• It has come to our attention that some users may have been "banned" when they tried to change their passwords after the site was hacked due to a glitch in the old vBulletin software. This would have occurred around the end of February and does not apply after the site was converted to Xenforo. If you believe you were affected by this, please contact a staff member or use the Contact us link at the bottom of any forum page.

Irish Election Stats - Probability figure for each candidate


davehiggz

Well-known member
Joined
May 7, 2009
Messages
1,116
Hi folks,

I've just launched Irish Election Stats, a site you may be interested in. It brings together various data sources into a single statistical model to forecast seats for the election.

First time I've attempted something like this, but very pleased with the output so far. As new polls emerge, the forecasts will change.

I'm a huge fan of Nate Silver and his work on US elections. I've done my best to adapt his approach to the Irish system. We'll know on the 27th whether it's worked or not.
 

flavirostris

Well-known member
Joined
Dec 21, 2007
Messages
25,031
Wow . That is some awesome work
 

Carlos Danger

Well-known member
Joined
Nov 14, 2009
Messages
28,485
Website
www.youtube.com
Well done, Dave.

How do you manage to project second, third, fourth preferences? Nate Silver has a much easier task sine the US has a different electoral process.
 

ruserious

Well-known member
Joined
Jan 3, 2011
Messages
29,599
Any chance you'd buy this site? Well done.
 

Congalltee

Well-known member
Joined
Nov 10, 2009
Messages
6,211
The two SF candidates in DBN each have 80% plus probability of being elected?
Going from 0/6 to 2/5.
Really?
(There is a high chance I don't understand probability).
 

sadcitizen

Well-known member
Joined
Sep 27, 2011
Messages
4,330
Good work.

Just a warning, almost every table/graph except for the Dáil make-up graph are not showing up in IE11 or IE Metro on Windows 8.1. Most stats show that more people use IE than Firefox or Safari, so it'd be a good idea to make sure it works. I'll update IE here and see if it makes a difference.

Edit: Still definitely not on IE. I thought it was an iframe issue except it works on Windows Mobile IE, so I'm guessing it's a Datawrapper thing.
 
Last edited:

General Urko

Well-known member
Joined
Oct 24, 2012
Messages
15,749
Good work mate!
 

ShoutingIsLeadership

Well-known member
Joined
Jan 17, 2011
Messages
50,459
Does Regina Doherty get re-elected? If the answer is yes, then I hope your site gets shut down.


:)
 

Cathalio11

Well-known member
Joined
Jan 24, 2011
Messages
322
Sinn Fein 99% and Labour 92% to win a seat in Carlow/Kilkenny. Now that is audacious.
 

Dame_Enda

Well-known member
Joined
Dec 14, 2011
Messages
52,092
Hmmm have you factored in geographical considerations, like Mythen being the third candidate from Enniscorthy, which Ivan Yates thinks splits the Enniscorthy vote too much to elect Mythen?

I mean - Howlin only 1% chance of being elected? Even a Wallace canvasser have told me he will get in.

A good start and a great idea, but I think you need to adjust the model to take account of the role of geography and the innate conservatism of rural Ireland when it comes to leaving the Civil War parties, as well as the class breakdown. I suggest you consult the Census on social class. There will certainly not be as much of a swing away from Labour and FG among the middle classes as among the working class. The national swing will not be uniform. It hardly ever is.
 

Plebian

Well-known member
Joined
Feb 20, 2011
Messages
9,342
Hmmm have you factored in geographical considerations, like Mythen being the third candidate from Enniscorthy, which Ivan Yates thinks splits the Enniscorthy vote too much to elect Mythen?

I mean - Howlin only 1% chance of being elected? Even a Wallace canvasser have told me he will get in.

A good start and a great idea, but I think you need to adjust the model to take account of the role of geography and the innate conservatism of rural Ireland when it comes to leaving the Civil War parties, as well as the class breakdown. I suggest you consult the Census on social class. There will certainly not be as much of a swing away from Labour and FG among the middle classes as among the working class. The national swing will not be uniform. It hardly ever is.
Yates is too stuck in the past to realise that IF Mythen gets elected it will be on the SF brand being strong enough across the 5 seater to get him into contention. That same brand dynamic may cause chaos as regards the traditional geographical distribution of seats in a fair few constituencies.
 

Dame_Enda

Well-known member
Joined
Dec 14, 2011
Messages
52,092
Another factor is that the retirement of Liam Twomey has freed up some of the Wexford town vote for Howlin, according to Yates on Newstalk a few days ago. He said this "greatly improved" Howlins chances. This is a seat that survived the 1987 rout where Labour got 7% nationally.
 

Shiloh

Well-known member
Joined
Sep 12, 2008
Messages
323
This is very interesting. I don't know your methodology but it looks intriguing.

Just looking at the results though SF FF and independents seem to be getting more seats than anyone else is predicting. Also FG look a little shy.

Looking into things on a party basis it looks as if you are predicting than neither Joan Burton or Alan Kelly will win a seat - that would be some pickle for the laborious party.
 

Jack Walsh

Well-known member
Joined
Jun 26, 2015
Messages
9,327
Really top class effort.
Thanks

Will stay tuned and think you will get very close to actual result.
Any model that can strip away the skew caused by party bias is ok with me as is any model using Nate Silver methodology.
 

ireallyshouldknowbetter

Well-known member
Joined
Mar 7, 2014
Messages
9,975
Firstly, as a numbers freak, I welcome this site, it's quite fascinating and I look forward to checking in on its projections.

I would have a question regarding how you have verified its accuracy. Basically, I am wondering if you've backtested it. When building algorithms to project future outcomes, I would say the only way to be confident in future performance would be by running blind tests of previous outcomes using the model.

As others have pointed out, some of the constituencies are likely inaccurate. Much as I would love to think Joan Burton only has a 4% chance of election in Dublin West, she's at evens in the betting markets so factoring in Paddy's enormous over-round, the market gives her about a 40% chance of election.

Again, much as I'd love to think Kelly has only a 3% chance of election - and much as I don't believe either him or Burton will win election - he has more than the 3% probability you give him. He's at 1/3 to win a seat.

Howlin holds arguably Labour's safest seat, so a 1% probability is.... bold. The market has him at 1/10 reflecting, after over-round, a true probability in the region of 85%. You have the Shinner as a 100% probability; he's 2/1 in the markets.

All of which suggests possible structural issues with the model in terms of calculation of Labour candidates probability.
 

Seanie Lemass

Well-known member
Joined
Nov 26, 2010
Messages
20,159
Interesting!

You now need to follow Silver's other interest into sports, and advise us on betting for the GAA championship :)
 

crossman

Well-known member
Joined
Feb 16, 2011
Messages
1,573
I like the idea but it is giving some crazy results e.g. SF to get a seat in Cork SC with McGrath only a 30% chance or Nash with 0% in Louth when he is at least in with a shout.
 

Seanie Lemass

Well-known member
Joined
Nov 26, 2010
Messages
20,159
@ireallyshouldknowbetter

Your observations regarding betting odds and their predicted outcomes would be interesting to test against the results as some of the site's predictions throw up some good outside bets, including one you have noted.

That was how people used Silver's PECOTA ratings as factors in betting on baseball.
 

Dame_Enda

Well-known member
Joined
Dec 14, 2011
Messages
52,092
I recommend you consult electionsireland.org so you can see how national swings differed from constituency swings and factor that into your algorithm.

The recent Irish Times poll showed married people were far more likely to vote FG, whereas singles were voting for SF and Indos. As it happens Wexford is 51% single but its also probably older than average. On the other hand its 63% rural which contributes to a resistance to electoral change.
 
Top