What if we had Proportional Representation on August 21st?

There are many forms of PR: quotas, state by state,
allowing preferences etc.

Let us suppose that we just a simple national version for the 150
member House of Reps based on 1st preference votes only.

The quota to get any representation could be just the total
number of formal votes divided by 150 seats =
82,683 formal first preference votes.
The CDP makes it over the line, One Nation and
the Country Liberals do not make it.

Say we just ignore the votes that were wasted on the small
parties that did not make this quota. Thus 11,903,018
votes are not wasted and 499,365 votes are wasted.

Now let us divide up the 150 seats amongst the parties
that made it over the quota according to what
proportion of the 11,903,018 votes they got.

The seats in the new parliament [vs current] would be:

ALP 59 seats [actually now 72]
Greens 18 seats [1 seat]
Liberal 48 seats [44]
LNP(Qld) 14 seats [21]
Nationals(all states) 6 seats [7]
Family First 4 seats [0]
CDP 1 seat [0]

So we would also probably have an ALP-Green Coalition
government led by PM Gillard, but it would be
Greener, less Indy-ish and hold a slightly
more substantial majority.

The Two-Party Preferred Vote in O'Connor

The counting has been finished and the AEC reports that the two-party preferred (TPP) result in the seat of O'Connor is

Australian Labor Party

22,029 votes

27.00 %

Liberal/National Coalition

59,555 votes

73.00%

The margin is 37,526 votes.

However, this is not quite correct as it is presented.

The AEC actually distributed all the 81,584 formal votes in O'Connor into two piles: those that preferred Ian Bishop(ALP) over Tony Crook(Nat) and those that preferred Crook over Bishop.

The problem with this is that Tony Crook, along with other WA National Party candidates campaigned on a platform which included the intention of not joining any Liberal National Coalition. Mr Crook reaffirmed that after the election. Mr Crook has not joined any such Coalition since being elected (and it it is not clear there is any formal Coalition agreement now
anyway).

There was actually a sitting Coalition member for this seat, Wilson Tuckey, who was running for re-election and so Mr Crook was clearly not running as Coalition candidate.

It would thus seem more correct for the AEC to have distributed the votes in O'Connor on a TPP basis as Bishop versus Tuckey. This is what happens in other seats where there are Independents or Minor Party candidates in the final two positions.

What difference would that have made to the final figures. We can not tell exactly but below I attempt and estimate.

There are 9 candidates with first prefs as follows:
HUNTLEY, Andy: The Greens: 7232
ROBINSON, Jean: Citizens Electoral Council: 375
TUCKEY, Wilson: Liberal: 31294
SCALLAN, Pat: Family First: 1164
YOUNG, Jacky: Christian Democratic Party: 2221
CROOK, Tony: The Nationals: 23538
STOKES, Geoffrey: Independent: 1298
BISHOP, Ian: Australian Labor Party: 13962
SMITHSON, Neil Ramsay: Independent: 500

The TPP candidates are as follows:
BISHOP, Ian
CROOK, Tony

We can estimate how many votes get distributed to the two TPP candidates.
Best match of TPP booth results is with following preference flows:
From HUNTLEY, Andy to BISHOP, Ian: 79.74% which amounts to 5767 votes.
From ROBINSON, Jean to BISHOP, Ian: 52.88% which amounts to 198 votes.
From TUCKEY, Wilson to BISHOP, Ian: 1.81% which amounts to 566 votes.
From SCALLAN, Pat to BISHOP, Ian: 41.14% which amounts to 479 votes.
From YOUNG, Jacky to BISHOP, Ian: 5.06% which amounts to 112 votes.
From CROOK, Tony to BISHOP, Ian: 0% which amounts to 0 votes.
From STOKES, Geoffrey to BISHOP, Ian: 36.93% which amounts to 479 votes.
From BISHOP, Ian to BISHOP, Ian: 100% which amounts to 13962 votes.
From SMITHSON, Neil Ramsay to BISHOP, Ian: 93.11% which amounts to 466 votes.

