Serving the Quantitative Finance Community

 
User avatar
tags
Posts: 3162
Joined: February 21st, 2010, 12:58 pm

Re: General Election 2019

December 13th, 2019, 1:36 pm

Clear BREXIT mandate in England.
Not so in NI nor Scotland.
And, as usual, nobody cares about Wales.
Wales got electricity last year. They are fully on their own now, I assume.
Last edited by tags on December 13th, 2019, 2:06 pm, edited 1 time in total.
 
User avatar
katastrofa
Posts: 7440
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: General Election 2019

December 13th, 2019, 2:05 pm

Tagoma, where did you catch this lightly sarcastic sense of humour? Spent a weekend in Norway?
 
User avatar
Cuchulainn
Posts: 20254
Joined: July 16th, 2004, 7:38 am
Location: 20, 000

Re: General Election 2019

December 13th, 2019, 2:34 pm

Image
Tagoma, where did you catch this lightly sarcastic sense of humour? 
 
User avatar
FaridMoussaoui
Posts: 327
Joined: June 20th, 2008, 10:05 am
Location: Genève, Genf, Ginevra, Geneva

Re: General Election 2019

December 13th, 2019, 3:02 pm

Clear BREXIT mandate in England.
Not so in NI nor Scotland.
Moving from Great Britain to Little England.
 
User avatar
Paul
Topic Author
Posts: 6604
Joined: July 20th, 2001, 3:28 pm

Re: General Election 2019

December 13th, 2019, 4:22 pm

I hear that across the land academics have not been turning up for coffee. So no "work" today then.
 
User avatar
Cuchulainn
Posts: 20254
Joined: July 16th, 2004, 7:38 am
Location: 20, 000

Re: General Election 2019

December 13th, 2019, 4:27 pm

I hear that across the land academics have not been turning up for coffee. So no "work" today then.
Are they on a 'go slow'? Sulking?
Last edited by Cuchulainn on December 13th, 2019, 4:31 pm, edited 1 time in total.
 
User avatar
Cuchulainn
Posts: 20254
Joined: July 16th, 2004, 7:38 am
Location: 20, 000

Re: General Election 2019

December 13th, 2019, 4:29 pm

BTW, Paul,
How accurate were your predictions in absolute terms?
 
User avatar
Paul
Topic Author
Posts: 6604
Joined: July 20th, 2001, 3:28 pm

Re: General Election 2019

December 13th, 2019, 4:32 pm

I hear that across the land academics have not been turning up for coffee. So no "work" today then.
Are they on a 'go slow'? Sulking?
I didn't know it was physically possible to go any slower. But academics are always making new discoveries like that.
 
User avatar
Paul
Topic Author
Posts: 6604
Joined: July 20th, 2001, 3:28 pm

Re: General Election 2019

December 13th, 2019, 4:48 pm

BTW, Paul,
How accurate were your predictions in absolute terms?
My last forecast was 22 majority and 90% chance of Conservative majority. It was a majority of 80.

The final Conservative lead was 12%. The average poll in the run up was around 9%, including the biggest. 

Last time the final lead was 2% with the polls at 7%. I.e the other way around!

Even those who got the seats right had the wrong percentages. And I suspect vice versa!

With the correct final polls I would have estimated a majority of around 32.

It's fun! What is most important to model, the percentages, how the percentages turn into seats (even if your percentages are wrong!) or go straight to seats (but then how do you use the only info you have, the polls?)?
 
User avatar
tags
Posts: 3162
Joined: February 21st, 2010, 12:58 pm

Re: General Election 2019

December 13th, 2019, 4:58 pm

BTW, Paul,
How accurate were your predictions in absolute terms?
My last forecast was 22 majority and 90% chance of Conservative majority. It was a majority of 60.

The final Conservative lead was 12%. The average poll in the run up was around 9%, including the biggest. 

Last time the final lead was 2% with the polls at 7%. I.e the other way around!

Even those who got the seats right had the wrong percentages. And I suspect vice versa!

With the correct final polls I would have estimated a majority of around 32.

It's fun! What is most important to model, the percentages, how the percentages turn into seats (even if your percentages are wrong!) or go straight to seats (but then how do you use the only info you have, the polls?)?
Hi Paul. Is your forecasting methodology for this specific vote exposed/discussed somewhere, please?
 
User avatar
Paul
Topic Author
Posts: 6604
Joined: July 20th, 2001, 3:28 pm

Re: General Election 2019

December 13th, 2019, 5:14 pm

No! Only if I'd done better!

I did try to simplify things. So I focused more this time on how the polls would turn into final votes nationwide, not by seat (like I'd done before). 

I had always wondered if a more qualitative approach might be better. E.g. focus on marginal seats. But that's clearly not going to work judging by the number of seats going blue that had been red since the dawn of time!

There's also the issue that pollsters only do small samples generally (sells more newspapers if they bounce around and cheaper?). But even aggregating them, or only looking at the large polls, doesn't help.

It's a really interesting problem!
 
User avatar
katastrofa
Posts: 7440
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: General Election 2019

December 13th, 2019, 5:30 pm

I believe it's your pet project, but if you're interested in a little collaboration before the next elections, we could gather available data and run a microsimulation including all details of the demographic changes, like here for pensions: https://arxiv.org/pdf/1802.09427.pdf
 
User avatar
Paul
Topic Author
Posts: 6604
Joined: July 20th, 2001, 3:28 pm

Re: General Election 2019

December 13th, 2019, 5:40 pm

Yes, Yes, YES! I wish you’d said this a few weeks ago!

The trick is not so much the modeling (which is really, really hard) but being brave enough to look stupid in real time! Have you got what it takes?!
 
User avatar
katastrofa
Posts: 7440
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: General Election 2019

December 13th, 2019, 5:49 pm

Do you really have to ask? :-)
 
User avatar
Paul
Topic Author
Posts: 6604
Joined: July 20th, 2001, 3:28 pm

Re: General Election 2019

December 13th, 2019, 5:57 pm

Ocean’s 2!