Prediction Thread - Ontario Election 2014

NorthReport wrote:

Rise above, if you are actually interested in learning about pollsters and aggregators try spending some time reading what Nate Silver, of New York Times fame, and America's most accurate aggregator has discoverd with his research. it's a fascinating journey of discovery into the the dirty world of politics. 

You do realize that sounds somewhat patronizing, yes?  You make it sound as if you know more than I do and that I'm some novice to politics who just learned about it this week.

I read Nate Silver all throughout the 2012 election of Obama vs. Romney, as well as all his Senate race projections.  He was right on nearly all of them.  I know a lot about his personal background as well, including his religious and sexual orientation, as well as which neighbourhood in New York he lives in.

Btw, Silver inputed almost every single poll into his model - he assigned each a different weighing value, even those which lean a particular way.  For example, although Rasmussen is pro-Republican, he still included them in his projections.

As of now:

Liberals           46

Conservatives  40

NDP                 21

Josh, your predictions are within 1 or 2 seats of Eric Grenier's current seat projection, so if the polling model at 308 is right, you could be right on the money.

Ciabatta makes some interesting predictions.

> I'm not sure the Liberals are going to take away all 3 of those NDP seats - they may not even take any of them.  As always, it depends upon the splits and the peculiarities of our electoral system.  I'm hoping the Liberals win Trinity-Spadina because it could mean the federal by-election campaign is doing well there, too.  Not sure if the Liberals will win back Davenport, and Parkdale-High Park, while possible in theory, is probably not on the table in this election.

> Etobicoke-Lakeshore has another re-match between the 2 titans again, but while Professor Barry Kay gives the Liberals the edge, I'm not so sure it will revert back yet.

> I agree that the Liberals will probably lose Sudbury (although it's not totally gone since Wynne was there this week).

> The Liberals could certainly lose Kitchener Centre, Brampton-Springdale & Etobicoke Centre, but I don't see any danger yet in Brampton West, or London North Centre.

> Also no imminent danger for the Liberals in Ajax-Pickering, Brant or Oak Ridges.

> Ottawa-Orleans was in danger of going PC a couple weeks ago, but since I'm working on the Eastern Ontario campaign, I can tell you that it's now back in play for the Liberals.  So is St. Catharines.

> I don't see the NDP losing Bramalea-Gore-Malton yet.  Or Parkdale High-Park.  As I said above, the Liberals are targetting Trinity-Spadina and Han Dong is getting a boost from Adam Vaughan & Justin Trudeau because of the by-election.  But Marchese has narrowly hung on in the past, and he may do so again.  Davenport could also stay NDP.

> According to 308 & others, it does appear as if Niagara Falls & Kitchener-Waterloo could go PC.  Sudbury will likely go NDP, as mentioned above.  It's the weakest of the Liberal seats in Northern Ontario.

My prediction:

NDP: 104

Socialist Party of Ontario: 2

Green: 1

Liberals and PCs: 0

I can dream, can't I?

Debater wrote:

> The Liberals could certainly lose Kitchener Centre, Brampton-Springdale & Etobicoke Centre, but I don't see any danger yet in Brampton West, or London North Centre.

> Also no imminent danger for the Liberals in Ajax-Pickering, Brant or Oak Ridges.

Agree with London - I thought I had taken London North Centre out of the Liberal loss column but seems I had forgotten.  (I've since edited above.)  I think NDP strength in St Catherines and Brant (where they're giving it a real good go, as opposed to Cambridge where they are invisible) and the NDP's Brampton-wide push is a problem for the Liberals.  Oak Ridges is a frankenstein of a riding and Ajax-Pickering is a toss up.

Aristotleded24 wrote:

Electionprediction.org now predicts a Liberal victory in Davenport, while Parkdale-High Park and London West have moved into the "too close to call" category.

They do this every election with some more marginal NDP seats, and every election they are almost always wrong.

Election Prediction is basically just a matter of opinion and is more for fun, so I don't take it too seriously myself, and it does lean-Liberal.

But the projections at 308 are of more significance because they're based on actual numbers by all the pollsters.

Btw, as Eric Grenier and Bryan Breguet point out, it's hard to know how much stock to put into the Ipsos likely voters model because each pollster uses a different one, and apparently the Ipsos panelists are not random, and may have been paid $$ for their contributions.  Raises some potential red flags.

Here are the Ipsos Regional Breakdowns:

Toronto: L 44, PC 25, NDP 22.

In the 905: L 38, PC 34, NDP 21.

In SouthWest: PC 38, L 28, NDP 26.

Central: PC 46, L 34, NDP 15

Eastern: PC 56. L 25. NDP 14.