We also have a two-candidate count in O'Connor which was distributed on a Tuckey versus Crook basis. The result was

TUCKEY, Wilson

Liberal

37,891

46.44 %

CROOK, Tony

The Nationals

43,693

53.56 %

The estimated preference flows are:
Best match of TCP booth results is with following preference flows:
From HUNTLEY, Andy to CROOK, Tony: 70.17%
From ROBINSON, Jean to CROOK, Tony: 0.58%
From TUCKEY, Wilson to CROOK, Tony: 0%
From SCALLAN, Pat to CROOK, Tony: 29.43%
From YOUNG, Jacky to CROOK, Tony: 26.97%
From CROOK, Tony to CROOK, Tony: 100%
From STOKES, Geoffrey to CROOK, Tony: 10.68%
From BISHOP, Ian to CROOK, Tony: 100%
From SMITHSON, Neil Ramsay to CROOK, Tony: 7%

So my estimates of the preference distribution between Tuckey and Bishop would be:

From HUNTLEY, Andy to BISHOP, Ian: 79.74% which amounts to 5767 votes.
From ROBINSON, Jean to BISHOP, Ian: 52.88% which amounts to 198 votes.
From TUCKEY, Wilson to BISHOP, Ian: 0% which amounts to 0 votes.
From SCALLAN, Pat to BISHOP, Ian: 41.14% which amounts to 479 votes.
From YOUNG, Jacky to BISHOP, Ian: 5.06% which amounts to 112 votes.
From CROOK, Tony to BISHOP, Ian: 20% which amounts to 4708 votes.
From STOKES, Geoffrey to BISHOP, Ian: 36.93% which amounts to 479 votes.
From BISHOP, Ian to BISHOP, Ian: 100% which amounts to 13962 votes.
From SMITHSON, Neil Ramsay to BISHOP, Ian: 93.11% which amounts to 466 votes.

The main figure here is the 20% flow from Crook to Bishop (over Tuckey) in keeping with the sort of flow measured in O'Connor in 2007.

The end result is:
Bishop 26,171 or 32.08%
Tuckey 55,413 or 67.92%

The margin is 29,242 votes.

That is better margin for the ALP and would have given them an extra 8,284 votes in their National TPP win.

The final National Tally would have then been
ALP 50.19% versus LNP 49.81%

Labor Wins the Two-Party Preferred Vote

ALP 50.12% to LNP 49.88%

see the AEC site now that the counting is over
http://vtr.aec.gov.au/HouseStateFirstPrefsByParty-15508-NAT.htm

See the short ABC item
http://www.abc.net.au/news/stories/2010/09/17/3015006.htm

Compare with their earlier erroneous headlines
http://rage.net.au/news/video/2010/08/30/2997877.htm

Why is the latter bigger?

Current First-Preference Vote Count HOR (almost finished)

When Julia Gillard went to the Governor General this week she
took a list of names of MHRs who had promised to support her in parliament. Together they represented a collection of parties and independents which attracted a certain number of first-preference votes across the country. How many people voted with their first preference for one of those parties or independents, those that support the government?

Government supporters: ALP+Green+Wilkie+Oakeshott+Windsor
6,280,629 first preference votes (50.6% of all formal votes)

Supporters of Abbott: Lib+LNP+Nat+CLP+Katter
5,446,806 first preference votes (43.9% of all formal votes)

Margin: 833,823 votes to the current Government

How does that compare with the number of first-preference votes represented at previous elections?

