X Factor 2012: Assessing the evidence on the new voting system

In the first show of this year, when we realised that the phone lines being open at the start of each show meant that Dermot was telling us the full running order upfront, we thought this would be a great help in figuring out what producers were up to before even the first act had performed.

Sod’s law, then, that the phone lines being open from the start should also have thrown into confusion all our old assumptions. It took only a show or two to make clear that the “early bad, late good” heuristic which has served us so well over the years was going to need updating. But to what, exactly?

We’re now four weeks into the new system – enough for us to start drawing some tentative conclusions. To start with, here are the stats for the first four public votes over the eight series that have had at least 12 acts – it shows the position in the running order of acts finishing in the bottom two (or three, as applicable).

2012
Week 1 – 9, 12 (of 13)
Week 2 – 7, 12 (of 12)
Week 3 – 2, 9 (of 11)
Week 4 – 2, 8 (of 9)

2011 (no public vote week 1)
Week 2 – 1, 5 (of 12)
Week 3 – 3, 7 (of 11)
Week 4 – 3, 5 (of 10)
Week 5 – 1, 4, 6 (of 9)

2010
Week 1 – 1, 10, 12 (of 16)
Week 2 – 1, 7, 12 (of 14)
Week 3 – 2, 7 (of 12)
Week 4 – 3, 9 (of 11)

2009
Week 1 – 1, 2 (of 12)
Week 2 – 4, 9 (of 11)
Week 3 – 3, 7 (of 10)
Week 4 – 4, 7 (of 9)

2008
Week 1 – 1, 9 (of 12)
Week 2 – 3, 4 (of 11)
Week 3 – 1, 2 (of 10)
Week 4 – 1, 2 (of 9)

2007
Week 1 – 1, 9 (of 12)
Week 2 – 4, 6 (of 11)
Week 3 – 2, 7 (of 10)
Week 4 – 3, 4 (of 9)

2006
Week 1 – 7, 8 (of 12)
Week 2 – 6, 8 (of 11)
Week 3 – 2, 7, 8 (of 10)
Week 4 – 3, 8 (of 8)

2005
Week 1 – 1, 3 (of 12)
Week 2 – 2, 3 (of 11)
Week 3 – 1, 4 (of 10)
Week 4 – 2, 4 (of 9)

Over the seven preceding series, we’d had only one act singing either last or second last (Ashley McKenzie, week 3 in 2006) hitting the danger zone in the first four public votes. That’s now happened three times in four weeks under the new system (Carolynne Poole, Melanie Masson, Jade Ellis). In all, six of the eight acts who’ve hit singoffs so far this year have sung in the second half of the show.

Jade’s treatment on Saturday suggests that producers may be getting to grips with this new reality. Having attempted to nobble District 3 and Chris Maloney from early slots, they stuck their week 4 target, poor old Jade, second-from-last.

But if late is the new early, we can’t equally say that early is the new late, as the other two of the eight singer-offerers this year both sang second. Going close to the start of the show is clearly not the near-guarantee of safety that going close to the end used to be.

So far, the safest place to be has been early-to-middle. It’s possible to dream up theories about why that might be – maybe that’s when most people are watching given how late the show has been on; maybe some viewers feel like after five or six acts they’ve seen enough to vote for their favourite so far – but these hardly seem compelling. With limited data so far, it might easily be coincidence.

So if we no longer have any easy heuristics based on position in the order, what else should we being look at?

The memory hole and the big-name sandwich

All credit to Sofabet commenter JScouser2002 for quickly spotting that if where you are in the order matters less, who you’re next to in the order may matter more.

The memory hole has long been a staple of our analysis on Sofabet – when a vulnerable act is immediately followed by a big-name act, they’re quickly forgotten. It’s a tactic that in previous years worked especially well in combination with an early slot, when lines didn’t open till the end of the show. But it still seems potent. Notably, MK1’s bottom two appearance came in the week when they were immediately followed by the exceptionally powerful VT about Jahmene’s troubled upbringing.

Unfortunately, it’s not always immediately apparent from the running order announcement where the memory holes will be. At the start of show 3, for example, you wouldn’t necessarily have expected Kye to be memory-holed by District 3, but an unexpectedly strong performance from the boyband made it feel that way as the show unfolded.

It’s also looking like coming after a big-name act may have become more unhelpful under the new system. District 3 and Melanie in week 2 both hit the singoff after immediately following one of the two more talked-about acts, in the shape of Lucy and Rylan respectively.

Logic thus dictates, and many have noted in the comments, that it may be especially disadvantageous to be sandwiched between two big-name acts. Carolynne’s surprise singoff appearance in week 1 came after she was wedged between Ella and Jahmene.

Again, Jade’s treatment appears to indicate that producers have come to the same conclusion – when they set out to get her on Saturday, they stuck her in a Jahmene-James sandwich. Interestingly, this is exactly what they’d done in week 3, when the triple-J of Jahmene-Jade-James sang in positions 3-4-5 compared to 7-8-9 this week. Perhaps producers were hoping to isolate the effects of early-to-middle running order from the effects of the big-name sandwich?

Is the older demographic now more powerful?

In a comment last week, TommySauce offers some fascinating thoughts about how the new voting may be playing out. One of them:

In the old days the “lines open now” used to immediately cause the system to clog and (certainly in our experience) you’d have to try multiple times to register your vote. Might the new system have changed this, spreading the load over the whole show? If so, uncommitted viewers who might have voted on a whim in previous years, but given up the attempt after being unable to register their vote, might be adding to the totals this year (will be interesting to see the total votes cast comparisons to last year under the new system). Might this mean some demographics – the older and technophobic are having their votes counted more?

