Double Blind

Based off of the preliminary results of the election (I will update the numbers of this post if the results change), my absolute error in seat count was 62.

Party – Actual – Predicted – Error
Conservatives – 167 – 160 – 7
Liberal – 34 – 39 – 5
NDP – 102 – 79 – 23
Bloc – 4 – 28 – 24
Ind – 0 – 2 – 2
Green – 1 – 0 – 1

Provincial Split – Actual vs. Predicted (CON/Lib/NDP/BQ/Ind)
BC: 23/1/12/0/0 – 21/2/12/0/0 and one GRN
Alberta: 26/0/1/0/1 – 27/0/1/0/0
Saskatchewan: 13/1/0/0/0 – 13/1/0/0/0
Manitoba: 10/0/4/0/0 – 11/1/2/0/0
Ontario: 70/16/20/0/0 – 73/11/22/0/0
Quebec: 5/9/32/28/1 – 6/7/58/4/0/0
New Brunswick: 6/2/2/0/0 – 8/1/1/0/0
Nova Scotia: 3/3/5/0/0 – 4/4/3/0/0
PEI: 1/3/0/0/0 – 1/3/0/0/0
Newfoundland and Labrador: 2/3/2/0/0 – 1/4/2/0/0
Territories: 1/1/1/0/0 – 2/0/1/0/0

Surprises/Analysis this election (only two):
1. The collapse of the Bloc Quebecois, mostly at the expense of the NDP. Conversely, the holding up of the NDP vote in Quebec, which I expected to have been diluted a few percentage points from what the pollsters predicted. Even if the NDP and Bloc vote swung 3% (in opposite directions, so a net effect of 6%), it only would have impacted the result of 2 seats that the Bloc were trailing within 6% or less from the NDP (Montmorency–Charlevoix–Haute-Côte-Nord, Gaspésie–Îles-de-la-Madeleine) but for the rest of the seats the NDP victories were crushing. This was the biggest mistake that my prediction model made, by far.

2. Elizabeth May. She has absolutely proven me wrong when I incessantly harped about how that along as she is leader of the Green party that the party will never win an elected seat in the House of Commons. She won with 46% of the vote against Conservative Gary Lunn, who lost with 36% of the vote. I would be highly curious to know what happened with the local campaign there during the election. Although I disagree with mostly everything that comes out of May’s mouth, she has to be congratulated for her victory in Saanich–Gulf Islands.

Other pundits out there:
Actual result – 167/34/102/4/0/1
ThreeHundredEight.com (Eric Grenier) – 143/60/78/27/0/1, error: 24/26/24/23/1/1 = 99
DemocraticSpace.com (Greg Morrow) – 155/47/86/20/0/1, error: 12/13/16/16/1/1 = 59
ElectionPrediction.org (Milton Chan) – 146/63/65/33/0/1, error: 21/29/37/29/1/1 = 118
LangleyPolitics.com (Jordan Bateman) – 136/52/79/39/0/2, error: 31/18/23/35/1/2 = 110

I wonder if anybody else has done a comprehensive compilation of pundits out there.

I managed two campaigns in Richmond, in 2008 and now 2011:

Year – Richmond Result / BC Conservative Vote / Outperformance
2006 – 38.7% / 37.3%
2008 – 49.8% / 44.4% / +4.0%
2011 – 58.3% / 45.5% / +7.4%

(Outperformance is measured by the difference between the local result and the provincial vote average)

Our campaign turned in a performance that was the third best in the lower mainland (behind Abbotsford – 64.8%, Langley – 64.1%) and in relative outperformance, we were the best in BC. When you consider our geographical proximity, we absolutely did fantastic. Our campaign team, with our candidate, supporters and volunteers is the best in Canada.

There will be a couple days to relax and dismantle the campaign office, but after this we will continue the work we have been doing in Richmond.

Here is my election prediction. I have tried to do this with as much of a non-partisan analysis as possible, but keep in mind I have been neck-deep in the campaign in Richmond – I’ve only had today to seriously look at the publicly available data (there is no insider information I am basing this on). This has also been a very strange campaign.

Conservative – 160
Liberal – 39
NDP – 79
Bloc – 28
Independent – 2
Greens – ZERO (Elizabeth May, 25%)

Provincial Split (CON/Lib/NDP/Bloc/Ind):
BC: 23/1/12/0/0
Alberta: 26/0/1/0/1
Saskatchewan: 13/1/0/0/0
Manitoba: 10/0/4/0/0
Ontario: 70/16/20/0/0
Quebec: 5/9/32/28/1
New Brunswick: 6/2/2/0/0
Nova Scotia: 3/3/5/0/0
PEI: 1/3/0/0/0
Newfoundland and Labrador: 2/3/2/0/0
Territories: 1/1/1/0/0

Notes: The NDP is suffering from a problem of vote diffusion, similar to the Green Party. This should disproportionately favour the Bloc. Note that all it took for the Bloc to consistently take 50 seats out of play in Quebec was 40% of the vote.

The Liberals have concentration in Toronto, Montreal and pockets of Atlantic Canada. The notable change this election should be the area surrounding Toronto – the Vaughn by-election was a precursor for this.

Ontario remains the key battleground, which reminds me of a reduced version of the 1997 election when the PCs and Reform combined took 38% of the vote, but the Liberals took 101/103 seats with 49.5% of the vote.

In BC, the Liberals continue to get squeezed by the Conservatives on the right and the NDP on the left. There is not much room left in the spectrum for the Liberals anymore.

Their April 29th poll declaring a 46% Conservative victory is brave. This poll is the classic case of an outlier. Assuming the averages of the other polling agencies (approximately 37%) is the actual result, the 9% variance would have meant that about 82 of the 750 people interviewed were more biased toward the Conservatives than the baseline number.

My election prediction will be coming up tonight.