2016 US Presidential Election

Any general discussion an go here
martinkil
Posts: 3415
Joined: Sun Jan 11, 2015 9:55 am

Re: 2016 US Presidential Election

Postby martinkil » Mon Mar 14, 2016 11:13 am

A big day on Tuesday with five states holding primaries on the Democratic side. In my first post I'll focus on Illinois

In anticipation of a better result in Michigan than expected (I wasn't expecting a win, but it was a pleasant surprise), early last week I started laying Clinton at 1.04 in the Illinois Democratic Primary (at this point the latest polls gave her a lead of between 29% and 35%). Even after the Michigan win I was still able to lay Clinton in Illinois at an average price of 1.1.

The rationale of my further bets were that there had probably been a surge in Sanders numbers which I expected to be reflected in this weekend polls, and I was also expecting the major polling companies to have a more accurate picture than the Michigan debacle, since I had noticed that especially the last CBS Michigan survey had only about half the required number of Independents included.

The weekend polls came out and they were much better than I expected. The first Illinois poll (NBC/WSJ/Marist) had cut Clinton's lead from 30% to just 5%, and was followed by a (CBS/YouGov) which actually gave Sanders a 2 point lead in Illinois.

Having looked at the numbers behind the polls I still feel these polls may be under counting Sanders support, especially amongst Independents, but not as disastrously as the Michigan polls.

In Illinois the demograph is both African American (1.33 million eligible voters), and Hispanic (0.95 million). Eligible voters is not the same as registered voters, but lets hope Sanders ground game has been getting the young registered over the last few months. The CBS Poll have them breaking 50/45 for Clinton, but the NBC polls had them breaking 64/30 for Sanders which since 34% of them are in the 18-29 age group could be really important for Sanders.

From the numbers Sanders must have a really good chance of taking Illinois on Tuesday.

martinkil
Posts: 3415
Joined: Sun Jan 11, 2015 9:55 am

Re: 2016 US Presidential Election

Postby martinkil » Mon Mar 14, 2016 4:25 pm

Florida on Tuesday is a Clinton lockdown. Not only are most Democrats over 50 with a large retirement community, but it's a closed primary, so any independents would have had to have changed their registration by February 16th.

North Carolina has 22% African American, and will most likely follow the deep south states with the black vote breaking for Clinton 80/20. I wouldn't expect quite the same bloodbath as the deep south states but Clinton really should win this.

martinkil
Posts: 3415
Joined: Sun Jan 11, 2015 9:55 am

Re: 2016 US Presidential Election

Postby martinkil » Sat Apr 23, 2016 4:31 pm

Clinton has an almost unassailable lead in pledged delegates after winning the New York primary by 16%, 58/42. The demographics of New York, which is the State Clinton represented as a Senator for two terms are 17.6 African American, 18.6 Hispanic, and 8.5 Asian. This was a very restrictive Closed Primary. Voter wanting to change from a party to the Dems needed to have done it by October 2015. New registrations for the 19 April had to be in by 25 March. There was no Same Day Voter Registration.

On Tuesday 26 we have 5 primaries, 4 of which are closed, Connecticut, Delaware, Maryland, Pennsylvania and one semi-closed Rhode Island.

Maryland demographs are 29.3 African American, 8.1 Hispanic, and 5.5 Asian, so this is really a Clinton home run and the polls show her 20 points clear.

Delaware demographs are 21.2 African American and 9.0 Hispanic, so this is really a Clinton home. The is only one poll which gives Clinton a 7% 45/38 lead, but this poll has only 5% African American in the sample, so is next to useless. Based on the Demograph this is a likely Clinton win.

Pennsylvania demographs are 11.6 African American and 6.6 Hispanic which is similar to Missouri which Sanders only just lost 49.4 to 49.6. Recent polls have Clinton leading by 13, 17, 11, 6. Voter registration deadline was 27 March.

Connecticut demographs are 11.5 African American and 15 Hispanic, so the Hispanic breakdown will be very important. There have been just two poll. One has Sanders down 9% and the other down 6%. It's a closed Primary the voter registration deadline was 21 April and there is same day registration and if you are unaffiliated you can affiliate to the democratic party on polling day.

Of the above, Connecticut has Clinton available to lay at 1.1 and 1.13 which might represent a little value especially if Sanders can move the polls a little with TV adds.

martinkil
Posts: 3415
Joined: Sun Jan 11, 2015 9:55 am

Re: 2016 US Presidential Election

Postby martinkil » Mon Apr 25, 2016 11:10 pm

A very helpful Connecticut poll had sanders just two points behind Clinton which allowed Clinton to trade at 1.3 for a while.

Could well be interesting since 23,000 new under 30 yo's have registered as democrats this year (will be between 250,00 and 350,000 voter on Tuesday's democratic Connecticut contest).

