i actually think it's mostly pretty clear, but take my word carefully - and check as much information as you can before you make a choice.

the general rules are...

- heavily unionized ridings (hamilton, windsor) are ndp locks. liberals are not competitive in rural ridings in most of the country, outside of southern ontario - that has to do with things like gun rights. the other exceptions are the northern prairies and out east. there are also a handful of legitimately solid mps (like paul dewar) in what should otherwise by liberal ridings that have earned their seat. if you're in a riding with a multiple term high profile ndp mp, consider the incumbency effect.

- liberals are the more competitive option in almost all non-unionized urban and suburban ridings, and in rural ridings in southern ontario. the exceptions are ndp-liberal races, and some western ridings. it should be obvious to most people in saskatchwan that the ndp is more competitive, and it should be obvious to most people in toronto that the liberals are more competitive.

the places where it gets confusing are in mixed ridings. that kitchener riding is a good example. kitchener has some union workers, but it's also somewhere between a big town and a small city, so it gets the rural/urban split vote. but, it's the exception, not the rule. most small town ridings are still too small town for this to work itself out. but, i'd keep an eye on growing cities like kelowna and lethbridge in the next election.

so, thankfully, i really don't think the split is as much a concern as it seemed like it was going to be, because the vote seems to be aligning rationally. i was freaking out a few weeks ago that the ndp were going to get a boost in ontario, and it would just elect conservatives. but, thankfully, that's not actually happening.

excluding the mixed ridings, the only other place where i think this is going to be a big issue is alberta. the conservatives are down. it seems like they're polling in the high 30s and low 40s in a lot of places. but, the vote is not coalescing around either the liberals or the ndp. it seems like long time conservative supporters are giving the liberals a sort of a break, here. it remains to be seen whether ndp supporters line up or not. if they do, we could finally see edmonton and calgary align with the cross-country urban consensus. and, i think that would not just be good for the parliament, but good for the country. the really key argument that both leadnow and the liberal party need to make right now is to line up behind the liberals in calgary and most of edmonton (excluding those two ridings where the ndp are ahead).

there's also eastern quebec, where bloc/ndp splits may elect conservatives. that should also be keyed on, but it probably won't affect the outcome of the election.

http://www.theglobeandmail.com/news/politics/elections/the-big-experiment-of-voting-strategically-this-election/article26767617/

Formerlyrbfromcalgary
Outrageously funny misreading of Alberta data, I hope your incompetence is reflected in other areas.

There are precisely 6 out of 34 Alberta ridings, and that would be 6 out of 19 urban Alberta ridings where non-Tories have a good chance of winning.

Percentage of vote for the Tories, province wide AND in totalled in urban ridings will be well over 50%

The NDP support is clustered in 2 Edmonton ridings: elsewhere they are almost non existent.

Of the 28 Alberta ridings with the Tories polling in the lead, there are a total of 6 of those ridings where 100 % of the combined vote for the Libbies and NDP would exceed the Tories.

Given the massive buyers remorse at the provincial NDP victory, it would be fair to assume that any strategic voting in Alberta would have a quite different target - Mulcair.

A "big issue" in Alberta?

Too funny

deathtokoalas
i'm going to reiterate that if you look at the data more closely, you'll see that the conservatives are down a good ten points, at least. you can exaggerate that a little in the urban ridings. so, that takes them down to the high 30s and low 40s. strategic voting could pull off some upsets, if they're targeted well.

if i was the liberal party, i'd have trudeau on a plane to calgary. it's his best fight right now. let toronto's local candidates keep up the fight in the 905.

and, frankly, i think the idea of trudeau showing up and campaigning in harper's riding is excellent theatre across the country.

2011 results - 66%.
cbc poll average - 53%

Formerlyrbfromcalgary
You're numbers vis a vis Alberta are way out. Check out ThreeHundredEight.

Their history is actually pretty good. In the few ridings where Tory support is in the thirties, those ridings can be expected to go Liberal or NDP already.

Elsewhere, not even close for the vast majority.

Absolutely no groundswell for Libbies or dippers breaking through meaningfully
in Alberta.

In fact, a reasonable chance that NDP percentage of vote will be lower than last election

deathtokoalas
i have a background in math, and trust me when i say that those models are making a lot of guesses. they use a proportional model, and they argue that it's better across the board. i think the better argument is that you need to use a mixed model, based on what you understand about the specific cases. in alberta, you don't expect rural ridings to swing much, so you want to avoid the kind of proportionality they're using, which is going to underestimate the swing by distributing it province wide. you want the opposite effect. so, if they're down 10-15 points province wide, then they're down 20-25 points in the cities, to balance out for stasis in the rural regions. that means 55-65 falls down from 30-35, to 40-45.

but, you're not really getting what i'm saying - what i'm saying is that if strategic voting groups get behind the liberals (who are clearly polling higher in calgary and most of edmonton), then getting ndp voters to support the liberals could create some serious upsets.

the models haven't been very good with the last few alberta elections, and i think this is the reason why - they take the province wide numbers and proportionally decrease them. so, 66--->53 would be a 20% decrease. they would then conclude that they're down 20% across the province. so, if they got 80% in some rural riding in 2011, they would cut them down 20%, which is 16 points, to 64. but that's likely inaccurate. they're probably still running at 80% in those rural ridings.

conversely, if they won a riding with 55%, they would take that down 20%, which is 11 points, to 44. but considering that they're running down 20%, and they're probably not down at all in the rural ridings, what you want to actually do is amplify that so that the total gets to 20%. it turns out calgary and edmonton are about half of the population of alberta. so, what you want to actually do is double it 40%.

then 55% becomes 33%, and 65% becomes 43%.

(sorry. the 55-65 is urban. to be clear.)

but don't listen to me. nobody ever does. just wait for the result to prove me right, and watch edmontonians and calgarians bang their head against the wall for splitting the vote.