Nicholson in the Oz

welcome to               Mumble   =

Aug 26 Kwernsleend

Another day another bad poll for the Queensland Coalition. In the Courier Mail, primary support 47 to 40; no two party preferred in online version. 

Centrebet odds stuck on 9 to 1 for a Coalition win. My 7 to 1 bet of last week obviously a poor buy. But I still suspect it'll be closer than most seem to think.

Aug 25 A Queensland graph: Newspoll over the Goss government

Graph below shows Newspoll primary support from before the election of the Goss Labor government in 1989 until its virtual end in 1995. The beginning, which broke 30-odd years of National Party rule, was expected at the time. But not so the end; note the final survey, on the Wednesday-Thursday before the poll, showed a little movement but missed the big swing on election Saturday.

As you know, Goss hung on until a byelection finished him off early the next year.

'Genius' watch

If you google "John Howard" and "genius" you get all sorts of strange couplings. Yesterday, Paul Kelly, in an otherwise typically muscular piece, couldn't help but add to the collection. Why must people use the 'g' word?

• Spam causes me to change my email address, at left. 

Aug 23 Queensland: terms or years?

In yesterday's Crikey email, Charles Richardson wrote this:

Of this year's four state elections, Queensland is by far the most important psephologically. The other three – Tasmania and South Australia last March, Victoria still to come in November – confirm what we already know, that first- and second-term Labor governments in this cycle are unbeatable. But Queensland is uncharted territory; none of the other state governments has reached the end of its third term.

That's kind of how I tend to see things. The Beattie, Bracks and Rann (and probably Bacon/Lennon) state governments have achieved re-elections of record proportions. (These are fair dinkum 'records', not the pretend ones we hear about John Howard.) Carr had two huge wins and Gallop one ordinary one. (Both territory Labor governments also hold record majorities.)

Visit each of the six jurisdictions and you'll hear six variations of the same narrative: the Premier is a political wonder in touch with his electorate; and the opposition is a joke, out of touch. But it's more useful to recognise a pattern.

However, within this there might be two ways to see the Beattie government. It is, as Charles says, attempting to win a fourth term. But it is also an eight year old state government facing re-election - like Bob Carr's was in 2003. Perhaps time in government is more important than elections faced?

(We can also note that the Bracks government, although only coming to the end of its second term this year, is just a year younger than Beattie's.)

The September 9 results might fill the gaps. 

Aug 22 Qld: Newspoll says 54 to 46

From primary support of 45 to 38.

Aug 21A referendum on four year terms?

Time flies, and it's referendum consideration time again. Has it really been a year since last time? Not quite. [Update: cabinet rejected it the following day.]

Queensland by-election losses

Antony Green sees the Nats increasing their Gaven majority on September 9. Gaven was one of the three seats Labor lost in by-elections during the current term. Here's the pollbludger on Gaven. And an earlier post. Malcolm Mackerras doesn't expect any of the three to return to Labor.

I must be missing something. Surely all three will swing to the ALP (from those by-election results) and I reckon at least two (don't know which) will be regained, even if Labor loses the election.

Or what am I missing?

Aug 20 Four TNS Qld polls

In Sunday Mail, sample size 200 in each seat = 7% error margin. Take with care (or ignore altogether).

Aug 19 Galaxy: 53 to 47 in Courier Mail

Exhausting the votes

The other day I wrote that 'all pollsters' ignore the 'optional' in Queensland (and NSW's) OPV. Not true of ACNielsen, at least, whose preference question (asked of the minority not nominating a major party for their primary vote) differs slightly:

So Nielsen gives respondents the implicit option of 'exhausting', but as I don't think they're polling Qld this election (have they ever?), it's academic. Not sure about the other preference-getter, Morgan, but their headline results are notionally allocated these days anyway.

Aug 18 Qld: Courier Mail's colourful pendulum

Aug 17 Always read Malcolm ...

.. who was in yesterday's Australian, on Queensland. Here's his pendulum in a new window [100kb]; you might like to toggle.

Malcolm sees Labor gaining no seats, but I reckon a by-election loss or two (Chatsworth, Redcliffe or Gaven, all of which he puts on the Coalition side) is likely to return to the fold, even if Labor loses government.

Another Queensland graph: Morgan

Above is Morgan from the 2004 election to this week. They've been taking them regularly, so there are more of data points than Newspoll (below). But note that every one of Morgan's surveys has Labor further ahead than at the last election, which is ... counterintuitive.