2010: Gillard ALP supported by Green Party + Independents ... 50.6% of first-preference votes

2007: Rudd ALP ... 43.38% of first-preference votes

2004: Howard Liberal/National Coalition (incl CLP) ... 46.7% of first-preference votes

2001: Howard Liberal/National Coalition (incl CLP) ... 43.0% of first-preference votes

1998: Howard Liberal/National Coalition (incl CLP) ... 39.5% of first-preference votes

1996: Howard Liberal/National Coalition (incl CLP) ... 47.3% of first-preference votes

1993: Keating ALP ... 44.9%

1990: Hawke ALP ... 39.4%

1987: Hawke ALP 45.8%

1984: Hawke ALP 47.6%

1983: Hawke ALP 49.4%

1980: Fraser Coalition 46.4%

1977: Fraser Coalition 48.1%

1975: Fraser Coalition 53.1%

1974: Whitlam ALP 49.3%

1972: Whitlam ALP 49.6%

1969: Gorton Coalition 43.3%

1966: Holt Coalition 49.98%

1963: Menzies Coalition 46.0%

1961: Menzies Coalition 42.1%

1958: Menzies Coalition 46.6%

1955: Menzies Coalition 47.6% *

1954: Menzies Coalition 46.8% *

1951: Menzies Coalition 50.3% *

1949: Menzies Coalition 50.3%

1946: Chifley ALP 49.7%

1943: Curtin ALP 49.9% *

1941 October re-alignment: Curtin ALP supported by Independents 40.2%* plus ?

1941: Menzies UAP+CP supported by Independents 43.9%* plus ?

1937: Lyons UAP+CP 49.3% *

1934: Lyons UAP+CP 45.6% *

1931: Lyons UAP+CP 48.4% *

1929: Scullin ALP 48.8% *

1928: Bruce Nationalist/CP 49.6% *

1925: Bruce Nationalist/CP 53.2% *

1922: Bruce Nationalist/CP 47.8% *

1919: Hughes Nationalist 45.1% *

1917: Hughes Nationalist 54.2% *

1914: Fisher ALP 50.9% *

1913: Cook CLP 48.9%*

1910: Fisher ALP 49.97% *

1906: Deakin Protectionist+Labour 53.08%*

1903: Deakin Protectionist+Labour 60.7%*

1901: Barton Protectionist+Labour 52.5%*

*= some MHRs were elected unopposed in these elections and so
we do not include any votes for them

SO 2010 is the first time, in 35 years, that the Government has the first-preference support of more than half the population. It is only the tenth time in Australian Federal Government history that a Government commands support of a majority of first-preference votes. It is only the second time ever that an ALP Prime Minister has governed with a majority of first-preference votes behind him or her (the only other time was Andrew Fisher after the 1914 election).

The Two-Party Preferred Vote and Legitimacy

1)
Senator Nick Minchin 21st August

From the Sydney Morning Herald on August 22nd,

http://www.smh.com.au/federal-election/powers-in-balance-after-rise-of-the-crossbenchers-20100821-13a42.html

"Senior Liberal Nick Minchin said the independents should respect the major party that had the highest two-party-preferred vote and the most seats."

(This was apparently said on TV on election night.)

2)
Senator Brandis

8th September ABC 612 Brisbane Mornings with Chris O’Brien

Senator Brandis
"the Coalition actually won the majority of the votes, about 700,000 well, primary votes, and importantly it won several thousand more two-party preferred votes"

3)
Julie Bishop 31st August

http://www.abc.net.au/news/stories/2010/08/31/2997897.htm

From the ABC:
"Liberal deputy leader Julie Bishop says Prime Minister Julia Gillard has lost her mandate to form government after Labor has lost its lead in the two-party preferred vote."



4)
Mirabella sept 14th

http://www.dailytelegraph.com.au/news/breaking-news/gillard-to-explain-claim-to-leadership/story-e6freuyi-1225912314309

"If Ms Gillard's main moral pillar of seeking to form government was that she was ahead on the two-party preferred vote then that has dissipated," Liberal frontbencher Sophie Mirabella told Sky News.

5)
Tony Abbott

From Liberal news

http://www.liberal.org.au/Latest-News/2010/09/08/Tony-Abbott-interview-with-Alan-Jones.aspx

ALAN JONES:

Is it a healthy democracy when a Party wins the majority of the two party preferred, wins the majority of the primary vote and wins more seats in the Parliament than the other Party but the other Party forms government?

