On strategic failures

Following up on this post, let's take a closer look at how the recommendations of the strategic voting sites which claimed to offer the only chance to stop Stephen Harper matched up with the list of ridings being pointed to as having given the Harper Cons their majority - looking both at whether the riding was correctly identified as one likely to affect the election outcome, and whether the strategic voting sites actually recommended that voters support the party with the best chance to defeat the Cons.

Riding	Identified	Correct Opponent
Labrador	0	0
Nipissing-Timiskaming	1 (PD)	1 (PD)
Bramalea-Gore-Malton	2 (CAW, PD)	0
Etobicoke Centre	0	0
Saskatoon-Rosetown-Biggar	3	3
Elmwood-Transcona	1 (C22)	1 (C22)
Montmagny-L'islet-Kamouraska-Riveire-du-Loup	2 (C22, PD)	0
Lotbiniere-Chutes-de-la-Chaudiere	1 (PD)	1 (PD)
Don Valley West	3	3
Mississauga East-Cooksville	0	0
Winnipeg South Centre	2 (CAW, PD)	2 (CAW, PD)
Yukon	0	0
Desnethe-Missinippi-Churchill River	0	0
Palliser	3	3
Total	18	14

So what can we tell from the chart?

To start with, strategic voting sites did an absolutely woeful job of identifying the ridings which proved decisive. Given free rein to identify as many seats as they saw as justifying an endorsement, the sites had 40 total chances to identify which ridings could actually stop a Harper majority (allowing for the fact that the CAW seems to have avoided endorsing candidates in Quebec). And their success rate in doing so was less than 50%.

What's more, the fact that the sites recognized that a riding mattered didn't guarantee that they'd get it right, as two ridings saw the NDP emerge as the leading challenger to the Cons contrary to "strategic" recommendations. Which leaves the strategic voting sites with a total success rate of a measly 35% in making the correct recommendation for a riding which proved decisive (to go with an epic fail rate of 10% where they actually endorsed the wrong party).

In fact, as the election turned out, a simple "vote NDP" strategy would have resulted in the correct vote in more of the ultimate swing ridings than following the advice of strategic voting sites. (And indeed the same would have applied to a simple "vote Lib" strategy.)

Now, none of the above is to doubt the sincerity of those who put together the recommendations in the first place. But there should be less doubt than ever that the more important work is to be done in building a broader electoral movement, not in trying to game the electoral system.

[Edit: fixed wording.]