World Cup predictions are like noses: almost everyone, it seems, has one. And even the algorithms like Microsoft’s Bing are taking the field.
Microsoft said Tuesday that it is busy collating all sorts of factors, from social to sport, to build a comprehensive picture of the results of the group stages. At this point, however, Microsoft isn’t willing to go beyond the opening games. The World Cup itself kicks off Thursday with host Brazil taking on Croatia at 1 PM Pacific time.
Bin began trying to predict the outcome of “reality” shows like American Idol in 2008, after Bing researchers found that that search queries could predict the likelihood of snow days in the Seattle area. Microsoft released Bing Predicts in April and claims to have successfully predicted this year’s winner, Caleb Johnson, as well as the correct results each week since its launch.
For the World Cup, however, Bing couldn’t rely on social sentiments. “Rather, the actions and performances of a handful of individuals solely determine who wins, loses, or ties,” the Bing team wrote.
“For the tournament, our models evaluate the strength of each team through a variety of factors such as previous win/loss/tie record in qualification matches and other international competitions and margin of victory in these contests, adjusted for location since home field advantage is a known bias,” Microsoft said. “Further adjustments are made related to other factors which give one team advantages over another, such as home field (for Brazil) or proximity (South American teams), playing surface (hybrid grass), game-time weather conditions, and other such factors.”
And if that isn’t enough, Bing checks its own predictions against Las Vegas. Presumably, if its own algorithm gets too out of sync with the betting markets, Microsoft will adjust it.
A quick exit for the United States?
So how will the United States fare in group play, within the “Group of Death,” Group G? At this stage, it looks dire:
Bing breaks down the results into three possibilities: win, lose, or draw. Naturally, an even split would be 33.3 percent. So while the American chances against Ghana look pretty good, Microsoft says that the United States will finish dead last in the group and exit the World Cup without notching a single point. (Three points are awarded for a win, one for a tie or draw, and none for a loss.)
Other predictions are equally harsh
Of course, Bing isn’t the only voice providing World Cup predictions. While Google has yet to wade into the World Cup predictions game, data wonks have probably already seized on FiveThirtyEight’s simulations, which ran 10,000 simulated games and came up with this outcome for the Group G stage:
Note that ESPN owns FiveThirtyEight, so ESPN’s algorithms essentially use FiveThirtyEight’s model.
Bloomberg paints a different outcome, but again, the United States finishes third in group play. As a bonus, however, Bloomberg’s data nerds tell us who they expect the hoist the Cup itself.
And of, course, there are other algorithms that slice the data different ways. SBNation’s chance-quality model puts the United States dead last in the group stage again. The Economist‘s World Cup model is, as you might expect, virtually impenetrable.
But what Chris Anderson and David Sally note in their book The Numbers Game is that about 50 percent of a team’s success is based on luck alone. So much for the model.
In any event, every team enters the tournament with hope. And all the United States needs is a lucky bounce, a key injury, or a opponent’s overconfidence to break through. and when the odds are against you, science be damned. After all, an octopus could do it.