Seat Predictions Ontario 2018 Election

NDP-70

pc-40

liberal-14

It looks as though the predictions sites are all under heavy pressure(or perhaps some sort of direct orders)NOT to show the PC's falling short of a majority in the seat count.  

why you lumping all Canadians into electing doug ford?

Ontario is doing it all themselves to themselves if they elect Ford nation.

my gma always used to say "Ontarians might as well be American" hope Ontarians prove her wrong.

robbie_dee wrote:

Oh, we’ve mostly been nuts from the get go for sure, but new communications technology like social media that lets our different flavors of nuts connect with each other and magnify their collective voices, along with eroding social cohesion and increasing economic insecurity, have made the problem worse.

Yeah that.

Now I will go out on a limb and predict:

Many of the votes that were/are going to the PCs are "The Liberals have to go and the PCs have always been the most likely alternative". Now that the NDP are a serious contender more people will be giving them a serious look.

I predict a higher than predicted turn out from millenials in reaction to a possible Ford win and the rise of the NDP.

I don't know what the seat distribution will be but I predict Horwath will be the next premier of Ontario.

Not to mention Preston Manning, the original "drain the swamp and refill it with Nazis" candidate. But then, as an immigrant, it's been my observation that many Canadians don't actually live in Canada. They inhabit more of a "Canada of the mind", devoid of the unpleasant facts of its actual history, with an idealized political culture which didn't really vote for Harper, or Ford, or Mulroney. It's why Trudeau could give his "Canada is back!" victory speech without many raised eyebrows. The things we don't like about Canada aren't really Canada, see. Racism's certainly not a part of our Canadian heritage.

Agreed that seat numbers are really no more than guessing.

I'm fairly certain that Ford will win a majority.

I also think the Wynne will end up second due but I understand how crazy that sounds right now. But I think the NDP has risen too early, the Liberals have some very concentrated support, and the debate is the start of the come back narrative that the media has nicely left open for the leader that best suits their business.

Not at all. I'm just someone with a different analysis, who thinks that:

a) current NDP support levels are a result of vote-parking while people wait to see who will rise during the last third of the campaign

b) the NDP has peaked too early

c) the media has been setting up a St. Kathleen narrative to manufacture a rise for Wynne

d) the debate was not beneficial for Horwath beyond her base (and particularly not with voters open to Wynne)

e) the power of incumbancy and demopgrahics,

f) that the Liberal vote (given riding and vote patterns) is more efficient than the other two parties (a point I believe that you and I disagreed on previously.)

In another thread someone said Wynne's good debate could only get her "a piece of a piece of a piece" and I agree - I think that's all the Liberals need to come in second or a not super distant third.  Also, I'm basing the prediction on a rise in their polling (not current numbers). 

A lot can happen in 10 days.

There's simply no case for any progressive to vote Liberal.  There's another, MORE progressive party well ahead of them in the polls, in a dead heat at worse with the PC's and more often than not in the lead.  And probably half of the voters who'd habitually backed the OLP in the last few elections were people who were closer to the NDP on the issues but who believed that only the OLP could beat the PC's or keep them out after they'd been beaten.  Since the tactical case for progressive votes for the OLP no longer exists, and since there's a better party this time which is more likely to win, what remaining case even exists for voting OLP at all?  There's certainly no conviction-based case, since the OLP HAS no unchangeable convictions on anything.  If you vote ONDP, you vote for what you care about AND you vote for the progressive party with the best chance of winning.

Ciabatta2 wrote:

Not at all. I'm just someone with a different analysis, who thinks that:

a) current NDP support levels are a result of vote-parking while people wait to see who will rise during the last third of the campaign

b) the NDP has peaked too early

c) the media has been setting up a St. Kathleen narrative to manufacture a rise for Wynne

d) the debate was not beneficial for Horwath beyond her base (and particularly not with voters open to Wynne)

e) the power of incumbancy and demopgrahics,

f) that the Liberal vote (given riding and vote patterns) is more efficient than the other two parties (a point I believe that you and I disagreed on previously.)

In another thread someone said Wynne's good debate could only get her "a piece of a piece of a piece" and I agree - I think that's all the Liberals need to come in second or a not super distant third.  Also, I'm basing the prediction on a rise in their polling (not current numbers). 

A lot can happen in 10 days.

