My favourite place to analyse polls and elections is the island Australian state of Tasmania, and it has been since the run-up to their 2018 election, which I watched live (and subsequently have re-watched about 20 times in the intervening three years). I love it for the intrigue, the drama, and the electoral system - Hare-Clark, or as it's known everywhere else, multi-member STV.
Tasmania is guaranteed 5 seats to the Australian House of Representatives, a malapportionment guaranteed to them by their status as a founding state of the federation. It isn't a huge problem in the grand scheme of things - they'd be due 3 as a matter of their population, if not for that floor, although the Senate giving them equal membership as New South Wales is absurd - but it does create a fun circumstance where there are now always six politicians serving districts of the same name at one level of government or another. As a matter of course, Tasmanian elections are fought on the same boundaries as the five federal districts, with the same names and lines. The difference is each of those seats elects 5 members, for a Parliament of 25. You have to vote 1-5 on your ballot paper, meaning it is a form of semi-optional preferential voting - there are some preferences, but you do not need to number every square. Because it's a 5 member election, the minimum number of votes you need is one vote more than 1/6th of the total, although with the semi-optional nature of preferences fifth seats are often won with less than a quota because of exhausted ballots.
In 2018, the Liberals won a second straight majority government with 13 seats - a bare majority, yes, but also one that's a more impressive accomplishment than Scott Morrison's 3 seat majority in Canberra elected in 2019. The government won 13 by getting 3 seats in each of Bass, Braddon, and Lyons - the first two northern seats, and Lyons a sprawling electorate linking north and south of the state - and winning two seats in both of the southern seats, Franklin and Denison (which has now been renamed Clark). In Franklin, the Liberals came rather close to another seat, losing out narrowly to the Greens, who only got two seats in 2018, both southern seats.
In theory, the Liberal majority is rather vulnerable to a Labor or Green advance, but in reality, it isn't. Neither Green nor Labor are particularly close to winning any of those third Liberal seats in the north, and Labor are not nearly strong enough to win 3 in Clark to deprive the Liberals of a majority. On the other hand, the Liberals, riding a COVID popularity high and the fact that four state or territorial governments in Australia - and their neighbours across the ditch, New Zealand Labour - had all gotten re-elected, decided to go for an early election, thinking they could get a more reliable Parliament.
That desire for a more reliable Parliament was necessitated by the fact that the Liberals were in a Majority Government in name only for basically the entire term. Their second member in Denison/Clark, Sue Hickey, won the Speakership with Labor and Green votes after not being given a Cabinet position, and she would routinely vote against the government. That came to matter less because of a Labor member's retirement and replacement with a member who ended up sitting as an independent, and voting fairly regularly with the Government. It wasn't easy, but as far as (de facto) Hung Parliaments go, it was very easy. But, the Liberals took their shot at a more stable Parliament, and now we have the election Saturday. Given I said the majority wasn't vulnerable to Labor or the Greens, you'd be thinking they're home and dry. The problem is, I lied. There are two threats to the majority that threaten to make this election call the Australian equivalent of David Peterson going in 1990 in Ontario.
Polls
There hasn't been a single public media poll of the race, and the only data we have gotten was a uComms poll commissioned by a left-leaning pressure group, which had the Liberals at 41%, which is the point at which they're not getting a majority. Kevin Bonham, whose work I greatly respect, has the definitive write up on this poll here, which should be considered and taken seriously, even if I have absolutely no belief it is true. The poll would result in a minority government if strictly born out - the Liberals would almost assuredly lose their majority, with either Labor or the Greens winning that last seat in Lyons, and maybe even the Greens gaining the seat they used to hold in Bass off of the Liberals, not Labor. It would be a catastrophic failure for the party, and would go against every single trend of post-March 2020 elections globally. That global trend also suggests the government will beat their polls, as New Zealand Labor did, and even in a losing effort, Donald Trump did, but these samples are tiny, and should be treated with caution.
I've been made aware of private polling in the final week suggesting the Liberals are holding their 2018 vote, the Labor vote is down, the Greens are up a touch and the vote for others is up (which we will get to in a minute in more detail), which matches what Bonham has suggested he's seen, and what older polls said. If that is true, then the Liberals would gain another seat in Franklin at the expense of Labor, and Labor would be at risk of losing their second seat in Bass to the Greens. What you'll note is that I'm not mentioning Clark at all here, and that's where the second threat to the majority comes - a pair of independents who threaten to make everything haywire.
Clark
Clark is the most urban seat in the state, but it is a seat with two distinct ends - Hobart, which is where some of the best Green votes anywhere in the country are cast, and then Glenorchy, which is much less inner city. The Mayor of Glenorchy, Kristie Johnston, is running for the seat as an independent, as is Sue Hickey, former Lord Mayor of Hobart and the disruptive ex-Liberal Speaker. Both of them are credible candidates, but both could easily poll 8% and not get over the line (quota being 16.7%). Clark last time was a simple count - 2 Labor, 2 Liberals, 1 Green - and now it is a complicated mess. We can be sure the Liberals and Labor will get one member, but beyond that, we know nothing, with 5 candidates - the second Labor, second Liberal, first Green, and then Johnston and Hickey - all fighting for 3 seats. We can guess at the kinds of voters who will vote for each of Hickey and Johnston, but we have no actual idea of their support levels, or who they're pulling votes from. Is Hickey taking Liberals away from the party, socially liberal rich people who are offended by the way the government treated her, or is she now appealing to Labor and Greens supporters? Even beyond the question of whether she can win, that question changes whether the Liberals get a second seat in Clark, which matters if they don't get 3 in Franklin.
Prediction
So, the lack of public, media commissioned polls makes me very nervous making any form of prediction, but I'm not that arsed about that, and I have a prediction anyways, so let's make it public. I think the Liberals get a majority government by holding their 3 seats in Bass, Braddon, and Lyons, and gaining a 3rd in Franklin from the Labor Party. I have the Greens gaining a seat in Bass off Labor and holding their seat in Franklin. And I have no idea about Clark, beyond what I've said that Labor and Liberal will get a single seat at minimum. My gut - with absolutely no backing for this - says that Clark elects 2 Labor, 1 Liberal, 1 Green, and Johnston, with Hickey losing.
Final* Predictions By Electorate
Bass: 3 Liberals, 1 Labor, 1 Green (Green Gain from Labor)
Braddon: 3 Liberals, 2 Labor (No Change)
Clark: 1 Liberal, 2 Labor, 1 Green, 1 Independent (Independent Gain From Liberal)
Franklin: 3 Liberals, 1 Labor, 1 Green (Liberal Gain from Labor)
Lyons: 3 Liberals, 2 Labor (No Change)
*If there's a poll I reserve the right to revise all of this
Final Prediction
13 Liberals, 8 Labor, 3 Greens, 1 Independent (Johnston), Liberal Majority Government.