TONY ABBOTT:

Again, Alan, it’s disappointing when all these things are true but it is our system.

from

http://www.dailytelegraph.com.au/news/tony-abbott-admits-disappointment-coalition-did-not-get-chance-to-govern/story-e6freuy9-1225915512964

"The coalition won more votes and more seats than our opponents but sadly we did not get the opportunity to form a government,'' he said.

Latest on the National TPP count

ALP 24,901 votes ahead out of 12.0 million votes counted.

I think that is about 96% of the expected number of final formal votes counted.

My latest estimates for the final result are:

Total number of formal votes in the 142 normal seats: 11,825,318

ALP to get 5,914,739 of them, that’s 50.02%.
LNP to get 5,910,579 of them, that’s 49.98%.
margin 4160 to ALP.

Then the eight special seats.

ALP to win Batman TPP by a margin of 40,281. (76%)
ALP to win Melbourne by 42,069. (74%) [counting underway]
ALP to win Grayndler by 36,626. (72%)
ALP won Denison by 20,525. (ALP 66%) [counting finished]
LNP to win Kennedy by 21,248. (ALP 37%)
LNP won Lyne by 21,163. (ALP 38%) [counting finished]
LNP to win O’Connor by 36,258. (ALP 28%)[counting underway]
LNP to win New England by 34,642. (ALP 31%)

Final total margin:
ALP by 30,350

Out of about 12,500,000 votes that’s a final
result of:

ALP 50.12% – LNP 49.88%

The TPP in Lyne

Here is my analysis of Oakeshott’s and Lyne’s voters using the
final TCP and TPP results for the seat. In the coming weeks
we will get some actual data on some of the preference flows
but in the meantime I am using estimates based on the flows
that best match the individual booth results.

So Rob Oakeshott got 47% (=40,061) of the first preference votes in Lyne.

About 57% of his 1st pref supporters (=22,835) went on to
preference the Nationals over the ALP.

The other 43% of his 1st pref supporters (=17,226) preferred
the ALP over the Nationals.

However, Oakeshott did not win the seat on 1st preferences,
he needed preferences from ALP, Greens and another
independent to keep the seat from his National opponent.

Oakeshott ended up with about 63% (=53,297) of the
two-candidate preferred (TCP) vote in Lyne to hold the
seat against the National.

How was Oakeshott’s winning 53,297 made up?

22,835 (or 43%) were his National leaning 1st pref voters.
17,266 (or 32%) were his ALP leaning 1st pref voters.
10,656 (or 20%) were ALP 1 voters who preferenced Oakeshott over Nationals.
2,520 (or 5%) were Green 1 voters who pref’d him.
The other 20 were prefs from the other Independent.

So that makes about 57% of Oakeshott’s supporters,
the ones who gave him the seat, being left-leaning voters.

It is thus obvious that Oakeshott stands clearly to the left of the
LNP (without even having to look in detail at his
election material).

Optional Preferential Voting and the 2010 election

Rod Hagen on Pollbludgers mentioned an article by Mumble

Rod pointed out that Mumble had claimed that
Optional Preferential Voting (OPV) would have delivered Australia
to Abbott in the 2010 election.

However, you pointed out that OPV might not be so bad for
the ALP as 1) more of the currently informal votes that
would become formal would favour the ALP and
2) the Greens would generally be more careful to make sure that
they indicated their preferences among the major parties.

You said that Mumble disagreed.

So, I had a look. I took the results of the vote across the country
and plugged them in with a couple of adjustments to
take account of your model of OPV. In particular I gave
the ALP 2% of the 5% informals and the LNP 1%.
I also assumed that 80% of the Greens would indicate
their preference ALP or LNP. I assumed that only 50%
of the minor party and independent voters would
indicate their preferences.

What are the results of this simulation?

All the 4 independents and Bandt and Crook still win.

141 of the other seats turn out exactly the same.

Only three seats change winners under OPV.
They are:

Forde gets won by the ALP instead of LNP.

Herbert gets won by the ALP instead of LNP.

Hasluck gets won by the ALP instead of LNP.

