The Algorithm That Can Predict Chart-Toppers


by Joseph Sarrington Smith

There’s no doubt it – technology is slowly taking over. Yes, we’ve been saying that for many years, but never before has this statement felt more true than when I heard earlier this week that there is now a machine that can recognise a hit song. Last year it forecast, with impressive accuracy, some of the most successful dance/pop songs to reach the charts. All results scored over 65%. You can almost picture the cigar-chomping suits sitting in the boardrooms, rubbing their hands with glee. I’ve even seen the written equation which the scientists came up with when they carried out their research. It made no sense to me whatsoever, but then again Maths was never my bag.

Furthermore, algorithms are also being used to specifically choose the tunes we want to hear on our iPod playlists. That way, it feels more personal. It knows what I like, not what the masses like. This is something the BBC started doing a few months ago. A daunting task, given the fact that a staggering 43,000 hours of music (although I have a feeling my record collection may surpass this) apparently exist in their archive.

However, I think there is a bit of an issue here…

Don’t get me wrong, computers are smart beasts. But surely it’s a bit depressing when an algorithm can predict what constitutes a future Number 1. Chart music has become too bland and formulaic. It seems like someone at some point just decided that there is a ‘thing’ that works, whether it’s to do with tempo, vocals, instrumentation, production etc. They wrote it, they released it and everyone else copied it. It would be great to see songwriters taking risks again in the pop hit department. We all know how rewarding it can be, but we shouldn’t be afraid to try thinking outside the box.

Be the first to comment

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.