Some people reckon Morgan's survey strategy, of visiting houses rather than phoning, under-samples country folk. I mentioned this in the AFR a few years ago, and it would obviously be relevant in the easily most rural state.

(Since the 2004 election, Morgan has continued asking respondents who they'll be preferencing, and publishing those results on their site, but headline numbers are like Newspoll' federal ones, distributed in bulk as per the last election. Same OPV preference misgivings as Newspoll, below.)

Aug 16 A Queensland graph: Newspoll

This (above) is Newspoll two party preferred results since just before the last Queensland election. (Labor red, Coalition blue.) The first one after the election was taken in 2004; the rest in 2005 and 2006. As graph shows, four in a row were 50 50, while most recent was 52 to 48.

Note: Newspoll's usual preference misbehaviour is exacerbated by the fact that, in Optional Preferential Voting (OPV) context they, like all pollsters, assume no votes 'exhaust'. (Or they assume each side's 'exhaust' at the same rate.)

Aug 15 part III Morgan says 56.5 to 43.5 in Qld

Which is, of course, not close at all.

Aug 15 part II Queensland off to the polls

In less than four weeks. This is the first Australian contest in years which could (in my book) go either way. Most pundits are favouring Labor, and so am I - but only just. Queensland is volatile - remember Wayne Goss in 1995, clinging on for a few months with about 48% two party preferred - and  naturally conservative: it and Tasmania are the only states with more seats outside the capital than in. Throw in an eight year old government, a few big by-election swings over the last year or so, and opinion polls showing it close.

Beattie got about 56 percent of the 2pp in 2004, and I reckon we could see a swing of between 4 and 8 percent to the Coalition on September 9, leaving us between 48 to 52 and 52 to 48. But inside that range, either side could easily win with a vote minority, ie either party getting get 52% but not government. Yes, I know the current pendulum indicates that the Coalition needs over 51 percent to win, but we shouldn't take it literally. (Ie don't you worry about that.)

Anyway, I just found amazing 7 to one odds on a Coalition victory at a certain online agency, which I snapped up. [update: a bad buy, as on election day they were 10 to 1]

You must, of course, visit Antony at the ABC and William.

Aug 15 Newspoll says 51 to 49

Very different to yesterday's Nielsen, obviously. Perhaps we should average 'em.

Aug 14 Nielsen: 53 to 47

In The Age.

Rooster Season

Lateline on Friday night included Wayne Swan defending his leader over last week's doorstop stoush with Wilson Tuckey (transcript, video, summarised at SMH). The quintessential Rooster and perpetual one-man election campaign, antenna up, damage-control on - " learn your lines mate, and recite 'em". But his furrowed demeanour actually reinforced the impression Beazley had done something dreadful.

Despite the commentators tut-tutting about behaviour unbecoming of a potential PM, I reckon the trivial incident if anything added a small but favourable dimension to Kim's public persona.

More on Roosters and Lemmings

Witness for the Audit

Found myself as a witness in a Standing Jommittee the other day at Parliament House, roped along by thesis supervisor to present part of the Democratic Audit's submission. Well, listened while supervisor did most of the presenting. Being a Friday, only half the committee was there, Masters of the Universe George Brandis and Brett Mason nattered with each other distractedly at times, and George made a gratuitous joke about Malcolm Mackerras ( I'm pleased to say I didn't laugh), but everyone was pleasant, especially the Chair, that nice Mr Lindsay from Townsville.

Will link transcript when it appears if it doesn't reveal me as rambling and incoherent. [Update: it does, but here it is anyway.]

Aug 8 Newspoll says 51 to 49

No link. From primary support of 41 to 42.

Update, having seen the hard copy Coalition is down one, Labor up one and Greens up two. Newspoll's wacky preference allocation strategy ignores changes in Green support (despite Dennis Shanahan implying otherwise). Anyway, it's an ok boost to Labor, if you're into the fortnight by fortnight form-guide.

But this raises another issue. As you know, Dennis's interpretation of Newspoll results ripples through the day's commentariat. The headline and intro directly influence Fran Kelly's questions to Michelle Grattan on ABC radio at 7:45am, and all the news bulletins. Mr Shanahan reckons this is a very good poll for the Coalition (despite it pointing to a probable election loss) and so, therefore, does pretty well everyone else - including, on available evidence, most federal MPS of all parties. Writes Dennis: "Labor had been expected to get a bigger boost from the interest rate rise". By who? Would we really expect a .25 percent interest rate rise to immediately flow into an opinion poll?