 

 

Ok -- I disagree with the concept of vote parking at this point. Not sure what evidence you have beyond wishful thinking that this is not a decision based on all the choices people have. There is no reason to take a vote from the NDP and give it to someone else when they are leading in the last ten days.

I agree (b) might be true although we have not yet seen the NDP momentum falter. If it is going to then this will be evident this week. But you are saying this based on absolutely no evidence (my reason for suggesting you were being hopeful).

I think that (c) is not new at this point and if anything more and more of the media are giving up on the idea that the Liberals are still in it right now.

I think on (d) you have half a point but that most are not open to Wynne and the the thing Horwath had to do was prevent Ford from winning.

I think (e) is marginal. Incumbancy is not a big thing when the government is as hated as this one is. I agree the level of hatred is irrational but it is there. Demographics might not be as big a factor as it has in the past if the younger generations decide they don't want a Trump-like premier for Ontario. (Yes I know Ford is not Trump but there are reminders.)

Your point in (f) is based on a complete misunderstanding of what effiecient means -- efficient is about distribution of votes in winable seats without losing too many in losing ones. It is totally collected to support level. A party that is efficient at 40% is extremely unlikely to be efficient at 25% or lower. If you run the math you will see how illogical it is to suggest that a party can be effieicnt at all support levels. The Liberals are very evenly spread in the province and this is why they are very efficient when they lead in the popular vote but it makes them very vulnerable when they trail. The reason is this support level is reproduced across the province more evenly. A pile of second place finishes would be considered inefficient - right? Well add five points acorss the board and then they are a pile of first place wins. A winning party that is efficient means they win by a little everywhere -- lost 15 points across the board then they lose everywhere. A party concentrated in strongholds will find it difficult to break out of them (this may be the NDP's difficulty here) but that also means when they are down to 20% they can still take seats. A party inefficient at 35% is liable to hold seats due to strongholds when down to 20%. a party efficient at 35% is liable to lose almost everywhere . That benefit of having no extra padding in lost votes at 35% is a huge risk when you lose one third or more of your vote. Really you should consider this. Votes are equal. Efficiency in a party is based on where it normally tracks and what votes are wasted. A big change in popular vote means a big difference in efficiency.

I agree that a lot can happen in ten days but your logic here seems very faulty -- especially considering your complete misunderstanding of what it means for a party to be efficient. -- it is distribution and level not a characteristic of the party.

Ken Burch wrote:

There's simply no case for any progressive to vote Liberal.  There's another, MORE progressive party well ahead of them in the polls, in a dead heat at worse with the PC's and more often than not in the lead.  And probably half of the voters who'd habitually backed the OLP in the last few elections were people who were closer to the NDP on the issues but who believed that only the OLP could beat the PC's or keep them out after they'd been beaten.  Since the tactical case for progressive votes for the OLP no longer exists, and since there's a better party this time which is more likely to win, what remaining case even exists for voting OLP at all?  There's certainly no conviction-based case, since the OLP HAS no unchangeable convictions on anything.  If you vote ONDP, you vote for what you care about AND you vote for the progressive party with the best chance of winning.

 

I think this is another thing that we might agree we would like but may not be true.

I dislike strategic voting but there are still seats where the Liberal may be positionned better than the NDP. I won't vote strategically but can you call a person who does not a progressive if they are trying to stop Ford? Even if they are mistaken?

We could always say what you say here that it is wrong to be a progressive and vote this way -- but don't we always consider those voting against us as wrong? You would hope you are voting the right way and that the others are wrong? I disagree with the Liberals and totally disagree with every position of the Conservatives but who am I to say someone's choices are objectively wrong? Even if I object to them in every way... The fact is we each have a vision of government and priorities and we do not share the same ones.

The only incorrect vote is one against the person's intention even if we disagree.

Pogo wrote:

At this point in a campaign voters have a general idea of how they rank the parties/candidates/leaders.  Any changes going forward will either be around fence sitters or strategic ("I really wanted to vote Green, but I don't think they are going to win").  This strategic step was fatal to the NDP last federal election. 

In this election I would see two contradictory strategic decisions occurring on the left/center.  People will be looking to the NDP as the vehicle to stop Ford and others will be looking at giving Liberals votes to ensure they have credible presence and/or keep the NDP from 'going too far'.  I would bet on more people going to the NDP to stop Ford.

