Animal Justice and Federation Party.  Johnston may increase her lead over Hickey in the Glenorchy prepoll booth.  All these numbers will change, but let's assume for now that they don't and see what might happen.

Presently, the Liberals are in fourth place.  Nobody has quota (the fifth time this has happened in the 25 seat system in some seat or other) so the race will proceed by cutting from the bottom up.  Minor Greens, Liberal and Labor candidates will elect Cassy O'Connor, Elise Archer and Ella Haddad.  Neither Labor nor the Greens will catch Hickey, so at some point we will be left with three candidates, Johnston, a Liberal and Hickey, unless one of these has crossed the line earlier.  The candidate in last place loses and the other two win.  

Currently Johnston needs to gain .19 quotas to catch the Liberals and Hickey needs .30 quotas.  This is out of .89 quotas in preferences, but the problem for them is votes getting to the end of the Labor or Greens ticket and exhausting because voters only voted 1-5 and stopped.  Plus, no party's supporters ever vote all the same way: some Greens and Labor voters will preference the Liberals, though given that there are two independents, maybe not many.

Where the independents benefit is leakage.  Because they are single candidates they cannot lose votes to leakage; all they do is improve their position through the cutup.  The Liberals, however, have 0.63 quotas to be thrown by candidates who will be excluded, even assuming Archer hits quota more or less exactly.  A 10% leakage rate would knock about .06 quotas off their position, with some of that going to their independent rivals.  So their real lead over Johnston might be, say, 0.11 quotas, and over Hickey, say, 0.20 quotas.  The question then is how many of the Labor and Green preferences flow and what sort of split there is in them between the surviving Liberal and the two independents.

Some insight into the problem of exhaust comes from the 2010 cutup in which Elise Archer narrowly defeated Andrew Wilkie.  On a surplus consisting of Lisa Singh (Labor) votes that had elected Scott Bacon, Archer (Lib) got 18.9%, Wilkie (IND) 18.0%, Burnet (Green) 17.0% and 46.1% exhausted.  The weak flow to Wilkie compared to Archer was probably affected by gender voting.  In any case, we can estimate that close to half the Labor vote exhausts.

Burnet (Green) was excluded and her preferences went Wilkie (IND) 43.4 Archer (Lib) 21.4 Exhaust 35.1%, so around a third of the Green vote exhausted. (This is also similar to what happens when Green preferences are thrown between Labor and Liberal candidates.)  Again I suspect there was a gender component in this flow.  

If these are regarded as reasonable benchmarks, around 0.23 quotas of the 0.89 currently available would exhaust just from the Labor and Green tickets, some small amount would also exhaust among the remaining candidates.  That leaves, maybe, 0.6 quotas to distribute.  If the Liberals get 20% of these and the rest split between Hickey and Johnston, then the Liberals gain 0.12 and the indies split the rest evenly, that's .24 apiece and Hickey falls .06 quotas short (700 or so votes).  

Maybe a 20% split estimate for the Liberals is miserly here because there are a few minor right wing candidates in the mix such as Shooters whose preferences will favour them, and we saw in the Wilkie example that the Labor vote didn't help an indie there against a Green at all, while the Green split wasn't all that strong.  But on the other hand, leakage when either Simon Behrakis or Madeleine Ogilvie is excluded could be higher than usual, because Ogilvie is a recent recruit with a Labor past, and because some Ogilvie voters may be voting for her personally and not the party.  So that could hurt the Liberals.  (I think leakage if Ogilvie is excluded is the bigger problem.)

Which of Behrakis and Ogilvie ends up as the candidate could also influence the preference flow.  Behrakis is hard-right and may get worse flows from Labor and Green voters than Ogilvie, but tempering that the major Labor and Green preference sources will be male, which may help him.  It's not clear which of Behrakis and Ogilvie will be the Liberal contender, because Behrakis may do better on the preferences of his male and more traditionally Liberal ticketmates (especially Coats) but Ogilvie is likely to outperform him on cross-ticket leakage and the preferences of ungrouped candidates because of her higher profile.  Presently Ogilvie leads Behrakis by 127 votes.

Another question is which independent wins assuming that only one of them does.  At present Johnston leads Hickey by 694 votes.  This isn't unbridgeable, but my suspicion is that Hickey will do better on Greens votes, Johnston on Labor votes, and it is challenging for Hickey to close the gap.

What we should keep an eye on is the raw quota total as more votes come in.  If the Liberal vote continues to build to over 1.9 quotas, it is very hard to see their final candidate getting caught by both independents.  But if this doesn't occur, we still have a live contest that looks close, and so the possible 1-1-1-2 result (with all the chaos it would unleash) is still alive.   I still tend to think getting both indies over the Liberals is tough, even without primary vote changes.  