Here's a better interpretation: yet another poll shows Labor ahead. On polls in general, re-reading this will be good for you.

Aug 4 Do not vote on an empty stomach

You have been warned. (From Q&A of AEC's Canadian counterpart, Elections Canada.)
[Update: Graeme from Qld notes this Canadian organisation, apparently a long-existing lobby-group which regularly changes hobby horses. See their name on the web address.]

Election 2007: Queensland cart before horse

Assuming no Howard ultra-trickiness, and barring a Lemming eruption, we appear headed for a Howard-Beazley contest late next year. Malcolm Mackerras in the Oz is picking a close Coalition victory - 'Howard's poorest win', which means a smaller than 13 seat (1998) majority. This is strategically clever of Malcolm, because it is unlikely to be way out of the ball-park.

As you know, I put Labor's chances of winning at better than even, but they need at least about 14 seats, and where will they come from? Well, for one thing a Labor victory would positively have to include a much better showing in Qld, at least a four or five percent two party preferred percent swing there. Being off a low-low base, it's not as silly as it might first appear.

Currently the ALP notionally holds just 6 from 29 seats, from 43 percent of the 2pp. They'll want to get another 6 electorates, just for starters.  A total of twelve would still be substantially less than half, and while the Qld portion of the notional pendulum (see pollbludger) suggests it would take an eight percent swing, I reckon they could do it - plus a bit more - and still sit on about 48 percent of the vote after preferences. (Labor gets a federal two party preferred majority in Queensland about once every thirty years.)

Other hunches: notwithstanding the return of their local boy to leadership, WA will prove reluctant, and so will SA, but NSW and Victoria to swing to Labor. So the biggest relative two party preferred pro-Labor swings might be in this order: Queensland, NSW and Victoria.

I also reckon (in NSW vernacular) Labor is more likely to get an Eden-Monaro (3.3) than a Lindsay (2.9); that is, the 'regions' rather than 'outer suburbs'. In fact, given likely continuing demographic changes, I even favour Page ( 5.5) and Cowper (6.6) ahead of Lindsay.

All for now.

Aug 1 Howard to stay

Yes, I've done my Brendan Nelson dough, and my earlier 67% chances of Howard retiring (recently adjusted to 50-50) have come to nothing.

Needless to say, Peter's odds of ever becoming PM now lurk in probability's bowels. And, yes, the possibility of the PM losing his seat next year remains (though it's less likely than an outright Labor win ).

Newspoll: 50 - 50

In the Oz. Which is possibly a bit generous to Labor, given the 5% Green vote, down two from the 2004 election.

July 28 Brian Loughnane's AEC whinge

Covered by Michael Costello in the Oz, mentioned by Malcolm Mackerras at Crikey (and Charles Richardson in Crikey email) and reported by Steve Lewis earlier this week.

Brian's conspiracy theory - that the redistribution commissioners deliberately constructed a pro-Labor 'bias' - is a silly try-on. However, the on-paper conditions do favour Labor compared with before the boundary redrawing (though not necessarily to the extent indicated by just looking at the Coalition's most marginal seats - see 'final point, at bottom of post).

Others have noted that the ALP still needs over 50% of the vote to win, but that's only in comparison with the last election, and every election is different, with different swings amongst different groups. If, for example, Australia only had elections every six years instead of three, and so the last voting data we had was from 2001, we would be saying that Labor only needs 49.6 two party preferred next year to form government.

This is because while 2004 behaved perfectly in terms of the Mackerras pendulum - a 1.7 percent swing picking up the precise number of predicted seats - it expanded the on paper margin by closer to 2.7. [Another way of putting it: had these boundary changes been put in place before the 2004 election, observers would have put Labor's required two party vote at 49.6, but after the election would have realised it was more like 50.6]

But if we used the 1998 election result as our 'base', we'd be back where we are today, saying Labor needs about 50 and a half % to win.

Notwithstanding Malcolm's pendulum, the 'uniform required swing' shouldn't be taken literally.

You might recall Table 2 on this page, constructed before the 2004 election, which in the right hand column has the two party preferred point which would have seen a different overall result. (The 2004 entry would be 51.7 for Labor and 48.3 to Coalition.)

