Saturday, May 26, 2018

Ontario - NDP wins on the math

Today marks the end of my trend based projections as I transition to math based projections. Despite this, as I've done in every election, my gut plays a role in that math. Weighted at one third, the popular vote figures of 42.3% for the NDP, 32.1% for the Tories, and 21.7% for the Liberals, have been applied to the pure math and produced the following projection:


63 - NDP - 39.17%
49 - PC - 34.47%
12 -  LIB - 20.83%





I've also made additional corrections to the sheets, so if you wish to make your own projections, download a copy. In particular, "turnout" figures for each riding should now be correct as previously I was drawing those figures from the incorrect location.

I've also done research on the factor a Ford candidacy has on Etobicoke North, and have slightly bumped up the Ford numbers in the riding, turning it Blue. Note, however, that at this point, the figures have been bumped by more than half; more than I've ever bumped any riding in any projection, and it still remains a close race with the NDP.

As we've been expecting an NDP win for some time, the main 'story' from this is that the Liberals are holding their ground better than expected, and may walk away with a dozen or more seats at the end of the day.

Thursday, May 24, 2018

Ontario NDP's Ceiling

How high could the NDP go?

After re-running the math based on some new evidence the NDP may actually be able to break away with a massive lead (which I've represented as 47% vs 32%) which would enable them to win far more seats than Bob Rae ever did.

Here are the new maximums for the NDP:

90 - NDP
26 - PC
8 - LIB



Wednesday, May 23, 2018

Ontario NDP back in Majority territory after accounting for trends

The party is trending up, and in a manner that is not easy to reverse. A Recent Ipsos poll shows that 20% of people think the Liberals are most likely to win, while many, many polls, show that a lot of Liberals are quite open to voting NDP as their 2nd choice.

In short; we are aware that this is a PC-NDP race because you and I are political animals. Many voters are not aware, yet, that this is a PC-NDP race, and, when they become aware, will stop wanting to vote Liberal, and will start wanting to vote NDP.

This is the trend of the election. We've seen voters switching Liberal to NDP since the writ was dropped. A Mainstreet poll from early in the year had the PC Party at 43%, the Liberals at 32%, and the NDP at 18%. The most recent poll, as of the time of writing, from Pollara, has the NDP at 38%, the PC Party at 37%, and the Liberals at 18%.

You and I may be fully aware the NDP is in the lead, but many voters are not. People are not nearly as engaged as most of us hope they would be, even during an election.

This is what I am correcting for.

This is the kind of thing I mean when I say I am correcting for the trendline.

I am estimating what the popular vote will be once the electorate is "up to speed", as they tend to do as the election draws closer. It is my supposition that once they are fully "up to speed", and all the Anti-Ford and Anti-Wynne people realize the NDP is a safe bet, that the NDP will be at or around 41%, the PC Party at or around 37%, and the Liberals at or around 19%.

This would give them the following result:

65 - NDP
47 - PC
12 - LIB

This would result in an NDP Majority.





Neither Ford nor Wynne would be returned; however, there are caveats. I am assuming there will be an Anti-Wynne vote as many people personally hold her responsible for the failures of her government. This is similar to how Premiers Charest and Getty lost their seats at various points, despite the raw math suggesting they should not. As for Ford, I had to bump up the PC total to try to estimate his personal popularity. It is possible I under-estimated it, and, he would win even in this scenario. Unfortunately, without additional data (such as a riding poll) it is difficult to tell, and we are still far enough from the election that I don't feel a compelling need to delve into hours of polls to ensure this seat is correct; I have other seats to do that for and a limited amount of time to do that in. Be assured that by June 6th and my final projection, this and all seats will have had time spent looking at it individually.

Also note that I only make actual changes to the base math that I feel are significant. As a result, you'll notice that Jack MacLaren does not appear whatsoever in my spreadsheet. I consider his election so unlikely that the work to add him would be a waste. I have had voters in a riding the NDP won in 2014 with 52% of the vote, and where they are projected to win with a much larger share, complain that the party I have in second in that riding is incorrect. Frankly, this is not a difference I feel is significant enough to justify a change to the math.