I have to quibble with this. Even though I did vote NDP in the last federal election, I disagree that the momentum to the Liberals was strategic. I think people bought what the Liberals were selling and not what the NDP was. The momentum to the Liberals shifted their way before it made any strategic sense. I think the anger at Mulcair and that campaign is evidence that the party understood this.

 Sean, I agree and disagree (not as cool as sorry/not sorry).  I have lots of anecdotal evidence of friends/co-workers making strategic decisions to pick the party best suited to get rid of Harper.  However to your point the Liberals only became this choice because they overcame a big NDP lead.  The NDP lost the election outside of any strategic voting but the magnitude of the loss was amplified by Liberals becoming the voter's choice to take on Harper.

Wynne's biggest problem now is how revolting Ford is to many of her erstwhile supporters. Lib/NDP swing voters who liked her debate performance and might be inclined to switch back still have to confront the knowledge that doing so could get Ford elected.

The NDP lost the last federal election because, at a critical moment, its Quebec base started to melt away. This in turn undermined its position as the default anti-Harper option and allowed Liberal voters to return home. It didn't help that Mulcair seemed stiff and dour trying to run a classic frontrunner campaign against the more dynamic Trudeau. Moreover Harper, while unpopular among a lot of people in the Liberal/NDP universe, was also probably still a less frightening boogey man than Ford because at least Harper appeared competent. But all that being said, I think the niqab issue is really what did it.

The niqab issue was really a sui generis event that ended up costing the NDP Quebec because the Party was out of step with a large number of its voting base over the issue of accomodating religious minorities in the public sphere and the supreme court dropped a critical, publicized decision on the exact issue in the middle of the campaign. I happen to think the NDP took the morally correct position on the issue, and ironically the end result was that the Liberals who agreed with the NDP's position reaped the majority of seats in Quebec, but the Liberals and NDP weren't really competing for the same voters in this regard and while the Liberals held their natural voting base the NDP took the loss on its exposed flank to the Bloc.

I just don't see an analogy to this in the provincial campaign. Maybe it could be the York University strike and the NDP's refusal to condone back to work legislation? That's obviously the one the Liberals are trying and it could work well among York University students and their parents but I think that's just too small of a group and it's hard to see the issue getting much broader legs. Besides where are the voters who are considering the NDP but disagree with them on this issue going to go? Back to the Liberals who were in government and allowed the issue to fester? Or to the Conservatives? I don't think so.

I think the parties' positions are pretty well settled with the voters now, much moreso than they were federally in September of 2015. I don't think the NDP will get the 47% that they polled with Forum and we might even see a slight pullback in the next poll from that organization. But things really do seem to be congealing at something in the mid to high 30s for the NDP, mid to low 30s for the PCs and low 20s for the Liberals and that is where I anticipate the final vote results will be. Then it will depend on riding-by-riding vote splits who ends up on top.

Knowing what the polling is looking like right now, why would any actual progressive in Toronto STILL vote OLP?  There's no "strategic" case for it.

This whole issue of some parties being more efficient in seats in generally false. It is about, as I have said before distribution, concentration and levels. But why do we have this concept despite the fact that an analysis of logic would show that this cannot be true? I want to explore that as a way to give a bit more of an indication of where things may go.

General comment on efficiency -- all parties will necessarily have various bands of support that are efficient depending on their level of concentration. So a party that is highly concentrated in strongholds would be remarkably efficient at a low level. One that is spread out would be inefficient at a low level but suddenly become efficient when this spread out support reaches critical mass in riding after riding. Then if the vote increases beyond that they will become inefficient again perhaps until support is high enough to get the next batch of seats. Example -- NDP in the 80s-90s. They were quite efficient at a low level but needed a significant boost to find a higher efficient level. They got that in 1990. If they had a little more they likely would not have gained much more as they did not have many near misses then -- they were mostly near wins. Had they obtained only 33% instead of 38% they would have been inefficient -- we know this becuase they barely won many seats and a reduction in support would have seen a pile of close seconds instead of firsts.

This is why in my discussion above I said that there is nothing about the Liberal party that is different -- they are no more efficient than other parties at all levels -- that is mathematically impossible. A pile of close wins means that a reduction in support would create a pile of close losses.

So where does the myth come form and why do scientific sites screw this up?

