World Cup Predictions Verified

All that was missing was the water for the players to plunge into. Still, referee Howard Webb must be a fan of the aquatic and artistic sport of diving, and he must have thought grass a reasonable substitute for a pool, because he ran over to judge appreciatively each performance of a player splaying himself onto the green.

We watched him write his ratings of form and originality down, and then show his judgments (9.7?, 9.5?) to the players on a little yellow card. Although he whipped out his card fourteen times, the TV cameras never caught what he had written. But all but two of the Dutch players got to see them, so maybe these men will tell us all later what Webb was thinking.

It’s true that the players mixed it up a bit—elbows, feet, and spikes were flying—but the misdeeds were equally meted out by both teams. Still, Webb thought it was all too much and, perhaps because of a nervous tick, quickly reached for his card after each “incident.” It only became absurd when the Dutch began to be awarded yellows for the Spanish players kicking themselves.

Anyway, it’s all over. And it will be four years—four years!—before the Netherlands can seek any kind of revenge. Or for the USA to finally put in an acceptable performance. Or for Germany to realize it’s potential. For France to remember why they play, for Italy to invent new hair styles.

But in four years it will all take place in Brazil, where it will require a miracle for that country’s team to lose. Meet me back here in four year’s time to see whether this prediction will be verified.

Meanwhile, we can look at how well other predictions fared.

First, Pulpo Paul, that slimy, eight-legged sea creature. He ate the Spanish clam, thereby forecasting Spain’s first-ever victory. Paul “The Psychic” Octopus has a good record, but, as we discussed a few days ago, his and some other prescient pets’ performances are not that unusual. The only question that remains is: will Paul be eaten, released into the wild, or stuck in a dark closet and forgotten? I say the first.

World Cup Statistics and Predictions for the Final

I created three models (for public consumption), which at least show that dry statistics can best emotions and produce predictions that beat wishcasting (see my USA-Ghana forecast for an example).

Prediction 1: Strictly using the observations from previous World Cups—and using no data from this one—I predicted the eventual number of Goals Per Game (GPG). The prediction was 2.38, or somewhere in the window 1.83 to 3.10.

There were 143 goals this Cup, and 64 games, which makes 2.23 GPG, which I hope you agree shows an accurate forecast. Our conclusion that scoring in World Cup games is becoming more difficult has been verified.

Running this model for 2014 predicts 2.30 GPG, in a window 1.79 to 2.97. Meaning that scoring will become even more difficult in the next Cup.

Prediction 2: The number of men sent off per game under the ignominy of the red card has been increasing. Our prediction was, over all 64 games, that about 0.44 (in the window 0.27 to 0.61) men per game would be sent off. I counted 17 (did I miss one?) early showers.

That makes 0.27 per game, which is just on the low end of the forecast. So, despite the final being a bit of a melee, this Cup was slightly more gentlemanly than the last.

Taking the model (without adjustments) to 2014 predicts 0.41 (0.23 to 0.61) men per game will be sent off. In Brazil. Seem likely to you? Does to me.

Prediction 3: The final game has seen lower and lower scoring through the years. We predicted that this trend would continue for this final, and said that there was a 47% chance that the winning goal differential would be just 1.

This is, of course, exactly what we saw. The model also predicted a 30% chance of a tie (which it nearly was), but only a 23% chance that the win would be 2 ore more goals.

The same model now predicts a 29% chance of a tie in the final in 2014; a 51% chance of a 1-goal victory, and only a 20% chance that the winning teams blows its opponent away by 2 or more goals.

But, like was said above, we have to wait four full years!—an eternity!—before witnessing whether these predictions will hold.