General Comments

I haven't posted in a few days. I will start again once the January polls start coming. Before Christmas the average spread tended to be about 6 points.

I haven't explained the format in a while, so here it is. Most of the time, I will post three lines for each poll. First is the raw poll results. Second is a seat projection. Third is a seat projection that incorporates the statistical probability of each party winning each seat. For regional polls the format is quite a bit different. First will be a table with the raw poll results. Next will be another table. It will first have the seat projection, then the projection with probability, then the results from the last election.

The reason for the two different projection methods is that doing the simple projection can lead to one party getting a disproportionate amount of the close seats (I'll call it the lucky effect). Since it is a projection it has variance, and the polls themselves have error margins, so incorporating probability will minimize the lucky effect. Of course, real elections have lucky effects (last election the Conservatives benefitted by five seats and the NDP paid the price), but to try to guess at this effect is a random shot.

Three weeks to go, welcome aboard.