So I’m with you Rod. Under reasonable assumptions
on the effects of OPV, the ALP wins 3 more seats.

Final result. ALP 75 seats, Green 1, Inds 4, NatWA 1,
Coalition 69.

Latest TPP counting.

LNP 2355 votes ahead out of 11.7 million votes counted for the 142 normal seats.

I think that is about 97.3% of the expected number of final formal votes counted
in those 142 seats.

My latest estimates for the final result are:

Total number of formal votes in the normal seats: 11986353

ALP to get 5988575 of them, that’s 49.96%.
LNP to get 5997778 of them, that’s 50.04%.
margin 9203 to LNP.

Then the eight special seats.

ALP to win Batman TPP by a margin of 41,902. (76%)
ALP to won Melbourne by 39,358. (71%)
ALP to win Grayndler by 38,132. (72%)
ALP to win Denison by 21,373. (66%)
LNP to win Kennedy by 20,785. (ALP 37%)
LNP to win Lyne by 20,877. (ALP 38%)
LNP to win O’Connor by 32,476. (ALP 31%)
LNP to win New England by 34,857. (ALP 31%)

Final total margin:
ALP by 22,477

Out of about 12,700,000 votes that’s a final
result of:

ALP 50.09% – LNP 49.91%

Federal Election 2010

Hung Parliament on 52-48 Scenario

August 17

Some say that with the state of the national polls
eg Newspoll ALP 52-48 it is impossible for
the ALP to not win majority government.

However, it is perfectly possible for the ALP
to win only 75 seats with 52% of the vote.

The following
scenario sets out how this might happen and
is roughly consistent with the latest national polls,
state-wide polls. individual seat polls and
individual seat betting markets.

State-by-state.

NSW (average 1.5% swing to Liberals, ALP end up on 51.9%)
Loss of Macquarie, Robertson, Gilmore, Macarthur
Bigger swing in Lindsay (-7%) loss
Lesser swings in Bennelong (-0.6%) hold

Qld (average 3.7% swing to LNP, ALP end up on 46.9%)
Loss of Flynn, Dickson, Dawson, Herbert
Bigger swing in Leichhardt (-5%) loss
Lesser swing in Forde (-2.5%) hold
Lesser swing in Longman (+2%) hold

Vic (average 3.2% swing to ALP, ALP end up on 57.2%)
Gain of McEwen
Lesser swing in La Trobe (1% to Libs) Lib hold
Greens win Mebourne

WA (average 0.3% swing to ALP, ALP end up on 47.0%)
Lesser swing in Swan (4% to Libs) Lib gain
Lesser swing in Hasluck (1% to Libs) Lib gain

SA (average +0.1% swing to ALP, ALP on 52.3%)
no change

Tas (av +0.6% swing to ALP)
no change

ACT (av 0% swing to ALP)
no change

NT (av -0.3% swing to Libs, ALP on 55.1%)
ALP lose Solomon

Overall TPP is 52.0%
(0.7% swing to Coalition)

Overall seat result:
ALP 75, Lib 71, Ind 3, GRN 1

Hung Parliament

Federal Election 2010: Analysis of Individual Seat Betting Aug 17th

Likely state of parties after the election according to the individual
seat odds of Sportingbet as of this morning (aug 17).

Overall situation: around ALP 75, LNP 71 , Ind 3, Green 1 => Hung Parliament

Note that the odds have changed a little over the last few days.
There are now less seats which are close to 50-50 probability.
There has also been a bit of a move by a few NSW seats to the
Coalition.

The change to a hung parliament today is a move
in the opposite direction to the moves on
the betting market for the overall government.

Details (according to current odds):

1)
ALP to lose Melbourne to the Greens
(Prob Green win =55%, Prob ALP =34%)

2)
ALP have 63 safe seats including Eden-Monaro and Longman.

3)
ALP look very likely to retain a further 7 of the following 9
(in which they have good odds of retaining):
Greenway (which was Lib before redistribution), Page, Dobell,
Bonner, Petrie, McEwen(a gain), Bass, Deakin, Brisbane,
But they may lose 2.