A very long way still to go on this one.

Updates

Sunday 5 am: The big Glenorchy prepoll came through just before 2 am and improved the Liberals to 1.88 quotas, Labor 1.32, Greens 1.21, Johnston 0.68. Hickey 0.59 rest 0.32 with 76% counted.  These changes were good news for the Liberals as they very much want to see the Greens excess decline, since Greens voters are more likely to vote all the way through than Labor voters.  Also the Liberals gained 0.02 Q and the preference pool dropped a bit to 0.85.

Sunday 9 am: I don't know the timing details but today's counting is scheduled to include mobile votes which favour the Liberals, and also the Granton (Clark) and Kingston (Clark) prepolls.   In 2018 mobile votes were around 1000 votes (not sure if COVID will affect this); I don't think the two electorate-fringe prepolls will amount to much.

Sunday 5 pm: TEC site hasn't updated yet but Poll Bludger has. Current quotas: Liberal 1.91 Labor 1.32 Greens 1.17 Johnston 0.68 Hickey 0.61 others 0.30.  The preference pool continues dropping, now down to 0.79 of a quota.  That is most of the prepolls gone.  In 2018 out of division prepolls favoured the Liberals very marginally and postals significantly, but late postals tend to be weaker than early postals so we may not see much movement beyond this.  Last night I was intending to call it if the Liberals got to 1.9 but I now think they need 2.0 for a call to be safe.  

I have also had some scrutineering figures from an excellent source that are very interesting.  For Ogilvie there is a 12.6% leak on a sample of 230, similar to Archer, but the issue here is that the leaks across the board from party tickets appear to have been increased by the indies.  (This is not good for the Liberals, though perhaps if Behrakis is excluded his voters could be more loyal.)  Also, the Liberal share on a three-way basis with the independents of leaked preferences from the Greens and Labor, albeit mostly from the Green and Labor ticket leaders who will not get excluded, is running at 13% and 21% respectively.  The source has run a detailed projection which also finds Hickey trailing, but by an even closer margin than mine from last night, just a few hundred votes.  

Sunday 7:30 pm: I have run another simulation off the current primaries in which I try to predict the eventual destination of each party's spare votes (I have not tried to do a full simulation of excluding individual candidates).    In my simulation the generic Liberal candidate was overtaken by Johnston but defeated Hickey by c. 800 votes.  This is close enough that I cannot be sure the Liberals will win, but if they continue advancing in the post-count they may reach a point where I can call it or nearly so.  Note I am continually seeing nothing that points to Hickey catching Johnston so unless something of that sort appears then Hickey's only path is beating the second Liberal.  

Monday: No new numbers til Thursday apparently but it is worth noting that Brad Stansfield who has worked on the Liberal campaign has said he thinks it's "certain" the Liberals will win the seat and that he believes it will be Ogilvie.  

Thursday: A substantial load of votes added. Liberals still 1.91 quotas, Labor still 1.32, Greens up 0.02 to 1.19, Johnston down 0.02 to 0.66, Hickey down 0.02 to 0.59, minor gains for AJP and Shooters. About 5000 votes added. These changes improve the Liberals' position; it would flow through as about 200 votes in my projection (taking the projected margin to about 1000) but also increases the chance that the final Liberal stays ahead of both indies (which protects them from an adverse Hickey-Johnston split).  I have had some reports of Ogilvie preferences spraying so if Behrakis is the final Liberal it could be closer.  

Friday: A very minor update with rechecking and a 225-vote increase in the total.  The Liberals are up 0.04% and Johnston is down 0.05% with Labor and the Greens up very slightly too; nobody's quota total is affected at even the second decimal place.  The gap from Johnston to Hickey has come down to 732 votes.  

Saturday:  For those with an eye on the tail end of the field, a colleague has informed me that Justin Stringer (Federation Party) has dropped from 200 votes to 150!  This turns out to be mostly a correction of an obvious typo in the Chigwell booth, where he was previously credited with an unlikely 29 votes (actually 2).  

Tuesday 11th: What to watch for: I've been writing guides on the cutup for some of the other seats but in Clark every single exclusion will be of some interest.  Those that will have the least bearing will be the minor Green and Labor exclusions until such time as the ticket leaders cross the line.  I think it will take until Bayley's exclusion to put O'Connor over meaning that a surplus of Bayley - O'Connor votes will be the biggest bundle leaving the Greens ticket at the end.  Haddad might cross on the third Labor exclusion (currently Clark).  Early on the preferences of the ungrouped candidates (especially Ewin) and Animal Justice will be interesting and I'll also be watching to see if the Liberals pull anything over the indies when the second Shooter is excluded.  In general things to watch here are: what rate of preferences are the Liberals getting, is Hickey catching Johnston, and is Ogilvie moving away from Behrakis?  Later on we get two minor Liberal exclusions that will have a massive bearing on the Ogilvie/Behrakis contest.  The final act in the cutup will probably be the exclusion of either Ogilvie or Behrakis and the distribution of their preferences.  At this stage we will have an idea what sort of leakage and to whom might be needed if the independents are to both win, or by that stage it could well be completely clear that that will not happen.  

