US statistical genius Nate Silver got UK 2010 election results wrong

In the aftermath of the US presidential election, polling guru Nate Silver has been enjoying widespread plaudits having called the outcome in each of the fifty states correctly. However, it’s worth remembering that with the 2010 General Election in the UK, Silver didn’t do quite so well.

In the run up to election day, Silver anticipated heavy losses for Labour, and big gains for the Liberal Democrats. One of his predictions had Nick Clegg’s party with 101 seats — a gain of 39. Of course, what actually happened was that the yellow party lost five seats, in an anticlimactic comedown from “Cleggmania”.

And whilst Labour suffered a bruising defeat, their final total of 258 seats wasn’t quite as bad as Nate Silver had indicated. His doom-mongering calculations had Labour on 214. So there are 44 red MPs feeling very grateful that he wasn’t right.

On his FiveThirtyEight blog, Silver wrote:

“its very hard to see how [Labour] could thread the needle in such a way that the outcome wouldn’t be devastating to them.”

Scrapbook wonders whether “Mystic Nate” can do better in 2015.

8 Comments

  1. Tom says:

    Well not really a surprise given how much less polling there is in the UK on a local level in comparison to the state by state stuff done in the US. Thought Clegg was looking like doing that well until the last few days before the election anyway?

  2. RJGFox says:

    Old News – and echo comments above. Intensity and scale of polling in the USA is totally different to here.

    Don’t think an error like that in any way wipes out the significance of what he has done – predict 49 out of 50, then 50 out of 50, in two elections, just by doing the easy things.

    Amazing.

  3. David says:

    As has been said above, the two cannot be compared. It’s all about the polls:

    In the US, every swing state is polled dozens of times by numerous pollsters. The equivalent would be repeated polling of every marginal constituency – instead, there are just a handful of national pollsters whose results are applied based on an assumed (but often very inaccurate) Uniform National Swing.

    It’s no surprise a polling aggregation service – which is essentially all Fivethirtyeight is – is less accurate when the quality of the data is so much worse.

  4. Bellaire says:

    Seem to remember reading about this 30 months ago, on the 538 blog itself. (A post-election “What went wrong” article)

    Keep up, PS! Where’s the news and gossip?

  5. Jimmy says:

    It’s unfair to call it just an aggregator. Silver predicted Florida for Obama despite the polls, also his prediction of a 2.5% overall lead in the popular vote, much higher than the polls suggested, was spot on. The boy done good.

  6. Martin says:

    Let’s be honest – it’s hard work making accurate assessments of the UK voting pattern compared to the US because we have more than two parties combined with a crummy electoral system that forces rational voters into making a ‘least worst’ decision.

  7. Nicola Prigg says:

    I mean Nate Silver was testing out a new stats model and lets be fair unlike the US, there is not a lot of local polling done at a constituency level.

    Also, if you look at the numbers he was putting into the swingometer, notice what percentage of the vote he put the parties on.

    The vote share that might have led to the LibDems taking 101 seats was 29.1% with Labour on 26.9% & Conservatives on 34%. In case you didn’t notice that didn’t happen. Although at one point in the campaign, polls showed that those numbers might have been possible.

    Those are the same numbers that his swingometer gave Labour 214 seats. Again, notice those numbers never materialised and Nate Silver never said that they were going to happen.

    In his article he was comparing a swingometer with a uniform national swing with a different calculation.

    There weren’t predictions.

  8. Really don’t visit live until eventually you’ve exercised on a demo bank account. James taped him or her self and the display for virtually any weeks precious time since they assessed a market segments and also become a member of tradings instantaneously, providing a new complete commentary within the products they has been considering as well as the reason why he / she needed the transactions the guy managed. There can be nothing at all cherry chosen. James Edward is usually a experienced trader which made dunes last year whilst commenced operating certain are living investor instruction classes.

Post a Comment

Your email is never published nor shared.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

  • Follow us on Twitter