I’m a Watford fan. You may already know this about me.
Being a football fan, particularly being a fan of a smaller club, is a tribal thing. A thing of faith. A secular religion.
Every so often The Algorithms suggest that I should be interested in other football teams. Google offers me news articles about competitors and rivals. Twitter suggests I should follow other clubs.
This is akin to Google and Twitter saying “I see you follow Jesus. Why not follow Mohammed?”
Yes, really.
After the latest incident with Twitter suggesting that I follow a Crystal Palace (wash my mouth out)-related account I posted about how I seemed to have broken their algorithm, only partly in jest.
Someone I know at Twitter posted a potential explanation.
If you looked at the data of normal football fans, you’d probably see the tribalism. I follow Watford-related accounts. I don’t follow Crystal Palace (wash my mouth out) accounts. Most football fans on Twitter would be the same.
But alongside normal football fans there are Football Journalists. Football Journalists are a different tribe who will follow lots of accounts for lots of teams because, well, that’s their job in 2021.
Trouble is that these Football Journalists will be often followed by big audiences of football fans. They are in social network terms “People of Influence”. And so because people of influence in the broader world of football follow lots of accounts from lots of teams, the Twitter algorithm assumes that is desired behaviour and therefore recommend stupid accounts from stupid people who support stupid teams to partisan (and undoubtedly immature) football fans like me.
It’s a neat little encapsulation about how why algorithms alone are not enough. Algorithms are stupid and don’t actually understand human behaviour. They just turn it into mathematical models. They miss the nuance. Sometimes as embarrassingly wide of the mark as many of the strikers at my beloved club over the years.