Tuesday 11:30 No new numbers have been posted yet.

Tuesday 12:05 Final primaries are posted.  Johnston leads Hickey by 733 votes.  The Liberals have a notional lead over Hickey of 3406 votes, which will drop as preferences are distributed, especially when Liberal candidates are excluded.  The preference pool of votes from other parties is 8950 votes.  Some of this will exhaust.  

Tuesday 1:00 Stringer (right-wing) and Gershwin excluded.  Gains Liberals 47 Johnston 51 Hickey 27.  

Tuesday 1:21 Leaks from first Shooters exclusion (Bennett) are Liberals 27 Hickey 4 Johnston 3 with a curiously large leak to Sam Mitchell (43) so presumably there is some mystery connection between Bennett and Mitchell.  Now excluding Volf (Greens).  Current Ogilvie to Behrakis gap is 274.

Tuesday 1:48 Leaks from Volf exclusion: Johnston 14 Hickey 4 Liberals 4 - which all went to Ogilvie.  Johnston lead over Hickey now 775. Now we have a fairly important exclusion: Ewin.  That said a fair few of Ewin's votes may flow to Dutta (also ungrouped and fellow Hobart Councillor).  

Tuesday 2:18 Ewin exclusion Johnston 82 Hickey 34 Liberals 11 (out of 597!).  A lot of Ewin's votes went to the Greens or Labor and some of these will be back in action later on.  

Tuesday 2:56 There's a delay at the moment because the Smith exclusion appears to have created a tie between Large and Dutta and this is being checked. If necessary countback will be applied, excluding Dutta. (Confirmed: Dutta excluded.)

Hickey now 817 behind Johnston and continuing to miss opportunities to gain. I'll check how she's going against Liberals overall soon.

3:42 Johnston has increased her lead by another 21 off Dutta. 

4:10 A bit of a stocktake during the quite important exclusion of Large (Shooters).  Since the start of the day Johnston's lead over Hickey has expanded to 848 votes.  The Liberals' lead over Hickey is now 3377 votes, but the votes coming into the ticket today are more likely to leak out of it later.  At this stage, Hickey's rate of gain has been very slow.  The preference pool is now 8232 votes, because a lot of votes from the minor candidates so far are going to the Greens and Labor.  The Liberals have overall gained 25.2% of votes that have gone to them, Hickey or Johnston.   The best news for them is that the split of preferences that are going to indies is favouring Johnston.  On the Liberal side, Ogilvie leads Behrakis by 281. 

Note that Antony Green is doing graphic updates.  Some of his comments have been misinterpreted as predicting a 2 IND result as likely.   

4:50 Large (SFF) excluded, the only remaining right-wing preference source to leave a party (though there will also be Liberal leakage later on).  Liberals gained 350 (39.2%) with Johnston gaining 100 and Hickey 128.  Labor gained 177, the rest to Greens, AJP and exhaust.  

5:57 Westcott (AJP) excluded. On this exclusion the Greens gained 388, Labor 213, Hickey 180, Liberals 172 and Johnston 162.  It's interesting that the Liberals did about as well here as a group as the indies, but those votes that went to the non-final Liberals have another chance to leak later.  I had thought O'Connor wouldn't cross quota until Bayley's exclusion but constant pickup from minor candidates means she will cross off Taylor's.  All the remaining exclusions now are Labor (4), Greens (2) and Liberal (3) plus small surpluses as one from each of these crosses the line.  

6:31 Carnes excluded with about a 10% leak which gave a few dozen each to the Liberals, Hickey and Johnston.  On the Liberal side, Ogilvie's lead is now 317, so she has been gradually pulling away from Behrakis as expected, but it will be very interesting to see what happens when the minor Liberals are excluded as Behrakis may make gains there.  Johnston leads Hickey by 797 and the Liberals notionally lead Hickey by 3547.  The preference pool is now 7022.

7:43 I've had another go at the same projection I did before now that there are fewer moving parts and it still has the Liberals winning by 1000 or so.  That is assuming 20% leakage on the final Liberal exclusion.  If it is Behrakis it could well be less than that.  The only danger I see is if the final exclusion is Ogilvie and her votes are ridiculously leaky (like, say, 35% leak.)  