This is more realistic, talking about how 'close' the election was to a different outcome, but of course then it's just academic. Not as exciting as anticipating the next result.

Final point: to be rigorous about all this we would look at changes to every vaguely winnable seat on both sides, rather than just the government-held ones on the inside of Malcolm's 'point of pendulum'. They won't all fall first.

If you know what I mean.

• And here's a new Nicholson cartoon

July 27 Malcolm's pendulum

I'm behind the times, but Malcolm Mackerras's pendulum is up on his website, although not yet a reader-friendly horseshoe. He assumes no changes to AEC proposed boundaries. Malcolm is always required reading, even in third person as here.

When all is done and the federal parliamentary library has made its final 2pp calculations, I shall filch numbers and begin constructing pendulum(s), similar to last time.

See also MM at Crikey, in which he lists, in ascending likelihood of losing respective upcoming elections: Steve Bracks, Morris Iemma, Peter Beattie and John Howard. (That is, Howard is most likely to go down.) This is a little different to his assessment to the Oz's Imre Salusinzky on May 1:

"I think there is one chance in three that Iemma will go, one chance in four that Howard will go, one chance in five that Beattie will go and one chance in 50 that Bracks will go."

Malcolm has therefore shifted Iemma from most likely chop-getter to number three, leaving Howard at the front.

July 26 Furphy watch: van Errington on Howard

Peter van Onselen and Wayne Errington (or 'van Errington', collectively) who are bio-ing our PM, have a piece in today's Oz. As you know, over-estimating Howard's electoral/political prowess is an all too common crime, against which we must constantly be vigilant. A couple of things grab the attention.

Write van Errington:.

• "Before he even entered parliament Howard was a master of marginal seat campaigning. As campaign manager in 1963 he secured Tom Hughes's elevation into federal parliament when Hughes had agreed to contest the seat only because it was considered unwinnable. "

The 1963 2pp swing that won Hughes Parkes was 5.9 percent. 1963 was a big victory for the Menzies government (Menzies' last, although not nearly as big as Harold Holt's 1966 one), off the back of 1961's one seater. The national swing was 3.1, against which 5.9 looks impressive. However, the NSW component was 5.5. And Hughes might himself like to claim some credit.

In addition, Howard's first election as opposition leader, in 1987, saw a positive one percent swing translated into a net loss of seats, which does not bespeak marginal seat soundness.

Still, let's grudgingly half-pay that one. The same can't be said for the following:.

• "One of the great shifts in Howard's approach to political campaigning from earlier prime ministers has been his targeting of Opposition-held seats."

This is just Howard-hyperbole (Howperbole?), as all government campaigns target vulnerable opposition seats, often succeeding. Most starkly see the 1987 (again) and 1993 results, when federal Labor increased its majority, but even when a government goes backwards net-wise it often picks up a seat or several along the way. Eg in 1990 Hawke did well in Qld and northern NSW. And should we really give all the glory to the PM, rather than the countless, faceless ones behind the scenes?

Still, van Errington are only mild Howperbolists in the grand scheme. Looking forward to the book.

July 19  George and John in The Walkley Magazine

I have a piece in current issue of Walkley Magazine. Cartoon by Bill Leak. Be warned, nothing to do with electoral business. Contains a 600kb image. As you would know, lead-in not written by me. Here.

July 18 Newspoll also says 52 to 48

It's difficult to imagine the Oz commissioning a survey on attitudes towards Mr Howard's honesty, but apart from that, similar stuff to yesterday's Nielsen.

July 17 Nielsen says 52 to 48

(The primary numbers would perhaps give closer to 53 to 47.)

Nicholson animations

Peter "Rubbery Figures" Nicholson at the Oz now makes fortnightly animations. Here's his latest; there's a facility at the end if you want notification of every new one.

Odds, bets etc

I've previously put the PM's chances of retiring before next election at about 67 percent. I now reckon it's about 50-50. And if that happens, Peter Costello's odds of taking over have roughly gone from 90% to 50%.

However, my 10 to 1 wager on Brendan Nelson being PM at next election, which had been looking forlorn recently, now appears healthier. That is, Howard retires but engineers an 'anyone but Peter' scenario.

July 11 John, Peter and the one shot in Glenn's locker

John Doyle delivered last year's Andrew Olle lecture for the ABC, in which he said this:

I remember working with a high profile journalist on a commercial network a few years ago who stated that his ambition was to work as Peter Costello’s press secretary when he assumes the Prime Ministership. His newspaper column has been nothing but full of praise for the Treasurer ever since.

