Note
Since people seem to think I have some level of special insight into election numbers (judging by the number of questions I got asked on the subject last night), my take on the situation:
- Labor is pretty much guaranteed 70 seats.
- Labor is highly likely to win Lindsay and Corangamite, although the ABC computer is saying both are very close - they are assuming the postal votes will behave as they did last time, but both seats were Liberal seats going into 2007 and Labor going into 2010, which means Labor should do better on postals this time than they did last time. (Incumbents normally have a bit of an advantage on postal votes because they have the benefit of an office to send the applications out; otherwise postals and prepolls tend to be a bit more conservative than the electorate as a whole, because a higher proportion of them are elderly).
- to get to 73 (which is what I think Labor will need to form a government), Labor will need to win at least one of Boothby, Hasluck or Brisbane (in decreasing order of probability). I think Boothby is the best chance of these. Labor are currently ~700 votes ahead there, though last time the Liberals won the postals by more than that. (The ABC computer factors this in, which is why they are currently showing the Liberals ahead).
- the fate of Denison is anyone's guess and anyone who tells you otherwise is either being premature or don't know what they're talking about. At the moment the Liberals, Andrew Wilkie and the Greens are within about 3% of each other, but the Greens will probably bridge at least some of that through the preferences of a socialist who got a couple of percent, and postals/prepolls (which the Greens normally do very well on in Tasmania, although as this is colloqially referred to as the 'bushwalker vote' there may not be as many of them in August as there are in November). If Wilkie doesn't come fourth he will probably be elected on some combination of Greens/Liberal preferences, if he does come fourth I think it will be Labor versus Greens (a wild card is that I suspect a reasonable number of Liberal supporters voted tactically for Wilkie so there may be more Wilkie 1 Liberal 2 votes than one might otherwise expect). None of this makes a huge difference to the big picture as I am fairly certain that either Wilkie or a Green would support Labor on confidence and supply.
- the final Senate seat in Victoria is also interesting - it's currently being called for the DLP but could go to any of the DLP, Family First (yes, this means Steve Fielding getting back), Labor or Liberal. The DLP and Family First gave preferences to each other and are only slightly apart at a key point in the count; whichever one stays in the count at that point will probably end up getting elected on major-party preferences (the Greens were only just over a quota so their second candidate, who would otherwise have got Labor preferences, will be out of the count by then).
- while on the subject of Senate votes, below-the-line votes are not counted on the night. From my experience about 20% of Greens voters vote below the line but only about 5% of others do, so the final Senate figures will be stronger for the Greens than their current ones are. This won't affect their own result (except for the outside chance of getting a second person up in Tasmania) as they've clearly won one seat in each state, but could affect preference flows elsewhere.