that goofy poll is not so strange if you look at the undecideds. it may suggest that the liberal-ndp swing just imploded. that is frustrating, but not entirely surprising.

let's be clear: stephen harper is irrelevant in the election. stephen harper has been irrelevant in every election he's run in. he'll get around 30% - higher if turnout is low. he won't swing liberals. he won't swing dippers. his numbers are stagnant, have been stagnant since he started and will remain that way until he finally goes away. he is not a popular leader. his policies are not popular. when he wins elections, as he has, it has had very little to do with him. despite him winning several elections, canada has never voted explicitly for harper. canada is not going to vote explicitly for harper. this is for the simple reason that canada does not agree with much of what harper has done, has to say or promises to do.

rather, there's an epic struggle on the left that neither party seems able to win. once that resolves itself, harper will be instantly destroyed. nary a tear shall be shed.

something that's consistent is that harper has no potential amongst undecideds - if you're going to vote for him, you've decided to already. so, when you see a poll that has tories 30%, undecideds 20%, ndp 22%, liberals 20% (or whatever it was), you can instantly translate that to: ndp 34, tories 30, liberals 28%. and, then that's every other poll.

a little bit of sneaky bs from a conservative pr firm. it won't be the last time.

but, if there's something to this idea that the undecideds are really that large, then there's reason to worry on the left.

i've said from the start: there are two possible outcomes of this election. if the left splits (as such as an undecided vote might suggest), then harper wins a sweeping majority with 33% of the vote, or whatever it is. all he needs to be is the most popular in the riding. and if the ndp and liberals split their 60% majority 30-30 across the board, he needs to average 31% to win.

unfortunately, he can do that. because that 30% is the conservative floor - outside quebec.

if one of the parties on the left pulls ahead, they sweep. the liberals basically can't do this, they can only contribute to the splitting. but if the ndp get that boost to 35-36%, then they get over the floor enough to sweep.

there will not be a minority government, as some conservative-friendly models are predicting.

nor will the uccb swing the election.

===

the tories are going to reduce turnout as much as possible, and push their own base. they have no appeal outside their base. they have to try and stop their opponents from voting.

the ndp are going to want to swing a small - but uniform - percentage point or two from the liberals. if they can stabilize at 35, they have a good chance of winning a majority.

the liberals do not have a strategy, besides hoping that the conservative base suddenly dies of old age. they would need an unrealistically large uniform swing from the ndp - upwards of ten points to compensate for their inability to compete in quebec.