Northern: NDP 45. L 29. PC 22.

----

The Eastern Ontario number for the PC's seems a little high & rather at odds with other pollsters, but this is Ipsos, afterall.   Seems like the province is divided between the Liberals & PC's, with the NDP only leading in Northern Ontario.  And if the Liberals really do have twice the support of the NDP in the 416 and the NDP is down in 3rd there, perhaps it is possible the NDP could lose a couple seats in Toronto, afterall.

But tomorrow is another day, with no doubt another poll. 

If you can forget about the phony "prediction" aspect and can sift through the partisan trolling, there is some good local insight you get form the odd post on electionprediction.com that no one is reflecting in the media or on other websites.  It should be localelectiondynamics.org or something, but we all know that's no as compelling.

Election atlas has the NDP holding their current seats because it hasn't taken into account the latest polls.  I don't see anything but NDP losses (overall) as plausible.

I think Grenier's is the most likely to end up accurate, although let's be honest looking at general polling trends and pointing out each party's weak ridings is not rocket science if you pay attention to this sort of stuff.

Debater wrote:

Election Prediction is basically just a matter of opinion and is more for fun, so I don't take it too seriously myself, and it does lean-Liberal.

But the projections at 308 are of more significance because they're based on actual numbers by all the pollsters.

Btw, as Eric Grenier and Bryan Breguet point out, it's hard to know how much stock to put into the Ipsos likely voters model because each pollster uses a different one, and apparently the Ipsos panelists are not random, and may have been paid $$ for their contributions.  Raises some potential red flags.

Brequet's projection site:

http://blogues.journaldemontreal.com/tooclosetocall/

Right now, he has the Liberals with 51 seats, while Grenier has the Conservatives with 51 seats.

I hope you're right, although I would rather see those 22 Liberal seats go to the Greens - Mike Schreiner strikes me as an awesome leader and the Ontario Liberals don't deserve a single seat after their decade of debauchery.

scott16 wrote:

my new prediction

50 NDP

35 PC

22 Lib

Hopefully Andrea reaches out to the Libs new leader and forms a steady coalition.

josh wrote:

Debater wrote:

Election Prediction is basically just a matter of opinion and is more for fun, so I don't take it too seriously myself, and it does lean-Liberal.

But the projections at 308 are of more significance because they're based on actual numbers by all the pollsters.

Btw, as Eric Grenier and Bryan Breguet point out, it's hard to know how much stock to put into the Ipsos likely voters model because each pollster uses a different one, and apparently the Ipsos panelists are not random, and may have been paid $$ for their contributions.  Raises some potential red flags.

Brequet's projection site:

http://blogues.journaldemontreal.com/tooclosetocall/

Right now, he has the Liberals with 51 seats, while Grenier has the Conservatives with 51 seats.

Grenier had the Liberals with 49 seats yesterday, the reason the Conservatives are at 51 seats today is because of the Ipsos poll.

I'm sure the next poll will change things again in this crazy polling election. 

Jacob Two-Two wrote:

Debater wrote:

Winston, that is ridiculous - 'decade of debauchery'?  The Ontario Liberals have been far better than the decade we had under the Harris/Eves PC's.  Would you prefer to go back to Version 2 of Mike Harris with Tim Hudak?

You're basically making a case for voting NDP.

I don't see how you read that into what I wrote above.

The Ontario Liberals have done a far better job of running Ontario than the cut & slash everything in site that Mike Harris did ("you're gonna see one mean, mad Mike Harris").

No pollster or seat projection model shows the NDP anywhere close to 1st place right now, let alone 2nd, so it cannot beat the PC's.  Kathleen Wynne is a far more sensible and progressive leader than Hudak, and perhaps more so than Horwath too, according to Judy Rebick & others.

josh wrote:

Brequet also included Ipsos, but he didn't give anywhere as much weight to Ipsos's "till death do us part" voter sample.

Yeah, well part of the problem is that you have to put a lot of trust in Ipsos's voter sample, and ignore the fact that they tend to be a Conservative pollster that is run by Conservatives (Darrell Bricker & John Wright) and their history of inflating Conservative numbers, particularly at the Federal level going all the way back to the 2000 Election where they predicted a week before Jéan Chrétien's 3rd Majority that he would probably only win a Minority.  They've been playing these games for years.  They had Harper at about 42% going into the 2011 Election and he didn't even end up hitting 40%.

The latest Ipsos poll also shows massive numbers for the Cons (56%) in Eastern Ontario, which is my region of the province that I know well, and it's totally at odds with the other pollsters.  It also has them winning seats by larger margins than their Federal cousins despite the fact that the Provincial Liberals are in stronger shape than the Federal Liberals were under Ignatieff.

