Kiwi Poll Guy calculates the probabilities
One of the major goals of this site is to try and predict election results based on recent relevant political polling. This is intended to include not just the total number of seats won by each party, but also viable coalition possibilities and electorate level results.
Today I present the simulation results at the candidate level, including probabilities for each major or minor party candidate to be elected to parliament by either winning an electorate or being selected off their party list.
So what are the odds he calculates? Note that these are based purely on probability calculations based on poll data, and don’t take into account local circumstances such as popularity of a local MP. And the probabilities below are of them holding their electorate seats – not whether or not they would also come in on the list.
His probabilities are:
100% chance of retaining seats
- National – Key, English, Brownlee, Power, Smith N, Ryall, Collins, Williamson, Tolley, McCully, Smith L, Mapp, Guy, Tisch, Wong, Carter J, Heatley, Hutchison, Ardern, Goudie, Roy, Coleman, Tremain, Borrows, Foss, Peachey, Goodhew, Dean, Auchinvole, Bennett D, King, Hayes, Bridges, Upston, McClay, Macindoe, Kaye
- Labour – King, Sio, Shearer, Hawkins
- Maori – Turia, Sharples, Harawira, Flavell
90% to 99% chance of retaining seats
- National – Bennett P 98%, Young 96%
- Labour – Goff 95%, Carter C 97%, Dalziel 97%, Anderton 98%, Robertson R 97%, Curran 93%
- ACT – Hide 95%
50% to 89% chance of retaining seats
- National – Lotu-Iiga 87%,
- Labour – Horomira 76%, Hodgson 86%, Cunliffe 65%, Mahuta 62%, Laban 74%,
- United Future – Dunne 73%
- Maori – Katene 60%
Under 50% chance of retaining seats
- Labour – Dyson 19%, Mallard 37%, Cosgrove 4%, Hipkins 24%, Burns 25%, Robertson G 9%, Lees-Galloway 10%,
Now as I said, this is a straight probability calculation based on current polls showing National more than 20% ahead of Labour on the party vote, and applying that to the electorate vote. Kiwi Poll Guy also works out the probability of each MP and Candidate coming in as a List MP, based on teh 2008 lists.
The probabilities are not predictions. For example I would give Grant Robertson a far higher chance than 9% of retaining his seat – but that is based not on a uniform swing to National, but local knowledge of the Wellington Central seat.
As we get closer to the election, and we also get electorate polling data into the mix, I’ll be checking Kiwi Poll Guy’s numbers more and more often.
February 9th, 2010 at 11:25 am
Interesting precentage for Dunne at 73%. If we take even a 10% swing to National in Ohariu Dunne is gone and the seat goes blue.
Would be great to know how he is polling now as he was well behind before Nov 08 until he jumped ship as a Minster of the Larbour mob and joined Blue to save his backside.
Probably better for him to bugger off to Canada and save face.
February 9th, 2010 at 11:42 am
In politics there is no such thing as a 100% certainty.
Gravity is a 100% certainty, reelection isn’t
February 9th, 2010 at 12:46 pm
How interesting that he rates Ms Bennett’s likelihood of retaining her seat as higher than Mr Goff’s chances.
February 9th, 2010 at 2:16 pm
Oh yes, yes, music to my eyes. Not a single National MP at risk of not being re-elected. Fucking suffer, Labour drones.
February 9th, 2010 at 2:45 pm
Murray,
The 100% probability is calculated from the simulation, which as DPF points out is based on a model that assumes a uniform swing in each electorate proportional to the swing towards National in the polls relative to the last election. Under this model any electorate that went National in either 2005 or 2008 is predicted to go to National with 100% probability because the polling swing in favour of National since either of the last two elections is much larger than the margin of error on the polling.
Whether or not the model itself is accurate is a different question. DPF obviously doesn’t think so, and I’m also quite happy to agree. The model is nothing more than a stop-gap measure until we get some electorate level polling in the lead up to the next election.
February 9th, 2010 at 2:56 pm
Tim,
The reason is that that is how the model determines the electorate winners. Please refer to my reply to Murray above.
At the last election Goff won Mt. Roskill with a margin of 19% points (55% vs. 36%). Bennett won Waitakere by a margin of 2% points (44% vs. 42%). Assuming a uniform swing since the last election towards National of about 11% points then the model would see Bennett as safer.
Again though, the model itself is just a stop-gap measure. Personally I wouldn’t think Goff’s seat would be in any danger.
February 9th, 2010 at 3:26 pm
I’d force of a default to 99% myself. It might be mathmatically accurate but that shows its not real world based. It doesn’t allow for things like having a heart attack and dying the day before the election which can slow your chances down.
Not of voting if you’re a democrat in Ohio maybe, but certainly as a candidate.