I will, always, take a look at ridings that are close, and adjust them as needed. There are many reasons for adjustments. They can include but are not limited to:

  • A very popular candidate ran for a party last time, but the current candidate for that party is much lesser known.
  • A popular candidate is running this time, but did not run last time.
  • A candidate has managed to get themselves into a scandal of some sort.
  • A party has managed to get themselves (un)popular in the local area due to their (lack of) support for a particular policy

These are things I look for an try to take into account, but if the party is not able to win the particular riding in question, the impact on what people generally want to know most of all - who is going to win the riding - is minimal. As such, I do not put resources (IE time) into refining these aspects when I feel that my resources could be better spent elsewhere.

I hope this answers a lot of misconceptions that I've seen over this election campaign. Keep in mind that you can always make your own projections using the same math I do:

https://docs.google.com/spreadsheets/d/1WvTuHrZOiqcVBQ4VNWPMO03oqfPJCY1H9DGliUtsgL4/edit#gid=1072939327

Copy this sheet and change the numbers in the black coloured cells.

Remember to check back for any updates I may have made to the sheet in the interim so you can grab a fresh copy with the revised math.

Tuesday, May 22, 2018

Venezuela's "Election"

There's not much to say here beyond the election was clearly a sham.

It's difficult to say if more ballots were actually cast for Falcon or not, as the actual opposition candidate was not permitted to run, and Falcon was seen by some as a stooge of the regime.

What is clear is that the official results webpage, for some reason, will not load any results from states won by the opposition in 2013, but loads the states won by the regime in 2013 just fine.

As such its impossible to even tell what the "real" election results are in terms of ballots actually cast; much less the ballots that could have been cast if the opposition candidate was allowed to run.

Regardless, the country is so far from a democracy at this point that it's no longer worth covering seriously.

Sunday, May 20, 2018

Ontario, Maximums and Minimums

After additional feedback, I decided to share what I currently have for the maximums and minimums of each party in this election.


-----------------------

NDP

78 - NDP
40 - PC
6 - LIB





-----------------------

PC

85 - PC
31 - NDP
8 - LIB





-----------------------

LIB


48 - LIB
42 - PC
34 - NDP





-----------------------

There is no Green map as, at current, their maximum is likely 1, nor are there "minimum" maps, as, the effective minimum for each party is shown in the maximums for the other parties; IE around 40 for the Tories, and around 30 for the NDP, with the Libs minimum being a small handful of seats.


Ontario Update - Better math

I've overhauled the math I use to make projections and present a better projection of the coming election using the same basic rules as in my last post.

PC - 59 - 39%
NDP - 57 - 39%
LIB - 8 - 17%






Hopefully the ridings won will make more sense now that the math has been corrected

Saturday, May 19, 2018

Ontario update - (the lack of) movement

Due to reader feedback, I've decided to move back to a full trendline system; which means this is what I expect the end result to be on election day. As such, this result and projections by others will move closer and closer towards one another as election day approaches. I simply expect it will be they moving towards these numbers, and not the other way around.

As such, this is an update to this earlier projection.

In short, what has happened is the NDP has lost 4 seats.

Additionally, as I improve the base math I use to make these projections, some riding winners have been shuffled around.

60 - NDP
57 - PC
7 - LIB






There are a few things to take away.

1 - NDP not growing fast enough
Compared to what I expected, the NDP's growth has been a bit slow over the past three days.

2 - Tories are not falling away yet
In order for the NDP to win a majority they have to knock a few points off the PC Party, this is not yet happening

3 - Liberals still holding ground
They really need to be lower than what we are seeing for the full NDP win to emerge

4 - NDP vote is shifting
The NDP is unreasonably high in the GTA. I've yet to track down exactly why, but will continue to investigate. This contrasts with them being unreasonably low in Eastern Ontario.