Strategic voting won't help you

Strategic voting is a mirage. It's like gambling, or predicting the stock market -- there's a whole industry of people trying to figure out how to game the system, and most of them fail.

Progressives have been hand-wringing for the last decade about how to unite the left (without actually, you know, uniting all the left-leaning parties). We really really want to kick Harper out this time, and strategic voting seems like our last option in a system where the cards are stacked against us.

Remember when Joyce Murray ran for the Liberal leadership, proposing that progressive parties unite behind one candidate in key ridings? Nathan Cullen, vying for the NDP leadership, also pushed for cooperation to defeat the Conservatives. They both lost.

Now we have Trudeau and Mulcair at the helm of their respective parties, and neither is interested in electoral cooperation. That ship has sailed. A minority NDP or Liberal government may need support from the other party to pass legislation in parliament, but there's no way they'll be endorsing each others' candidates in this election race.

With that door shut, we've turned to strategic voting as the next best thing. But it's not the next best thing. It's not even a good thing. It's hardly even a thing.

Polling can't help you in tight ridings

The aggregate poll results at ThreeHundredEight.com, run by writer-journalist Éric Grenier, have been an excellent resource over the last three elections. The site dredges through the panoply of polls across the country to come up with an informed outlook of what could happen on election day. It's more than just an average of polls -- Grenier uses a projection model that gives a percentage likelihood of each riding's winner.

So, that's great news! We have a smart statistics-man that did all the hard work to tell us who we need to vote for to defeat Harper! Right?

Wrong.

Polls and projection models can be pretty accurate on a nation-wide or regional level, but they lose accuracy fast when you're trying to predict results for a particular riding. This is the Achilles' heel of strategic voting -- Canadians don't vote collectively as a country. We're carved up into 338 ridings, each a world unto itself. And when you get down to the riding level, there's so much uncertainty that it's difficult to figure out how to vote strategically.

According to the ThreeHundredEight projections, ridings with tight two- or three-way races have less confident predictions about who will win -- and these are the only ones that matter for strategic voting. In my riding of Kitchener-Centre, where the NDP, Liberals, and Conservatives are in a dead heat, the polls are fluctuating between a 50-60% chance of a Liberal victory. That's the statistical equivalent of throwing up your hands and shrugging your shoulders.

What about Leadnow?

This election, the advocacy group Leadnow has a shiny new campaign called VoteTogether. They hope to mobilize a hivemind of progressive voters in key ridings to tip the scales on voting day and stop Conservatives from getting elected.

I don't doubt that Leadnow has very smart people working for them, and they're trying their hardest to give people accurate information. But it didn't work in 2011, and Leadnow isn't doing anything substantially different this time around.

In 2011, there were two organizations trying to mobilize the strategic vote. They didn't do so well:

1. Project Democracy ended up making the wrong call in 19 swing ridings, were caught off guard by 10 ridings where the Conservatives gained a seat, and failed to make a dent in 33 targeted ridings where Conservative MPs were re-elected.
2. Catch-22, only endorsed 2 candidates that ended up defeating a Conservative incumbent. 34 of their picks came second, and 8 came third.

But maybe, just maybe, Leadnow has it figured out. The strategic voting movement for #elxn42 has coalesced around VoteTogether and they claim to have identified 72 Conservative swing ridings where a united left can keep the Conservatives out.

I wish them all the best, but the math of strategic voting is working against them. They say the only poll that matters is the one on election day -- and the only poll that matters for strategic voting is the one just before election day. Leadnow won't have any useful polling data until a couple weeks before the election. Hopefully, this gives their on-the-ground teams enough time to collectively choose which candidate to support, get the word out, and finally, cross their fingers in hopes that everyone obeys their pledge.

It's a tall order. For the strategy to work, the popular vote intention can't change much in the last stretch of the campaign. If we see another late surge of support, like the NDP enjoyed in 2011, it will be impossible to navigate all those shifting goalposts in each riding.

I don't want to diminish the real and valid mission behind a coordinated anything-but-Conservative effort. It's a fantastic display of solidarity against the dismal legacy of Harper's government. But the strategic vote is like herding cats, and has a dangerous potential to mislead.

Sam Nabi