Markov Chain Monte Carlo and the Eurovision Song Contest

  • Interesting part: "In 2009, it was reported that the 43 Azerbaijanis who texted in votes for Armenia in that year’s contest were summoned to the National Security Ministry to explain their actions".

    I'm surprised that this country was allowed to stay in the EBU after such a behavior. Really shocking.

    Lebanon was more or less thrown out of the competition for not playing by the rules (https://en.wikipedia.org/wiki/Lebanon_in_the_Eurovision_Song...), so I'm surprised the EBU did not do the same here.

  • For the 2012 contest, I assume that all friendship values remain the same as in previous years.

    Being in Europe, I'd say that the love for several countries has been in constant review over the past year. Just how and by how much will be clear after we do the exercise again, after Eurosong 2012.

    P.S. that title ("I've got Eurosong fever, Ted") is from a TV series called "Father Ted". I recommend.

  • Gotta love statistician logic! The increase in certainty is probably due to additional information gleaned from actually listening to the song.

  • Living in a multinational household Eurovision is always a fun evening of cheesy euro-pop and European geopolitics. I've always preferred Terry Wogan's humourous coverage of the event, which he's attributed to getting progressively more drunk as the night goes on[1]. I imagine this year that no matter how good the song, Greece will be getting nil point from Germany and vice versa. Of course, Cyprus and Greece get nil point in our household, thanks to the Turkish element. Likewise, the French get nil point from the English side. I imagine for Americans it must all be rather bemusing and strange.

    The post misses a substantial point about the Czech song Aven Romale[2], that many eastern european countries that would normally bloc vote would be against voting for a Gypsy song based on cultural reasons. I do like the point about the Azeris and Nagorno Karabakh though.

    [1] - http://www.guardian.co.uk/media/2009/may/15/eurovision-song-...

    [2] - http://www.youtube.com/watch?v=w3CgUmzl4sk

  • Very interesting, and not too hard to follow with little knowledge of statistics. I might even start taking an interest in the contest, despite the Netherlands coming up with terrible songs to enter every year.

  • Does anyone know what twitter (data) is predicting about the final? Twitter has done quite will in picking recent elections, sporting events etc

    The recent crisis in Europe should make the voting interesting.

  • slight sidestep but it would be interesting to see how the friendship factors have changed over the years. Could be fun to map on a news timeline to see how it evolves with the geopolitical events.

  • There's also a Million Song Dataset Challenge going on on Kaggle now: http://www.kaggle.com/c/msdchallenge

  • The Israel numbers seem off - they have a large issue getting votes right now, as evidenced by this semifinal when they (subjectively) had one of the top three songs but didn't make it to the finals.

    There may be a big change in public opinion, but I don't remember Israel being popular in Europe for at least the last few years...

    (From one of the very few Americans to actually know about Eurovision, much less watch it.)

  • In a similar note... ( American Idol prediction based on social media ) http://news.ycombinator.com/item?id=4015553

  • tongues, not AI.

    God says... you're_wonderful fun CIA these_cans_are_defective It_grieves_me it'd_take_a_miracle little_fish now_that_I_think_about_it talk_to_my_lawyer the_enquirer spending dufas biggot Dudly_Doright car middle_class revolution clever hype