The fact that all the acts most obviously aimed at a younger demographic – District 3, MK1 and now Union J – have hit the bottom two suggests there may be something in this. And if we give credence to the Star’s leak that had Maloney topping the first two votes, that would fit, too.

Is the previous week’s performance now important?

Here’s a theory for which we see a little bit of evidence, albeit tentative and inconclusive – does an act’s performance in the previous week’s show influence how well they do in this week’s vote?

The thinking behind this theory is that some voters may now vote for their favourite act right at the start of show, as soon as lines are open. If their favourite act did well the previous week, they may be more fired up to do so; if their favourite act disappointed the previous week, they may not.

Evidence? District 3 hit the bottom two in week 2, after their disappointing performance in week 1. MK1’s Glee-style performance didn’t land them in the singoff in week 2, as many expected – but they were there a week later, in week 3. Their singoff opponent in week 3, Kye, also had a dreadful performance in week 2. Many of us were surprised that Rylan avoided the bottom two in week 3 after a mediocre showing; could he have been helped by clocking up some start-of-show votes inspired by his fun-filled week 2 performance?

This week, District 3 were down to joint-favourites for the chop after the Saturday show, but escaped the singoff; could they have got some votes in the bank at the start of the show from fans remembering  their week 3 barnstormer? Similarly, Union J had a relatively forgettable week 3 – might this have led fewer fans to dial in right at the start of the week 4 show?

Of course, there are counterexamples (Rylan’s survival this week, to name just one); there are many other factors involved in each case, and it’s very early days to be placing any weight on this line of thinking. But it could be something worth bearing in mind. If there’s anything in it, you would think District 3 might be vulnerable in week 5 on the basis of the week 4 performances.

Agree or disagree with any of the above? What are your own tentative conclusions about the how lines being open from the start might have changed things? As ever, do keep the conversation going below.

8 comments to X Factor 2012: Assessing the evidence on the new voting system

  • Kevin O Reilly

    How could Barlow ever justify his comment when Jade Ellis has said so often how much she wants to change her life for her daughter’s sake? Laughable and disgusting treatment of a lovely lady from what I have seen of her. I hope she does well and gets her garden. The sing off showed her vocal ability. There does seem to be backlash of sorts and as ever trying to get an angle I think Lucy is about to get the brunt of it. I am waiting for the bookies to price up bottom 2 and I think 5/2 (hoping for 3/1) is worth grabbing.

  • AlisonR

    Suggested most of this in earlier posts already!

    Last week’s performance having an effect: http://sofabet.com/2012/10/21/x-factor-2012-week-3-post-mortem-mk-gone/ Oct 23rd 8:53.

    A third to half way through being the best place to be: http://sofabet.com/2012/10/14/x-factor-2012-live-show-2-home-maloney/ Oct 14th 3:08.

    Two thirds through to penultimate seems to be the worst place to be if the last one is a good alpha, whereas if the alpha is near the end but not last then anyone following is in a poor slot. The start is a good place to stick your betas: if you lose them you don’t mind, but you’re not particularly trying to (http://sofabet.com/2012/10/20/x-factor-2012-week-2-elimination-betting-overs-and-out/ Oct 20th 1:36).

    Jade was clearly the target this week – poor draw and song. They also partly covered her face, something which has been used to good effect before to deramp/lose an act (e.g. Kitty, Katie) – harder to emotionally connect. Given she was also trailing in the social media stats and chucked in a weak performance as well, she should’ve been odds-on to go. Although I did fear for James because the first half of his song was dreary and I wondered whether we would get an “Aidan” shocker. So, for TPTB, target hit – the first successful one I think. I expect the targets now to become easier to spot week by week. I would have thought that next to go will be a male as there are too many of them left. Will be checking social media stats on Thurs/Fri then will match up with the draw for Saturday.

    • You did, Alison. Profuse and grovelling apologies for inadvertently nicking your ideas here. As you’ll be aware, we do always try to be ultra scrupulous about giving credit where it’s due and linking to the relevant comment when we want to bring it to attention in a post, but sometimes I don’t remember where (or even, evidently, if) I’ve read something before. Rather embarrassingly, lightning appears to have struck thrice here. Props for concluding all this before we did, and sincere apologies again.

      Good point on the facial obscuring, too – something also used on District 3 this week.

    • Edie

      100% with you on the face obscuring thing AlisonR!
      This is something I have noted with the Australian version of XF, the less favoured acts get really limited face time when singing, masks, long-range shots, focus on dancers etc, which often seems to end with them in the bottom two.

  • Agree with every word of this. Good roundup. Not only have producers got to grips with the new reality – Sofabet has now too! Re: the above comment, I don’t think Andrew was nicking anyone’s ideas per se because this article’s basically a summary of the conclusions and consensus that people in the comments section (and the X Factor betting community as a whole) have been coming to over the past few weeks. So I don’t think that much individual credit needs to be given when so many people have been articulating the ideas outlined.

    • Kevin O Reilly

      as you suggest elsewhere Union J were most likely bottom of the vote, which maybe explains the need for notes being passed? What this suggests to me is the level of votes for the boy bands may be paper thin. I think one will go very soon and any decent price is worth considering.

  • Dizzy

    The older demographic is almost certainly becoming more powerful (as evidenced by the support for Baloney), but not for the reasons stated above.

    The time slot has had a major impact this year in terms of viewing figures. The peak audience is at around 9.00pm, tailing off by 10.20pm. This not only affects the voting numbers, but also who is voting.

    The younger demographic will have either a) gone to bed (if in early teens) or b) gone on their night out. This leaves just the stay at home parents and older generation watching the second half of the show to vote.

    That would help to explain the weighting towards acts in the front half / middle of the show, as well as the type of acts falling into the bottom two so far.

 Leave a reply...