4)
ALP should be able to win 2 of the following 3 seats as they are just ahead:
Corangamite, Forde, Bennelong
But they may not get 1 of these.

5)
ALP should be able to win at least 2 of the following 5 seats
where they are just behind, but they will most likely lose
the other 3:
Hasluck(ALP), Solomon(ALP), Lindsay(ALP),
Herbert(Lib/notional ALP), La Trobe (Lib).


6)
ALP should be able to win 1 of the following 5 seats where they are behind
but have the cumulative probability of one win:
Hughes(Lib), Macarthur(Lib, notional ALP), Canning(Lib), Swan(Lib, notional ALP), Dawson(ALP)

7)
There are 6 seats which were in doubt but now seem likely
to be won by the Liberals. Likely 0 to ALP:
Macquarie(Lib), Gilmore(Lib, notional ALP), Boothby(Lib),
Dickson(Lib, notional ALP), Ryan(Lib), Flynn(Lib)

7)
There are 55 safe L/NP seats.

8)
There are 3 safe independents.

9)
Note that there are 75 seats which the odds say that the
ALP are more likely to win than not win, but because there
are more seats just on the ALP side than just on the
Coalition side, the expected number of wins is
a bit less 74.7.

Total 150 seats. Here is a table (click for a larger version).

























This also includes estimates of swings to the ALP
based on the bookies odds.

Federal Election 2010: Poll of Polls 15th August

We've had a few polls published this weekend.

1)
There are also the latest individual seat betting odds.
http://dr--good.blogspot.com/2010/08/federal-election-2010-analysis-of_14.html

2)
There is a Neilsen 53-47
http://blogs.crikey.com.au/pollbludger/2010/08/13/nielsen-53-47-to-labor/
showing a 0.3% swing to ALP if taken on a uniform national
basis.


3)
There is a Morgan phone poll 51-49 with state by state breakdowns
http://www.roymorgan.com/news/polls/2010/4559/
It suggests swings of
-6.3, +4.2, -3.9, +0.8, -0.4, -3.7 across NSW, Vic, Qld, WA, SA
and Tas respectively. I also include a recent individual seat poll
by Morgan which gives Solomon to the CLP.

4)
There is a newspoll marginals poll
http://blogs.crikey.com.au/pollbludger/2010/08/14/newspoll-marginals-poll-nsw-queensland-victoria/
It suggests swings of
-1.3, +6.2, -3.4 across NSW, Vic, Qld, respectively

5)
There is the dodgy monster Galaxy
http://blogs.crikey.com.au/pollbludger/2010/08/15/galaxy-marginals-polls-and-the-rest/
which is pointing towards an ALP 51-49 TPP but
some claim also suggests swings of
-2.4, +1.6, -5.4, -2.1, 0 across NSW, Vic, Qld, WA and SA respectively.
I think that they have not chosen sensible electorates
to get a picture of the statewide swings but anyway ...

Let's compare their predictions.

Here is a table of electorates with the respective
predictions and a total at the end of how many
of our 5 recent oracles, are predicting an ALP
win. Click for a bigger version. This only includes
marginals.























There are 57 safe ALP seats not listed so
you can use the number in the first
column to total up the overall ALP
wins. ( I don't include Melbourne)

Looks like about 79 to the ALP plus or
minus a few.

Federal Election 2010: Analysis of Individual Seat Betting Aug 15th

Likely state of parties after the election according to the individual
seat odds of Sportingbet as of this morning (aug 15).

Overall situation: around ALP 76, LNP 70 , Ind 3, Green 1 => ALP majority government

Details (according to current odds):

1)
ALP to lose Melbourne to the Greens
(Prob Green win =55%, Prob ALP =34%)

2)
ALP have 63 safe seats including Eden-Monaro and Longman.

3)
ALP look very likely to retain a further 7 of the following 10
(in which they have good odds of retaining):
Greenway (which was Lib before redistribution), Page, Dobell,
Bonner, Petrie, McEwen(a gain), Deakin, Brisbane, Bennelong, Lindsay
But they may lose 3.