Also between 1 jan and 12 April - , 13,722 unaffiliated voters joined the Democratic Party - there may have been some more since you can change from unaffiliated to Democratic until 12 noon on 25 April.

martinkil
Posts: 3415
Joined: Sun Jan 11, 2015 9:55 am

Re: 2016 US Presidential Election

Postby martinkil » Sun May 01, 2016 3:14 pm

It felt so close in Connecticut even though Clinton eventually won by 51.8/46.4. I had managed to get my stake back before voting started thanks to the helpful Connecticut poll which had sanders just two points behind Clinton which allowed Clinton to trade at 1.3 for a while.
During the count when 50% of the vote had been counted Sanders was 1.8% ahead still. Unfortunately many Clinton strongholds had still to be counted.

On Tuesday we have Indiana and in an effort to understand how it's Demographic could determine how the voting could unfold I've had a look at the three adjacent states which have already voted

State - African American%, Hispanic%, Asian%, Primary type, Clinton/Sanders
Michigan - 14.2% - 2.3% - 2.4% - Open Primary - 48.2/49.8
Illinois - 14.7% - 15.8% - 5.3% - Open Primary - 50.6/48.6
Ohio - 12.2% - 3.1% - 1.7% - Semi-open Primary - 56.1/43.1

Indiana - 9.6% - 6.6% - 1.7% - Open Primary - ?? / ??

Of the three which have voted Ohio looks the odd one out. Sanders lost by 13%, but there were a couple of reasons why that may have been so. Firstly the Ohio State University was on Spring Break and that wouldn't have helped, also Kasich was in a close run with Trump and many potential Independent Sanders supporters may have decided to vote in the GOP contest for Kasich, especially since 63% (2,043,043) of primary voters voted in the GOP contest, and 37% (1,202,163) in the Dems. In 2008 it was the other way around.

The latest poll has Trump leading Cruz by 15% - 49% to 34%, so it's unlikely many independents think they can make a difference this time. On the Dems side that latest poll has Clinton just 4% in front of Sanders 50/46, so many independents could well think thatr they could tip the scales.


On the face of it, Sanders should have a good chance of being within a couple of percentage points of Clinton one way or the other.

1 - It's an open primary
2 - There are only 9.6% African Americans, lower than any of the other three states.
3 - Clinton has refrained from buying any ad in the local Indiana media Hubs.

In the market Sanders is available to back at 5.9 and Cliton to lay at 1.25. I would put the price at 3.0 and 1.5 - at the most.

Since I've already laid Clinton at 1.15 I encourage as many as you as possible to lay Clinton, and let me trade out between 1.3 and 1.5 :)

martinkil
Posts: 3415
Joined: Sun Jan 11, 2015 9:55 am

Re: 2016 US Presidential Election

Postby martinkil » Wed May 04, 2016 4:09 pm

Well that was a perfect storm.
Sanders losing 5 of the 6 preceding contests (losing contests were closed primaries).
His campaign laying off 200 staffers
Fundraising drops from 44/45 million in Feb Mar to 25 million in April.
All the polls showing Clinton winning ( most were 4/5% within error margin)
Demographics were favourable to Sanders
This is an Open Primary - independents can vote
Clinton's Campaigns had no Media spend ( Sanders 1.8 million)

Then as the votes were counted the Early Voting/absentee votes were counted first giving Clinton a 10 point lead.

Clinton price collapsed

CNN exit polls show Sanders with a 10% point lead (poll was later adjusted to show a 5% win)

Day votes start to be included in the totals and by the time 19% of the vote has been counted they were 50/50.

Sanders lead slowly grows.

In only all betting days were like that :) :) :)

martinkil
Posts: 3415
Joined: Sun Jan 11, 2015 9:55 am

Re: 2016 US Presidential Election

Postby martinkil » Sun May 08, 2016 9:30 pm

We're on the home stretch of the Democratic Primary contest with just 12 more contests consisting of 926 Pledged delegates which is over 22% of the total pledged delegates. California alone accounts for 475 of those pledged delegates.

On Tuesday West Virginia with 29 pledged delegates votes. Sanders is a heavy favourite in a semi-closed primary where independents can vote. Real Clear Politics average currently has Sanders up by 6% with the last two polls having him up by 8% and 4%. With only 3% African American and 1.2% hispanic this very white state should be a Sander's banker, and the prices reflect that.

The following week on Tuesday 17 May both Kentucky and Oregon vote. Both are closed primaries which have been problematic for Sanders with independents not being allowed to vote.

Oregon though is a very liberal State and is thought to be a Sanders stronghold. With 2% African American and 12% Hispanic is demographics are very similar to it's northern neighbour Washington State which went for Sanders in a big way. Not surprisingly Sanders is a big favourite on the exchanges.