North Report, as usual, is trumpeting the Ipsos numbers above and is saying we should only look at the 'motivated voters' sample, and ignore everything else.  Ipsos looks at the world from a Conservative point of view, so one has to keep this in mind.

Most models show the NDP as being weaker in York South-Weston this year than they were in 2011, but who knows.  Perpetual NDP candidate Paul Ferreria is running there again for the 500th time hoping that the wind goes his way one of these times.

But seriously, I don't see the NDP losing all those other seats.  Why would they lose ridings they currently hold & have been strong in during previous elections like Trinity-Spadina, Parkdale-High Park & Davenport if they're capable of winning back York South-Weston?

Jacob Two-Two wrote:

In fact, the NDP could easily come first if the Liberal vote collapses, and polls indicate that it is quite soft. Is it likely? Can't say, but it is absolutely possible. If you cared at all about the province of Ontario you would be pushing for that scenario instead of trying to reward the Liberals for being just barely better than the party that you claim is so awful it must be avoided at all costs.

What about another scenario where the Liberal vote collapses completely and goes PC to the point that the Liberals end up in third and the NDP the Official Opposition without picking up that many seats? Kind of like the wrap-around effect of the last 2 Nova Scotia elections where the third place party ended up as the Official Opposition?

I'm one of the least partisan posters here.

That's what you don't seem to realize.

I don't view all parties & leaders in terms of black & white, good & evil, like the hardcore NDPers.  I view the world more along the way most of the population does outside of the NDP hardcore base that even the NDP doesn't take seriously.

One of the reasons both opposition parties have failed to defeat Harper is because most Canadians don't view Harper as evil - they view him more in shades of gray and many Liberal & NDP candidates forget this.  Canadians see Harper as a good father, lover of hockey, Tim Hortons guy, etc.

You have to get outside of the bubble and start looking at some of these things the way voters do.  Justin Trudeau is not the person you portray him to be, either.  He's someone who loves his children & his wife, and has had moments of tragedy in his life (eg. the loss of his brother in a B.C. avalanche).  He has not lived a fairytale life.

Debater wrote:

I'm one of the least partisan posters here.

I don't view all parties & leaders in terms of black & white, good & evil, like the hardcore NDPers.  I view the world more along the way most of the population does outside of the NDP hardcore base that even the NDP doesn't take seriously.

You have to get outside of the bubble and start looking at some of these things the way voters do.  Justin Trudeau is not the person you portray him to be, either.  He's someone who loves his children & his wife, and has had moments of tragedy in his life (eg. the loss of his brother in a B.C. avalanche).  He has not lived a fairytale life.

Saving this for an eventual rabble babble hall of fame.

Lib 54 PC 45 NDP 11

Debater wrote:

I'm a partisan in the sense that I am a Liberal, but I'm actually less partisan than you & most of the other hardcore NDP posters here.  I've actually voted for other parties in the past and can engage in constructive criticism of my own party and see the strengths & weaknesses of each leader.  That's not something most hardcore NDPers here are willing to do.  They see everything in black & white terms.

Yup.  Liberals and Conservatives bad, NDP good.  That is my perspective.

Oh, and emoticons are bad too.

Ciabatta2 wrote:

If you can forget about the phony "prediction" aspect and can sift through the partisan trolling, there is some good local insight you get form the odd post on electionprediction.com that no one is reflecting in the media or on other websites.  It should be localelectiondynamics.org or something, but we all know that's no as compelling.

Election atlas has the NDP holding their current seats because it hasn't taken into account the latest polls.  I don't see anything but NDP losses (overall) as plausible.

I think Grenier's is the most likely to end up accurate, although let's be honest looking at general polling trends and pointing out each party's weak ridings is not rocket science if you pay attention to this sort of stuff.

 

Grenier's projections are all over the map. No way he is going to be the most accurate. In 2011, he only got 85% correct, and usually does worse in other provinces. Electionprediction.org got 91% last time and the Canadian Election Atlas got 92%.  tooclosetocall's projection looks very sound as well, I wouldn't be surprised if he is the most accurate this time.

twinklestar wrote:

Grenier's projections are all over the map. No way he is going to be the most accurate. In 2011, he only got 85% correct, and usually does worse in other provinces. Electionprediction.org got 91% last time and the Canadian Election Atlas got 92%.  tooclosetocall's projection looks very sound as well, I wouldn't be surprised if he is the most accurate this time.

Actually, his projections are generally quite good. Since they capture data over a longer period of time, however, his aggregates take a while to capture rapid changes such as the NDP surge in 2011. If he had completely dropped all the polls more than a few days old, he'd have done alright in 2011.