I’m a geek. I fully admit it.
Years and years ago, I developed a secret formula for working out seat counts for the upcoming federal election, based on regional breakdown of polling data.
What that means is that I have a prediction program which lets me plug in polling information, it applies the changes in voter support, and spits out a seat prediction for the upcoming election.
Seat predictions are very difficult to do, because regional vote concentrations means that regional vote shifts do not necessary apply at a local level…
However, I’ve been fairly accurate in the past, even if it’s more by luck than judgement.
So, this time around, I’ve done it again – I’ve plugged in the numbers based on the most recent poll released by Ipsos Reid, applied the shifts in support to the results of the 2006 election, and come up with the following:
Conservative: 142 Liberal: 98 NDP: 34 Bloc Quebecois: 34 Green: 0
Majority threshold is 155 seats.
Despite the gaffes, it looks like Harper and the Conservatives are trending in that direction; however anything can happen. The only poll that matters is on October 14.
Don’t ask me to break the results down any further than at the national level, because there really is no point. I’ve tried it before, and when you compare it to the final result, while my seat counts are fairly close, the provincial and riding-by-riding results are extremely … “wacky.”