Projection Update: Nanos 10/3-5, Mainstreet's 9/29 Longueuil--Saint-Hubert Poll

The following polls have been added to the model:
Nanos' 10/3-5 national poll (national results only; current weight among national polls: 41%)
Mainstreet's 9/29 Longueuil--Saint-Hubert poll
For a full list of included polls, see previous Projection posts.

The Nanos tracker shows marginal gains for the Liberals, while the Tories are stable. The NDP and Greens pull back a bit from recent gains and losses respectively. The Bloc loses 3 points in QC, providing more evidence (but far from conclusive proof) that Blanchet's good debate performance might not have done much more than solidify the Bloc's existing vote.

The Longueuil--Saint-Hubert poll resulted in an adjustment that, despite its very small size, flipped Longueuil--Saint-Hubert to the Liberals. The Bloc could well get it back, though.

Because the midpoint of this poll is October 4, this update includes a baseline adjustment for the Conservatives disavowing their candidate in Burnaby North--Seymour on that day. Because the candidate will remain on the ballot (likely as a Conservative, unless for some reason the ballots haven't yet been printed), she is likely to still receive a non-negligible number of votes, though of course her chances of winning are now extremely low. This change costs the Tories ~0.4 seats in the projection, and increases the Liberal and NDP counts by ~0.2 seats each.

This projection is to be taken with (even greater than usual) caution for the two following reasons:

1. Now that we're into October, the model is discounting older polls very aggressively. That's a problem currently: since the Mainstreet tracker wasn't publicly updated yesterday, Nanos is now 2 days more recent than any other poll (and 3 compared to any non-tracker poll). As a result, it gets a very high weight despite its relatively small sample, and the model's strong shift to the Liberals - who are back in pre-"blackface" territory - may be an overreaction. Nevertheless, if what Nanos caught was real, the model will shift even more when other pollsters concur with the movement.

2. There has been no public regional breakdown since the TVA debate. Therefore, it is quite possible that the Tory drop is concentrated in QC, in which case its QC vote projection would be too high, and its projection elsewhere would be too low.

Projection as of the latest national poll (midpoint: October 4)
LIB - 159.6 (34.8%)
CON - 136.5 (35.4%)
NDP - 20.0 (13.4%)
BQ - 18.3 (5.2%)
GRN - 2.6 (8.0%)

PPC - 0.5
IND - 0.5
If you're new to my 2019 projections, view key interpretive information here.

Seats changing hands since the last projection:
- In QC, the PPC retakes Beauce from CONs.
- In QC, CONs retake Drummond from the BQ.
- In QC, LIBs retake Longueuil--Saint-Hubert, Hochelaga and Saint-Jean from the BQ.
- In ON, LIBs get Richmond Hill and Oakville North--Burlington back from CONs.
- In BC, LIBs retake Vancouver Granville from Wilson-Raybould.
- In BC, LIBs get Steveston--Richmond East back from CONs.
- In BC, CONs regain Courtenay--Alberni from the NDP.