Joining those dots should only require one try. Go Glenn, go.

July 9 Sawford and Bennelong

1. Re Rod Sawford's formula, July 7 below,  Leo from Queensland has done the research and reckons:

"A glance at the data shows inflation rising from 1.8% in 1993 to 2.6% in 1996 (4.6% in 1995) and nominal interest rates rising slightly through the term. So Rod the Lemming's formula sort of works in 1996. It is also correct that inflation and unemployment were lower in 1990 than 1987, though unemployment had started to rise again before the election.

1974 and 1980 are two where he is definitely wrong. Inflation and nominal interest rates both rose in Whitlam's first term, but he was returned just 1 seat the poorer (from memory). All three rose between the 1977 and 1980 elections, but Fraser was returned as well.

Oh - also, the 2001 election ALMOST stuffs him, but for some very-late- term rate cutting by the RBA in 2001. If he's being consistent, Rod must reckon those interest rate decisions in the second half of 2001 cost Bomber the premiership.

Cheers, Leo"

2. Bennelong

Pollbludger has lots more on Bennelong and other electorates. Also, NSW Labor guy Shane Easson has this to say. I haven't read it fully yet, but he believes Howard's margin has barely changed.

And Malcolm Mackerras, always required reading, has new margin estimates over at Crikey.

July 7 Playing catchup

Computer problems. (Again!) Since we last spoke a Newspoll has shown 53 to 47 (after Australia going down to Italy, so not inconsistent with The Beautiful Game Pt II post on June 30, below) which has set tongues wagging about a government in decline.

There's this from today's SMH, about dreadful Lemming Rod "we warn Kim Beazley" Sawford's dad's formula for picking elections:

The Sawford formula relies on three indicators - the unemployment rate, inflation and interest rates.

If two or more of these rise over a full, three-year electoral cycle, the government will lose. Conversely, if two or more fall the government will be returned.

It has 'proved correct for every federal election result since ...1961'.

But just just off the top of the head - without checking -  it falls down at the 1990 and 1996 results, doesn't it? And how about 1980? [update: see July 9, above.]

And how about this, near the end:

As to retiring, Howard sniffs the breeze better than most; if he feels the Coalition's best chance rests with Costello, he may do the selfless thing and step down.

I'd take out "the Coalition's best chance rests with Costello" and replace it with "the Coalition might get done like a dinner, irrespective of who's leading it".

July 4 My bets

The other day I came across this bet I must have taken several months ago:

$50 @ 4 to 1 that John Howard leaves office in 2006. For something that is more likely to happen than not, this is wonderful value for money, and the AEC's NSW redistribution, which shaved a couple of percent off Howard's margin, just made it better.

Because of aberrational nature of 2004 Bennelong election, you would still have to favour the Liberals holding it in 2007. But despite Barrie Cassidy's Insiders assessment that the PM 'never runs away from a fight', Howard is nothing if not twitchy, and no-one likes even the possibility of losing their seat.

June 30 The beautiful game I Secrets of Italian soccer training here. (2MB video, probably better to save to your hard-drive first, by right-clicking and 'saving target as'.)

The beautiful game II

A slew of opinion polls this week and last, from Newspoll, Morgan and Nielsen, have governments in NSW, Victoria and Qld - and the Federal one - all making ground. 

Something to do with millions of Australians wandering around tired and a little elated (until Tuesday morning)? Not as silly as it sounds, and at least as sensible as the usual fortnightly/monthly explanations.

June 26 Changes to electoral law

Mr Poll Bludger runs a fine summary of the electoral law changes ticked off in the Senate this week. Alan Ramsey quotes Senator Robert Ray's version of Australian election law history. Despite Graham Richardson's boast in Whatever it Takes - noted several times recently by Government members - along the lines of 'we changed the electoral laws in 1983/4 to ensure we stayed in power as long as possible', Ray's version is pretty well correct. The changes two decades ago were praise-worthy: fair boundaries, independent Commission etc. This is not to say self-interest was absent in Labor's motivation, eg public funding, but that too was a positive development.

Let's not beat around the bush: the Government's 2006 changes are designed for electoral advantage. You don't need to believe in conspiracies, it's just always been the case that 'conservatives' favour a narrow franchise and 'liberals' a wide one. It was, for example, a Labor government (Whitlam's) which lowered the voting age from 21 to 18. 