4)
ALP should be able to win 3 of the following 4 seats as they are just ahead:
Forde, Bass, Corangamite, Hasluck
But they may not get 1 of these.

5)
ALP should be able to win at least 2 of the following 6 seats
where they are just behind, but they will most likely lose
the other 4:
Solomon(ALP), Hughes(Lib), Herbert(Lib), Macarthur(Lib),
Swan(Lib), Macquarie(ALP)

6)
ALP may win 1 of the following 9 seats where they are behind
but have a significant probability of a win (i.e. it adds up over the 9
seats):
La Trobe, Canning, Dawson(ALP), Bowman, Leichhardt, Gilmore,
Boothby, Dickson, Ryan

7)
There are 54 safe L/NP seats.

8)
There are 3 safe independents.

9)
Note that there are 77 seats which the odds say that the
ALP are more likely to win than not win, but because there
are more seats just on the ALP side than just on the
Coalition side, the expected number of wins is only 76.

Total 150 seats. Here is a table (click for a larger version).
























This also includes estimates of swings to the ALP
based on the bookies odds.

Federal Election 2010: Analysis of Individual Seat Betting Aug 11th

Likely state of parties after the election according to the individual
seat odds of Sportingbet as of today (aug 11).

Overall situation: around ALP 76, LNP 70 , Ind 3, Green 1 => ALP majority government

Details (according to current odds):

1)
ALP to lose Melbourne to the Greens
(Prob Green win =55%, Prob ALP =34%)

2)
ALP have 63 safe seats including Greenway (which was Lib before redistribution)
and Eden-Monaro.

3)
ALP look very likely to retain a further 7 of the following 9
(in which they have good odds of retaining):
Booner, Page, Longman, Petrie, Bennelong, Deakin, Dobell, Brisbane, McEwen(a gain)
But they may lose 2.

4)
ALP should be able to win 3 of the following 5 seats as they are just ahead:
Bass, Corangamite, Forde, Lindsay, Hasluck
But they may not get 2 of these.

5)
ALP should be able to win at least 2 of the following 8 seats
where they are just behind, but they will most likely lose
the other 6:
Macarthur(Lib), Solomon(ALP), Hughes(Lib), Herbert(Lib), Dickson(Lib), Flynn(ALP),
Swan(Lib), Macquarie(ALP)

6)
ALP may win 1 of the following 8 seats where they are behind
but have a significant probability of a win (i.e. it adds up over the 8
seats):
Leichhardt(ALP), Canning, La Trobe, Robertson (ALP),
Dawson(ALP), Bowman, Boothby, Gilmore

7)
There are 53 safe L/NP seats.

8)
There are 3 safe independents.

9)
Note that there are 77 seats which the odds say that the
ALP are more likely to win than not win, but because there
are more seats just on the ALP side than just on the
Coalition side, the expected number of wins is only 76.

Total 150 seats. Here is a table (click for a larger version).
























We can also roughly estimate the expected swings by using the
seat odds and current margin. See the last column.

How good are the Individual Seat Odds from the Bookies at predicting the winner of the individual seat?

Below are some results from the 2007 election. (Click for a larger image)

These show some of the seats with the winner, two-party preferred vote
after the election and before and the swing away from ALP. They
show the Sportingbet probability of ALP winning the seat as offered on
Nov 23rd, election eve. Thanks Simon Jackman for keeping that
data.

We also see the winner of the seat according to three different prediction
methods.

First is by using the uniform national swing reported in the Newspoll
of Nov 23.

Next is the winner according to Sportingbet. (Sportingbet didn't
give odds for Hinkler and Kennedy like many seats that
were assumed to be very safe. So we just assume that
Sportingbet was assuming that the current holder would
win.)

Last is the winner according to state by state uniform swings
as predicted by Newspoll on Nov 23rd.

