i think i'm ready to make a final forecast. i'll let you know if this changes.
these are nik's numbers for yesterday:
conservative - 28.8
liberals - 28.7
ndp 17.3
greens - 8.6
bloc - 5.6
people's - 1.6
lost - 0.4
undecided - 9.0
over the last week, the liberals and conservatives have been more or less flat, while the undecideds have moved to the ndp.
i think this is a bradley effect. but, i suspect a lot of these voters are also a little younger and unlikely to actually vote. so, i'm going to to split the difference, and readjust:
conservative - 28.8
liberals - 28.7
ndp 15.0
greens - 8.6
bloc - 5.6
people's - 1.6
lost - 0.4
undecided - 11.3
this is where turnout matters, and i where i introduce jessica's theorem, aka the fundamental principle of canadian elections: the conservatives do terrible with undecided voters. they get maybe 20% of them. the greens also tend to underperform, and while it looked otherwise earlier, it doesn't look that way anymore. let's get them 5%. i'm also just going to give the bloc an extra 1% and the ppc an extra 0.475% and leave it at that.
that leaves 7% to distribute between the liberals and the ndp, and the results of the election will depend on how that works.
as mentioned, i think there is a bradley effect, but i'm actually going to point more to the liberal ground game. i've yet to see an election where it works out the other way, at least. so, i'm going to give the liberals 5% of this and the ndp the rest.
lib - 33.7 %
con - 31.1 %
ndp - 17.0 %
greens - 9.2%
bloc - 6.6 %
ppc - 2.1 %
those numbers are not that strange. but, you need to look regionally to understand the results.
the ndp and conservatives are outperforming their average in the west, and underperforming it in the east; the liberals, the opposite. so, i'm going to split the country in half. there are 104 seats in the west, and 231 seats in the east. there are also 3 seats in the north.
in the west, including bc, the conservatives seem likely to win by large margins, while the ndp and liberals fight it out for second place. in 2015, the totals in the west were as follows:
liberals - 29
conservatives - 54
ndp - 20
greens - 1
at current numbers, ignoring any bradley effect, the ndp may be able to keep their existing seats if the decrease in liberal support does not benefit the conservatives too much, and the increase in green support does not affect them too negatively. i think this is all too optimistic for the ndp. rather, we're going to take them down by a few seats. it does not, however, appear as though the greens are likely to make a breakthrough, after all. rather, i would expect the conservatives to gain a number of seats from the liberals in the west, and the outcome will look something like:
liberals - 15
conservatives - 70
ndp - 15
greens - 4
defeating the liberals will be pyrrhic for the anti-pipeline forces in bc, but the question really remains open: does the federal ndp even oppose the pipeline in the first place? the greens are your best bet here, in the long run. but, i'm not voting for them on monday, either - i get it.
in the east, the 2015 totals were rather different:
liberals - 152
conservatives - 45
ndp - 24
bloc - 10
greens - 0
at current numbers, the ndp are likely to lose all of their 16 seats in quebec (where they are down at least 15 points) and some, but not all, of their seats in ontario (where they are roughly flat). it would not appear as though they will gain seats in ontario, anyways; how many they lose would appear to rely on the strength of the presumed bradley effect. i see little evidence that they're going to pick up seats in the east. that will leave them with around 5 seats, at best.
the liberals are also roughly flat in ontario, but the conservatives are down dramatically, at the apparent benefit of the greens. this combination - ndp & liberals flat, conservatives down, greens up - is just about the best outcome that the liberals could possibly ask for. they may only steal a seat or two from the ndp in ontario, but they should steal 25+ from the conservatives, and in areas that are usually conservative locks, too.
expect something like this in ontario:
liberals - 110
conservatives - 6
ndp - 5
in 2015, quebec was a complicated four-way race that was difficult to call, resulting in split races with weird results. with the retreat of the ndp (mostly to the benefit of the bloc), you should see something more predictable in this election: the bloc should sweep the rural areas, while the liberals sweep montreal, gatineau and the south shore urban spaces and the conservatives nab a few seats in quebec city. that means you should expect the conservatives to lose a few seats, actually - and that the liberals may be able to hold steady, if they're lucky. but, the bloc will be the big winners here.
i do want to temper this slightly, though. i haven't seen a poll with the bloc above 30%, yet. they used to poll way, way higher than that in the duceppe years, and it does open up a few questions - is there enough lingering ndp support to help the liberals on the split? how may seats can they actually win with 30%?
i'm going to suggest the following in quebec:
liberals - 40
bloc - 30
conservatives - 8
ndp - 0
in 2015, the liberals won all 32 eastern seats, and ran over 50% in all four provinces. nobody predicted that. they have nowhere to go but down, but they are still leading the region by a good margin in all of the polling that i've seen. the greens may beat the ndp in this region, but they are taking support from the ndp itself, and i don't see any evidence that they'll actually win anywhere. that said, the east is likely to actually like elizabeth may for who she is, so some space for an upset should be accounted for.
that said, i think the reality is that the conservatives are not actually running much higher here than they were in 2015, and that the movement on the left is largely to be mostly inconsequential. it would be very unlikely for the liberals to sweep the region a second time, but nobody else is really stepping up, either.
something like the following is likely:
liberals - 27
conservatives - 5
ndp - 0
greens - 0
that adds up to a big win for the liberals in the east:
liberals - 177
bloc- 30
conservative - 19
ndp - 5
yeah. ontario...
there's three seats in the north, and they usually vote liberal.
so, here's my prediction:
liberals - 195
conservatives - 89
bloc - 30
ndp - 20
greens - 4
that would be a liberal majority.
what are some issues to draw attention to?
as mentioned, the greens are a wild card out east, but they're also a wild card in rural ontario. if they underperform in ontario, you could see the conservatives lose a few less seats; if they overperform, they could lose a few more. likewise, the greens could produce some confusing results in the east, but i'm not predicting it.
there's some possibility that the ndp could overperform these results in bc, as well, but we're talking a few seats, and i would consider that error.
in 2015, i tallied up 169 seats for the liberals - one seat short of a majority. so, my prediction was a strong liberal minority or a weak liberal majority (subject to error). most of the models actually predicted a harper minority, although some of the smarter kids clued in at the very end - but nobody that i know of predicted a liberal majority, except one forum poll released a few hours before the vote. the reason they got it right was because they picked up a last minute mass movement to the liberals in quebec, specifically. and, that was my error - i was suggesting a stronger showing by the bloc. i overestimated the bloc tally by 15 seats - the amount the liberals were short by. i was also dead-on with my conservative numbers.
if i'm wrong in 2019, it will likely be an overestimate by the liberals of roughly the same amount, and probably in ontario. but, 195 seats is actually a fairly comfortable majority, and i can eat that and still get the outcome right.
so, that's my prediction: a weak or comfortable liberal majority, but a liberal majority, nonetheless.
the liberals are still supposed to do better than this