I want to test the Wisdom of Crowds theory that I talked about some time ago, so what better example than the fast approaching Irish General Election? I decided yesterday evening to spend some time looking at Paddy Power’s betting on the outcome, candidate by candidate.

Based strictly on the odds being offered and the number of seats available, the analysis breaks down this way:

Fianna Fail: 72-73*

Fine Gael: 49-50

Labour: 16-18

Green: 9

Sinn Féin: 8

Progressive Democrats: 2

Others: 8

(In two constituencies, Wicklow and Kerry South, the odds were the same for last place, so I have shown a range instead of a single number).

If this were true, high profile casualties would include Ruairi Quinn, Joan Burton, Donie Cassidy, Jackie-Healy Rae, Niall Blaney, Tom Parlon, Liz O’Donnell, Martin Mansergh and Michael Woods.

This scenario is something of a nightmare for the two major political alliances. It leaves Fianna Fail and the PD’s with 75 seats and a “rainbow” coalition of Fine Gael, Labour and the Greens with 77 seats: well shy of the 84 seats needed to form a new government. Both coalitions would need the support of independents and possibly even Sinn Féin.

If this prediction comes true, the only viable administration would be a Fianna Fáil / Labour alliance (and how long that would last is anyone’s guess).

In any case there is plenty of room for manoeuvre. The number of listed candidates that are almost certain to win is 75. 190 other candidates will have a fight on their hands if they are to win one of the remaining 91 seats**.

More detailed information on each candidate can be found in the attached PDF. Let’s see what transpires tomorrow, shall we?

Election Predictions 2007

* Rory O’Hanlon (FF) is already deemed elected since he was Ceann Comhairle (chairman) in the last Dáil.

** My calculations are based on a “definite” being a bet of 10-1 on (91%) or greater, and a “close call” being less than 10-1 on but greater than or equal to 4-1 (20%).

Update: I had to revise my calculations because I was previously showing probabilities of 2000%, etc, which is ridiculous – a probability can not be greater than one!