Jack I posted this on FLL's stats thread last night:

Asaparagus I would argue that Jack's stats support the Cher plunge as he has her as the biggest percentage (?) faller. Momentum is everything in these things imo and once you have lost it if you don't have a big core vote you are toast.

Cher vote = emperor's new clothes.

The trick with all these stats is gleaning all the information you can from them. I have said "percentage" but I think they are just actual numbers?

I'm a lot cheesed off with survation. They emailed me the poll at 8.29pm tonight that had Cher well bottom. Poll line closed by then I think. That will teach me to be honest and tick the "yes I gamble on this" box as clearly they delay the email for those who tick that box. Grrrrr [>o][>o][>o]

cher is a certainty to survive next week

Looking at your momentum stats Jack, 1D are only positive twice?

That fits with what I've found elsewhere. They're not winning over many, if any new fans.

Hello chaps, the original specials data nerd is back. Some truly great stuff on this thread but I feel you have yet to crack some hurdles.

I have used most of the tools going over the last 3 years - Polls, Google Trends, Twitter, Streamgraphs, Socialmention, Facebook Analysis, Trendrr (the best by a mile if used correctly but is sadly now a pay site) and many more.

I now think the key to this is momentum. Look at last years stats (http://xfactor.itv.com/2009/news/viewer/item_100312.htm) and you can see that it was about now that Joe pulled clear. I have a lovely graph overlaying Twitter mentions for all contestants for the whole 3 months against the actual voting figures - the relationship is non linear. However momentum shifts predict things a lot better.

After all this work I still feel I can predict the results better than my statistical models can - hence I don't bother much (except BGT where Twitter rules).

This is the week that the winner will show themselves - whoever blows them away with a late slot will win it imo. (and underestimate Mary at your peril).

all the info on the "order thread"  nick

Hi Occam good to see you back. I always thought your stats predictions were either spookily accurate or miles off which reflects the inconsistent nature of the beast.

What has puzzled me this year is the accuracy of fll's twitter stats given their demographic.

I agree with you and have said elsewhere that the Big Mo (as they call it in US elections) is everything. I will concentrate my number crunching on trying to establish this in future weeks.

I also agree that gut instinct is as good as anything. Just watch it with an open mind. Pick up the phone within 10 mins of the last act (like most do)and see where your vote goes.  If you can be truly objectve you won't go far wrong. 5 hours of number crunching to back up the gut feeling doesn't go amiss mind.

matt had pimp slot week 1, one d have had it too, mary hasnt....interesting, she may be due it this week, though with two songs it may make far less of a difference

seven left, two songs each, this is the week when it all comes together

when we see really who can possibly win this thing, and surely, surely, that doesnt apply to katie, mary, wagner or cher

I have listed weekly singing positions of remaining contestants on Order thread.

The way these shows pan out these days (esp XF, AI) I wouldn't assume anyone is "due" anything though.

Jack - you and Figgy agree that Cher is the value bet to go, but Mary is some way adrift on both sets of stats...

Surely this points to Mary finishing bottom of the vote?