Hindsight bias in football

The 2012/13 Champions League second round draw pitted Manchester United against Real Madrid – a draw seen as inevitable by many

A Twitter search for ‘knew we’d get Madrid’ in the immediate wake of the Champions League second round draw revealed a small army of Manchester United supporters who claimed to have correctly predicted their opponents for the first knockout stage*.

Of course, a search of ‘think we’ll get Madrid’ revealed a much smaller sum of people. Prior to cup draws, most people acknowledge their random nature; that apart from a few pre-defined exceptions, your team has an equal chance of drawing any other team in the competition.

Yet after the result, people tend to see the outcome as being more predictable than it was before it took place. Psychologists call this hindsight bias, and cup draws are a very transparent example of this cognitive tendency.

Indeed, our propensity to deal in narratives – that sport and football follows some sort of script – reinforces this tendency; not just for cup draws but for matches themselves.

Bookmakers’ odds reflect football’s uncertainty. Last season, in only 12% of matches did the favourite have more than 75% chance of winning**. But we often describe results to have been more certain than they actually were, even though in most games both teams have a relatively decent chance of victory.

Chelsea and Liverpool’s Champions League triumphs are examples of results described as ‘written in the stars’ after the event; in reality their chances of victory before the match were far more uncertain.

Admittedly, some might think this is a very cold and detached view of the game. But for anyone wishing to objectively analyse the sport – journalists, club analysts, managers – this perspective is vital. No outcome in football is ‘inevitable’. Otherwise, why would we bother watching?

*Twitter only shows search results from the past seven days, but example tweets include this, this and this.

**Premier League, odds from Bet365, with data from football-data.co.uk. Not adjusted for over-round.