it's an interesting thought. but, by some estimates, if even a quarter of the undecideds vote as a block, it could swing the election. you have quite a bit of room to move on this hypothesis in being absolutely right, and yet being absolutely wrong. for that reason, there's almost no way to write off the undecideds - even if they overwhelmingly abstain, the few that don't are still the key in who wins, should numbers stay static.

if you look at how the 30/30/30 split has developed since august, the conservatives have more or less been flat while the ndp has lost a few points to the liberals. all of the data suggests almost all of the swing is on the left. i would expect the undecideds to eventually push one of those two parties up by a few points.

depending on where it happens, it could be disastrous for harper or it could save his hide. if the liberals get the bump, and it's largely in ontario, it could push the conservatives into third. but if the bump the liberals get is mostly in quebec and tepid in ontario, then you end up with equally sized opposition parties in a harper minority. conversely, if the ndp get the bump and it's mostly in ontario, you end up with conservatives winning tons of splits; if they get the bump and it's in bc, it could win them a small minority.

right now, it seems to be the liberals that are trending. but, even so, their gains may come at the expense of allowing mulcair a chance at governance, as they push tory seat totals in ontario down, but not enough to get over ndp dominance in quebec.

the point is, the undecideds are still huge - even if the vast majority stay home, in the end.

www.cbc.ca/news/politics/canada-election-2015-grenier-undecideds-sep21-1.3237086