France presidential election 2012: prediction thread

[big bad fab]

As Gael will lock the official thread tomorrow and before the results thread is dutifully active, here is the prediction thread.

And here's mine:

Guys, I've posted on my blog the last graphs by pollster. May be of some interest if you want to try to predict all this stuff.

If you just follow the trends of my Big Bad Tracker, this should be something like this:
Hollande 27.8 -> 28
Sarkozy 27-> 26.5
Le Pen 15.7 -> 16
Mélenchon 13.8 -> 13.5
Bayrou 10.3 -> 10.5
Joly 2.4 -> 2.5
Dupont-Aignan 1.4 -> 1.5
Poutou 1.1 -> 1
Arthaud 0.5 -> 0.5
Cheminade 0.1 -> 0

Now, I've tried to add effects from poll and media momentum, mobilization, electoral sociology and its consequences on turnout, tactical vote in a "useful" way. And it gives me my personal prediction:
Sarkozy 27.6
Hollande 26.7
Le Pen 14.8
Mélenchon 13.8
Bayrou 9.6
Dupont-Aignan 2.5
Poutou 2.1
Joly 1.9
Arthaud 0.8
Cheminade 0.2

Of course, I'll be completely wrong, as usual when it's about French elections

[big bad fab]

You may be fair right, Swedish, as tunrout and leaked numbers are bad for Sarkozy and good for Hollande...

[big bad fab]

Hash hasn't reopened the official thread, so I'm posting here just to tell you that I've published a tracker on vote transfers (reports de voix) on my blog

So you can make your own calculations...

Of course, today's Big Bad Tracker is also published !
55.11 / 44.89

With polling datas on vote transfers, applied to real results:
54.59 / 45.41

And daily IFOP has a first 55 / 45 this week.

[big bad fab]

27 Hollande
26.5 Sarko
18.5 Le Pen
13 Melenchon
9 Bayrou
3 Joly

28.5
27
18
11
9
2

Pas mal.

Yeah, quite.

As for me, I was definitely too conservative in my prediction and failed to grasp any of the major trends at their real level. Though, at least, I didn't put Sarko ahead.

GO TO HELL !

At least, my tracker was right, though I didn't evcen trust it

[big bad fab]

Maybe you can re-post these graphs in the Official Thread, Hash has just unlocked it

And add another point on the graph for the tracker with vote transfers ? As you wish...

[big bad fab]

On my blog, I've also calculated totals of spreads (or deviations) for each candidate BUT in a ratio with the result of each candidate: a "mistake" of 0.5 point isn't very important for Hollande, for example.

Harris has published the best last poll, in fact: good order, quite fine results and, especially, least bad numbers on Le Pen and Mélenchon.

For me, it's a surprising winner, rather than TNS-Sofres, OpinionWay, IFOP and IPSOS, though they have done quite well, but each with something wrong. But Harris was already the best for the PS primary... keep that in mind for the next years to come...

CSA, LH2 and BVA were clearly less good than the others.

[big bad fab]

BUMP

This thread can welcome any discussion on Sarkozy's reelection or Hollande's greatness...

[big bad fab]

From the official thread:

BREAKING NEWS !
2012 Big Bad Tracker #55 - 4 May 2012   


My very last tracker !!!       

With raw datas :
Hollande 52.95
Sarkozy 47.05

With vote transfers datas applied to real first round results:
Hollande 52.97
Sarkozy 47.03

No, I haven't cheated.

Simple average of all the last polls of each pollster, without weighting: 52.75 / 47.25
Weighted average of all the last polls of each pollster: 52.70 / 47.30


And my own prediction: Hollande 52.2 / Sarkozy 47.8