Kentucky though is a different kettle of fish. It's far from a liberal state invariably returning a republican and has 8.2% African American and 3.4% Hispanic. Those demographic are actually more favourable to Sanders than Indiana his last winning state (by 5%), but that was an open primary where independents can vote, in Kentucky only registered Democrats can vote in the democratic primary.
There are no recent polls to guide us, and the markets indicate a slight preference for Clinton. Since I have no theory as to why Sanders can win this one I'll leave well alone.

martinkil
Posts: 3415
Joined: Sun Jan 11, 2015 9:55 am

Re: 2016 US Presidential Election

Postby martinkil » Tue May 10, 2016 11:28 am

Interesting development -
Rachel Maddow reports breaking news that the Clinton campaign has said that after effectively ending ad spending against Bernie Sanders in the Democratic primary that decision is being reconsidered and ad time will likely be purchased for at least some of the upcoming primary contests.

Having lost Indiana by 5%, maybe they consider that it would have been closer had they hit the airways.

I'll await developments with interest

martinkil
Posts: 3415
Joined: Sun Jan 11, 2015 9:55 am

Re: 2016 US Presidential Election

Postby martinkil » Tue May 10, 2016 11:31 am

Polling this season for the Democratic primaries have been mixed. There appear to be three classes of primaries, The Deep South, where the polls often over estimated Sanders numbers, Closed Primaries where they have been a lot more accurate, and Open semi open/closed primaries away from the Deep south where Sanders support has been underestimated by the polling averages.

Since the upcoming Californian Democratic primary is a semi-closed Primary where registered democrats and registered No Party Preference are allowed to vote I had a look at exactly how much Sanders support has been underestimated by these polls looking at states which have enough polls to form a view.

It's turns out that on average these polls underestimate Sanders chance by 8.7%

Detailed analysis:

If we compare the Polling average from three sources - RCP% - HuffPost - 538 against the actual results we exactly how much the polls in general have been overestimating Clinton support in open semi open/closed primaries away from the Deep south :

STATE - RCP% - HuffPost - 538
Michigan - +22.5% - +19.8 - +22.8
Oklahoma - +12.4 - +9.6 - +10.2
Indiana - + 11.8 - +12.7 - +12.1
Wisconsin - + 10.9 - +10.0 - +10.6
North Carolina - +10.2 - +6.2 - +8.2
New Hampshire - +9 - +9 - +8.6
Massachusetts - +5 - +7.7 - +6.3
Illinois - +0.5 - +5.9 - +5.2
Ohio - (5.8) - (1.8) - (3.1) (negative - i.e. overstated Sanders)


STATE - average(mean) Clinton Overestimation
Michigan - +21.7%
Indiana - +12.2%
Oklahoma - +10.7%
Wisconsin - +10.5%
New Hampshire - +8.9%
North Carolina - +8.2%
Massachusetts - +6.3%
Illinois - +3.9%
Ohio - -(3.6%)


Of the above, Michigan (2008 benchmark) and Ohio (Kasich's anti-Trump vote) are the outliners. If we remove them and find an average we get 8.7% (with them included it's 8.8%). So, I'm saying that on average, once we get away from the deep south the polls in Open or semi-open/closed primaries have on average underestimated Sanders chances by 8.7% - with a range of between 3.9% - 12.2%


For Californian poll I suggest you subtract 3.9 or 12.2 and the correct poll value will most likely lie somewhere between those values.
As of 8-May-2016:
The RPC poll average for California is 51/41.3 - +9.7 Clinton
The HuffPost poll average for California is 49.7/40.6 - +9.1 Clinton
The 538 poll average for California is 52.5/38.4 - +14.1 Clinton

Gives an average difference of 11.0 - Applying the 3.9 - 12.2 correction the true poll value is likely to lie somewhere between 7.1 for Clinton and 1.2 for Sanders - with 2.3% for Clinton the most likeliest value (11.0 - 8.7)

martinkil
Posts: 3415
Joined: Sun Jan 11, 2015 9:55 am

Re: 2016 US Presidential Election

Postby martinkil » Tue May 10, 2016 12:11 pm

Lets try out our model on West Virginia

RCP - HuffPost - 538
Sand +6% - +4% - 6.2%

Average is 16.2/3 = 5.4% for Sanders.
Expected win would be 5.4 + 8.7 = 14.1%
Within a range of 5.4 + 3.9 and 5.4 + 12.2 -
Sanders will win by at least 9.3% upto 17.6%
- with a likely advantage of 14.1%

Update:
In West Virginia Sanders won 51.4/35.8 - an advantage of 15.6 - very close to the 14.1 I expected and within the 9.3% to 17.6% range


Return to “Down The Pub”

Who is online

Users browsing this forum: No registered users and 8 guests