Wednesday: Today's first exclusion is Bec Taylor (Greens).  This will elect Cassy O'Connor with a surplus in the hundreds; most of these two throws combined will end up with Vica Bayley.  Then will come Harvey Lennon - that will be an interesting one to see what the flow to Behrakis cf Ogilvie is like, as Behrakis' only hope is a strong flow off either Lennon or Will Coats.

10:25 O'Connor elected and surplus distributed; Johnston leads Hickey by 818 votes.

10:55 Scrutineer report that Behrakis has gained only 50 over Ogilvie on the Lennon vote.  That is not enough and in all probability Madeleine Ogilvie will now be the final Liberal which also means she is very likely to win.

11:08 Leakage off Lennon was very low, only around 7.5% with Hickey getting 42 and Johnston 26.  Mitchell is being excluded now which won't be very informative.  However if Chris Clark gets enough to overtake Coats then Coats will be next and after that it could well be callable.  

11:40 Leaks off Mitchell to Hickey 52 Johnston 45 and very little for Liberals.  Coats is out now.  

12:30 I hear that Behrakis is not gaining off the size of the piles on Coats so I expect Ogilvie to beat Behrakis.  There might be enough of a flow from Coats to Archer to elect Archer on this count but I am not sure about that.  

12:48 Ogilvie has moved further ahead of Behrakis on Coats by 134 votes.  Hickey is continuing to outperform Johnston on Liberal leakage but not by enough to be an issue.  

1:06 Archer is elected with a small surplus that consists of several hundred papers.  After that will be Clark, probably followed by a Haddad surplus, and then the serious exclusions of Bayley and Davis. 

1:30 Seems pretty much callable that Ogilvie will win.  

2:00 Clark exclusion now.  This will elect Haddad with a surplus then the Bayley exclusion.   

2:44 Clark exclusion done and enough flowed to Davis that Haddad didn't get elected yet!  So now one of the big ones, the throw of Vica Bayley (Green) between two Liberals, two Labor, two independents and exhaust.  Totals before Bayley (2761) are Behrakis 4896 Ogilvie 5316 Davis 4006 Haddad 10434 Johnston 7842 Hickey 7072.  This might put Haddad over, but I'm not sure about that.  

3:54 The first 2367 votes from Bayley split Behrakis 53 Ogilvie 46 Davis 146 Haddad 634 Johnston 643 Hickey 420.  425 (18%) exhausted immediately, a rather low exhaust rate.  This puts Haddad over by 442 which will have to be thrown again, as will the 146 to Davis.  But of those that did not land with Labor, the Liberal share was a meagre 8.5% which is also a lot lower than my model.  So on we go, but Hickey needs all these four things to happen:

* Higher than expected Behrakis leakage, say 15%

* Lower than expected exhaust rate from Davis

* Davis votes to split strongly in her favour cf Johnston

* Liberal share of Labor preferences to be lower than normal in this circumstance and lower than scrutineering sampling

Of these the third is the hardest because we have not seen this happen in any significant distribution so far.  It would be a miracle to hit all these targets at once.  

4:48 Haddad surplus and Bailey minor votes thrown and the Liberals managed a miserable 3% of these with 18% exhausting.  But most of the votes went to Davis with not that many to the indies and the pattern continuing of Johnston getting more than Hickey.   So now the preferences are all just Davis and then we get the Behrakis throw to see how much leaks or doesn't leak.   

6:21 Davis done and now Johnston 9744 Hickey 8477 Ogilvie 5591 Behrakis 5249.  Even assuming a 15% leak with 8% to Hickey that's still a win by over 1000.  Even with a 20% leak with 12% to Hickey that's still a win to Ogilvie by several hundred.  Called (both positions).

7:44 All over and Ogilvie wins by 1429 over Hickey and 175 over Johnston.  Johnston wins by 1254 over Hickey and is the first independent winner at the ballot box in the 25 seat system.  Clark elects five women at a general election for the first time, but it previously had five serving MPs since Ogilvie's win on a recount in 2019.  Ogilvie wins another close one, beating Behrakis by 342 in the within-Liberal contest to go with wins over Cox by 201 in the 2019 Labor recount and Amos by 331 in the 2014 recount.  (She also lost a within-party contest to Haddad by 731 in 2018).  

3020 votes exhausted, about two-thirds of them at Davis' exclusion and most of the rest off Bayley.  Had these voters numbered all the boxes and had their preferences flowed similarly to other voters for their parties, Ogilvie would have still just won, but by a few hundred.   The leakage rate off Behrakis (13.2%) was low for this end of the count, when a candidate's vote always includes a fair amount of cross-party spray.