Most parties track for considerable time around a particular level of support. This is what creates the myth that they are somehow more efficient in all ranges rather than the ones we have seen them in. Scientific sites want to use data but the reality is that this does not exist at every support level in a meaningful way. So what they do is make presumptions that gains and losses are proportionate. they assume the concentration and distribution of support does not change as that support rises and falls. Of course this is illogical for a number of reasons. Let's look at one:

Assume for a moment that in riding A a party has 60% accessible voters and in riding B they have also 60% of accessible voters and in riding C they have only 30% of accessible voters and finally riding D where they also have 30% of accessible voters. These are the poeple that would consider them.

In riding A with a great local campaign, good workers, and a perfect platform for that riding they get 2/3 of the possible voters and win the seat.

Now with riding B a poor candidate and poor local campaign they can only get 20% of the vote.

Now riding C they have a great candidate and get more than 2/3 and 25% of the vote.

Now riding D with a poor candidate they only get 1/3 of the acccessible vote and 10% of the vote.

So now we suddenly get a better overall campaign. what will happen? The models will take the increasesed 15% provincially and apply it evenly. Riding A is still a win, Riding B is still a loss but they predict riding D as a win with this across the board boost.

However, instead Riding D was already clost to maximising its potential and still comes up as a loss but Riding B a longer shot wins. The science makes the mistake that the data we have about concentration and levels is a single bar that goes up and down and does not change with the action of an increase or decrease.

So let's look at this campaign and you see how the math plays out for each party:

The Conservatives are not that far from a predictable level but they were at a fairly efficient level and the vote was going up. The same assumption of efficiency is maintained and they are presumed to have this magical portable efficiency and so the predictors all say they will get more seats than their support suggests. Where do they come from?

Next up is the NDP. Presumptions are made about previous low and flatter levels of NDP support and that this distribution would remain identical at a higher level. So some NDP seats are tagged Conservative.

Then you have the Liberals. This one is really the head scratcher becuase there are two extremes here. On one side people presume that the distribution of Liberal support would be constant even with a huge decline and show the Liberals getting one or two seats. More likely, that decline would be towards seats that are stronger for them with better candidates and so the Liberals ought to be able to get more but the so-called science says only two seats. The reality may be that in decline the Liberals actuall could hold 6-8 seats but the across the board stuff does not see that. On the other extreme Liberals have the illusion that it is not the distribution but the magical efficiency that is portable and imagine that the Liberals in this low range would somehow retain 20 seats. Both extremes are wrong becuase they are both based on characteristics of level, distribution and efficiency that are deemed independently portable for a party rather than functions of each other without much to do with the party.

The one characteristic that is related to a party is the core support -- the people who would vote for them even in bad situations. Are their concentrations of this? It might make a party more efficient at a lower rather than higher level.

Just when you though this was easy there is another variable. That is who the competition is. The seats the Liberals would find safe in an NDP surge are nbot the same seats they woudl find safe in a Conservative surge. This means that depending on the concentrations a party could be wiped out or find some efficiency people did not expect based on which other party these concentrations were up against. The example I gave was Ottawa Vanier. This is a very, very safe Liberal seat from a Conservative surge. But if the surge is NDP, it might be a lot less safe. Ottawa South might seem safe for the Liberals, in a NDP surge but not in a Conservative one. You have to consider this dynamic as well.

For all these reasons these predictors are only valuable based on an assumed cross-the-board movement that almost never happens.

So let me predict this:

1) The Liberals will hang on to more seats than the 1-2 predicted since their support when very low will likely be more concentrated in pockets. This is only in part becuase in those pockets there will also not be as much strategic vote against them.

2) The difference in NDP vs conservative efficiency is BS. It could go either way in the election depending on changes in support level that we cannot see right now.

3) The likely voter thing about the NDP is exaggerated but it is a real thing. It means if the turnout is high then that is good news for the NDP and if it is low good news for the Conservatives.

4) Most parties generally have some support level at the low end where their core bases are retained. (The famous 1993 federal washout was due to the Reform Party which was a split in the PC party taking the fortresses away.) This means that washouts are possible but less likely.

5) Most parties are not very efficient in the low 30s with a party with more than 5 points above them. So if there are three majors in the low 30s anything can happen but if one is in the low 30s and one 5 points up in the higher 30s with a third way below -- that low 30s is going to lose disproportionately. Many political scientists have said that predictions of a majority come from the gap from 1st to 2nd more than the vote share.