There are 26 seats listed. All three methods agreed in their predictions
on all the other 124 seats and they were all correct
in those predictions. These are the seats where at least one method
got the winner wrong.





















So what was the best way to predict the winner of a seat?

Here are the results. Recall that the ALP won 83 of the
150 seats. (They won 110 of the seats not listed here).

Method 1: (use the latest Newspoll and assume a uniform
national swing)
You get 134 predictions correct and you think the ALP will
win 75 seats.

Method 2: (use sportingbet odds for each set)
You get 138 predictions correct and you think the ALP will
win 81 seats.

Method 3: (Use the latest Newspoll state by state swings,
with the national swing for NT, ACT and Tas)
You get 132 predictions correct and you think that the ALP
will win 83 seats.

So, the most accurate predictor of individual seat winners
was to use the bookies' odds.

You can do some more work and find out the expected
number of seats won by each party but we'll come
back to that later.

Federal Election 2010:
Analysis of Individual Seat Betting Aug 9th

Expected Swings in the most marginal seats.

The odds of winning individual seats tell us something about the swings expected in those seats. Eg, 50-50 odds tell us that the bookies and their punters expect the final result in that seat to be close to even. More on this in a future post.

Here are the swings (to the ALP) expected according to today's sportingbet odds.





















(click on the table to see a bigger image)

Federal Election 2010: Analysis of Individual Seat Betting Aug 8th

Likely state of parties after the election according to the individual
seat odds of Sportingbet as of today.

Overall situation: around ALP 77, LNP 69 , Ind 3, Green 1 => ALP majority government

Details (according to current odds):

1)
ALP to lose Melbourne to the Greens
(Prob Green win =50.4%, Prob ALP =36.7%)

2)
ALP have 63 safe seats including Greenway (which was Lib before redistribution)
and Eden-Monaro.

3)
ALP look very likely to retain a further 7 of the following 9
(in which they have good odds of retaining):
Lindsay, Page, Bonner, Longman, Petrie, Bennelong, Dobell, Deakin, Brisbane
But they may lose 2.

4)
ALP should be able to win 4 of the following 6 seats as they are just ahead:
Forde, Bass, McEwen(a gain), Hasluck, Hughes(a gain), Corangamite
But they may not get 2 of these.

5)
ALP should be able to win at least one of the following 4 seats
where they are just behind, but they will most likely lose
the other 3:
Solomon(ALP), Macarthur(Lib), Herbert(Lib), Dickson(Lib)

6)
ALP may win 2 of the following 11 seats where they are behind
but have a significant probability of a win (i.e. it adds up over the 11
seats):
Bowman, Flynn(ALP), Swan, Ryan, Macquarie(ALP), Leichhardt(ALP), Paterson,
La Trobe, Canning, Robertson(ALP), Dawson(ALP)

7)
There are 53 safe L/NP seats.

8)
There are 3 safe independents.

Total 150 seats.

Federal Election 2010: Analysis of Individual Seat Betting Aug 3rd

Likely state of parties after the election according to the individual
seat odds of Sportingbet as of today.

Overall situation: around ALP 77, LNP 69 , Ind 3, Green 1 => ALP majority government

Details (according to current odds):

ALP to lose Melbourne to the Greens

ALP have 64 safe seats including Greenway (which was Lib befre redistribution).

ALP look very likely to retain a further 7:
Bonner, Page, Longman, Dobell, Bennelong, Deakin and Brisbane

ALP should be able to hold 3 of the following 4 seats as they are ahead:
Petrie, Forde, Bass, Hasluck

ALP should be able to hold at least one of the following 3 ALP held seats
where they are even, close or just behind, but they will most likely lose
the other 2:
Corangamite, Solomon, Leichhardt,

Other likely ALP losses are:
Macquarie, Flynn, Dawson, Robertson

ALP likely to gain about 2 of the following seats:
McEwen, Hughes, Herbert, Macarthur

ALP probably won't gain any of the following seats
although they are close:
Dickson, Bowman, Swan

There are 55 safe L/NP seats.

There are 3 safe independents.

Total 150 seats.