Both sides of Australian politics believe that if lots of members of certain groups - young, in jail, don't always have a drivers' licence handy, change residence a lot or live overseas - drop (or stay) off the electoral roll, the net beneficiary will be the Coalition. And they're right.

My hat number 2 deals in part with the 19th century introduction of reforms in Australia reeled off by Senator Ray and others (see below) - Australian ballot, female suffrage etc etc.

South Australia was the biggest innovator of the six pre-federation 'countries', and I've just posted a paper on the development of a centralised electoral administration. More over there.

June 22 About me 

An obviously wise and thoughtful Senator Ruth Webber (WA) said nice things about me and my other hat in the Senate last Friday. (Here, pg 12, right column.)

She may have seen my short (and slightly fiery) submission to the Inquiry into civics and electoral education

June 21 'Put a 4 in front of it' part II

Further to this post of a few weeks ago. Nielsen this week had 51 to 49 two party preferred, from primary votes of 36 (ALP) to 41 (Coalition) and Greens on 11.

The 'it needs a 4 in front of it!!' crowd would say: 'There is no way in the world Labor can win an election with 36 percent of the primary vote'. They may have a point. 

But shift boot to other foot and ask: could the Coalition win an election with 41 percent if Labor and Greens have 47 between them? Answer: not bloody likely.

My understanding is that Nielsen, unlike Newspoll, reads out parties to respondents and asks them to choose, which might explain their often high minor party numbers. But Nielsen's 2pp at the last two elections was the closest of all to the actual result. Unlike Newspoll, they give preferences appropriate respect.

A long way of saying: watch the two party preferred, not the first preference.

June 20 Newspoll says 49 to 51

June 19 Nielsen says 51 to 49

Newspoll tomorrow.

June 16 Vanstone leadership? Quite a few readers have expressed incredulity (to put it mildly) at my inclusion of Senator Vanstone in the list below, at odds 5 to 1 to be PM one day. They say I'm too kind to Amanda, and they may have a point: she's in the Senate, for one thing, and can be a political klutz, for another. But she has a kind of knockabout persona and is 'tough', and may contend in a future political universe.

Perhaps. Ok, I've downgraded to 9 to 1. 

June 12 A list: Prime Ministerial odds

At any point in time, the chances of a particular politician becoming Prime Minister at some stage in their life are usually small. Those who have already succeeded might appear, with hindsight, to have always been on target, but life ain't really like that and timing and luck - as well as ability - are always large determinants.

Peter Costello, however, right now, is a rarity: someone who isn't PM but will probably have the job one day. This is because his leader will probably retire before the next election, and Costello would then almost certainly replace him. Using basic stats (the only type I know) it might go like this: the chances of Howard retiring this term are 0.7, and if that happens, the chances of Costello then becoming Liberal leader are 0.9. So Costello's chances of becoming PM this term are 0.7 x 0.9 = 0.63 - a little under two-thirds, or approximately two in three.

Actually, you'd have to also factor the odds of Howard staying on, winning the election, and Costello becoming PM after that (a third times a third times nine-tenths, perhaps?), plus all other permutations. Let's give Peter a round-ish .75 aggregate, which is another way of saying he has three chances in four of becoming PM one day. 

Below is a list of everyone vaguely in the running at this stage.

Chances of individuals eventually becoming PM

Peter Costello: 3 chances in 4
Kim Beazley: 1 in 3
Kevin Rudd: 1 in 4
Julia Gillard: 1 in 5
Brendan Nelson: 1 in 5
Malcolm Turnbull: 1 in 5
Amanda Vanstone: 1 in 6 [updated to 1 in 10; see June 12]
Bill Shorten: 1 in 7
Tony Abbott: 1 in 10

As you can see, next cab off the rank is daylight, followed by Beazley. I reckon he has a one in two chance of surviving as leader of the ALP until the next election, and then two chances in three of winning it. Multiply them together and you get .33. So Beazley has about one chance in three of being PM.

Shorten is down the list because there's a decent probability that once the smoke clears he'll be revealed as a dull and empty vessel. Abbott's chances are poor because making him opposition leader will always be risky: too 'divisive'; might give women voters the creeps. (A handover during government, if he's still around for the next Coalition spell, is be more likely.)

The end. I'm sure I've missed a few people.