How to Predict the FIFA World Cup

Currently, England are 200-1 to win the FIFA World Cup 2018 in Russia this summer. To explain what that means to those of you unfamiliar with bookmaking lingo, if you put a £10 bet on England to win, at the end of the tournament you will have lost £10.

I am not a gambler, but I know enough to guess that a bet, when all is said and done, is a punt: you are supposed to enjoy it. So the psychology of this is such that even if cold, hard logic suggests that Brazil has a far higher chance of winning, a self-respecting England fan might find it hard to go and put £100 on Brazil to win it. An England fan is more likely to put their money on England, then spend the summer holding Pandora’s box upside down and furiously shaking it, wondering why nothing is coming out. Bookmakers know this in group optimism bias, a version of the availability heuristic, and adjust their odds around the world accordingly!

We’re mere mortals and cannot trust ourselves: we need a computer to sort this Russian World Cup out. Computers are objective, impartial, data-driven and prone to fail only under the malign intervention of cyber-attacks from foreign powers like, errr… Anyway. Not only is the computer not subject to ingroup bias, to assure objectivity one notable incarnation comes from a nation that doesn’t even call the game football. “Soccerbot” uses “Soccermatics” to read current odds and team performance data, calculate key metrics and predict upcoming matches. It is claimed that if you had followed the bot’s predictions, the Soccerbot – I’d like to imagine that the bot’s American developers had enough soccer “street” to give it Peter Crouch’s body – would have left you 1,800% up over the bookies over the past 3 years. (This return does involve a huge number of games, so the relative return per game is much more modest. But, nonetheless, it does suggest the predictive power of the bot is impressive).

Still, what does a robot really know? And don’t get me started on metrics and league tables. How could a computer predict the sorcery of a Gareth Bale overhead kick, or the Antigonian fatalism of a David Beckham underhand kick? Magic or genius, or something, somewhere in between, turns games. Sometimes the gods, or even the hand of God, intervenes. And we have hope and luck, and luck and hope: we know from football history that Terry Butcher can secure a goalless draw with Italy with a bloodied headband, goalkeepers can play FA Cup finals with broken necks. And we know Russian linesmen (or at least that Russians will be controlling the cameras for goal-line technology). Would a supercomputer predict any of that? I don’t think so. Who predicted Trump (aside from that strange, Bulgarian granny)? Who predicted Brexit? Who predicted they’d change Marathon to Snickers and Boris Johnson would be Home Secretary? The host nation is, after all, famously a mystery wrapped up in an enigma. The universe is one, big, capricious ball of hypothesis. This could be England’s year…

Hope means that we know all of this when we place our bets. If maths, bots, and experts can’t predict the World Cup, we need some other method of divination. And in recent years the predictive power of a range of different animals has loomed front and centre of media sports coverage. How does this work? In brief, animals either wear flags of competing countries and engage in some sort of race, or they nibble a bit of food or attach themselves to something with one or other competing nation’s flag on it (first nibble equals winner). Or, in a rather unsavoury version, choose to defecate on the prospective loser’s flag to signal the outcome. All hail the experimental method!

A few words about these animal predictions. First, they only really come to the fore when a tournament is in the knockout stages. A cynic might suggest this is because there are more gaps between games and there needs to be some sort of guff to fill the media vacuum. Or it may be that previous rounds have literally sorted the sheep from the goats and weeded out species with sub-standard psychic abilities. Animals with high psychic power: octopi (or is it octopusses?), otters, bears (polar, but not grizzly), meerkats.

Frankly, I’d be suspicious of meerkats given that they are clearly in bed with insurance companies which, as we know, have their own supercomputers plugging away at stats. Animals with low psychic abilities: chickens, jellyfish and wildebeest. Wildebeest are particularly bad. This list of animals with poor psychic abilities is compiled from personal observations but that feels entirely legitimate given that it is based on my own psychic instincts which I challenge anyone to contest. The industry around animals predicting sporting outcomes is just crying out for a decent randomised control trial (RCT). Why has no one done this yet? UK Research and Innovation (UKRI), you are frittering our money away.

A European supercomputer has England going out to Germany 3-2 in the quarter finals. That, frankly, feels like a respectable result. But perhaps it’s neuroscience, not these flaky computers, that will help us to predict the outcome when all is said and done. So here, for nostalgia and perhaps vindication through posterity, is Gareth Southgate’s sad face once again. But hope springs eternal so I found a scan of Gareth’s happy face too. Maybe, just maybe, we’ll see it this summer. But don’t bet on it. Gamble responsibly.

2016, “A Kick in the Neuros”: England vs Germany and penalties…