6) While almost all are calling for a majority -- this could be incorrect if they are lowballing Liberal seats and the Liberals get 5-7 seats instead of the 1-2 -- and if the NDP and Conservatives are within a couple points of each other. The deciding point here will be whether the surviving Liberals are in Lib-PC fights or Lib-NDP fights. In other words which party the Liberals are holding off to keep their handful.

7) The NDP does not need to be in the mid 40s to have a majority. They need to have 6-8 points more than the Conservatives. So if the Conservatives get 33% then the NDP only need 39-41%. If the Conservatives manage 35% then the NDP will need the 41-43% for a majority. If the Conservatives get 37% the NDP would need 43-45%. As you can see this leaves less and less room for the Liberals and Greens and smaller parties. This means effectively that the NDP likely needs 40% AND for the Conservatives to be held below 34%.

8) For the Conservatives the math is not much better. They also would need near 40% AND to keep the NDP from crossing 34%.

9) If the parties are very close -- within 4 points -- either one could lead but are unlikely to get a majority.

10) Without seeing the Conservatives -- if the NDP and the Liberals are both doing well in seat-count -- for their level of support (efficiency) then the Conservatives need a higher raw vote to win. So this means that the early returns not in all parties' best seats can still tell you soemthing.

So when you are watching -- don't just look for the totals of one party or make presumptions about efficiency. Look for the gaps. Hope this helps.

Sorry but I still cannot see what the basic gaps are to be able to predict. I will try before the polls close.

I do not see last minute guesses as cheating. I see earlier guesses without data as baloney.

The too-close-to-call predicter is very screwed up with the bias built into it.

Let the idea of a Conservative majority coming from this sink in:

L 21% C 34% N 37% G 5%

Seats L 6 C 65 N 55 G 0

A party will not LIKELY get a majority losing by 3% in the popular vote AND at only 34%.

Think what a shocker the Bob Rae majority was at 38% leading by 6%.

This is not to say this is impossible -- but a predictor is supposed to be what is likely not what would be the greatest seat count ever recorded in Canada for a party with the lowest percentage and losing the popular vote by 3%. that takes some pretty shady math principles.

This is bad science making presumptions about distribution/concentration remaining unchanged as a party support rises and presumptions about portable efficiency for a party no matter what support level they are at.

Any decent understanding of math will tell you that if you are more efficient at one level of support wou will be less efficient at another.

Then consider which party is more likely to build up regions of wasted votes -- the Conservatives or the NDP? Given that there are parts of the province where you could run a bunny rabbit as a Conservative and win -- I think you can see what I am getting at.

But if you need more data to see how cracked this is consider this:

Enter the results of the 2014 election. the predictor shows:

L-PC-N

61-42-21=124

seats % 49-34-17

Actual results

58-28-21 = 106

Seats % 55-26-20

So you see an apparent bias away from the NDP -- even when the numbers put in are identical.

Mighty Middle wrote:

Could there be a possibility that we could wind up with only two parties represented in the legislature with official status? Even if the Libreals win maybe 2 or 3 seats, could they possibly join the NDP and then it would be like BC? That it could look like this

NDP - 81

Conservatives - 42

Green - 1

How would it change things if it was just a two party system in the Ontario Legislature (notwithstanding the lone Green Party MPP)

That is a fantasy. For the Greens to win a seat, the Liberals cannot be so low in Guelph. The Greens need the Liberals to hang on to some support as if their support goes NDP or Conservative at that rate one of the two would win. If the Liberals hold enough to allow the Greens to come up the middle then they have more than enough support to hold seats elsewhere.

But it is likely that the Liberals are in danger of losing party status -- even if they get 5-6 seats which is more likely than the predictions of 0-2.

The Liberals might be able to retain 3 in the GTA and 2 in Eastern Ontario even if the numbers are bad.

Very Far Away wrote:

Sean in Ottawa,

A few years ago, I said almost the same thing: You are the main reason I come to this forum. (Because of the poor design of the site and some endless fights among members, I haven't been visiting it regularly for a long time).

Your analyses and the way you explain them are great. Thank you very much.

 

 

Thank you for your kind words. I am glad that the effort of trying to work through these things sometimes creates a value .