If You Ain’t First You’re Last

Eric Schmidt
Analyzing the World Cup using Google Cloud
4 min readJul 14, 2018

Only two more matches remain as we continue our analysis of World Cup using Google Cloud.

The third place match in World Cup play has come under increased scrutiny with each passing tournament. More and more, with each subsequent World Cup, the question has become: why in the hell is the third place match still a thing? The short answer: there’s too much money in it not to have one. At least that was the prevailing theme prior to the most recent third place match in Brazil four years ago (before the Netherlands defeated host country Brazil 3–0)

So what is the mindset and/or strategy for each team coming in? Hard to say…

Taking absolutely nothing away from the fervor surrounding the Belgium National Team after its semi-final loss to France — but it seems as if the England defeat has taken a much greater toll on its players, coaches and rabid fans. Belgium was beaten 1–0 by a well rounded effort from France that included a clean sheet. England looked to have the upper hand — took a 1–0 lead — then allowed a late second half equalizer to Croatia before folding down the stretch in extra time. There is an argument to be made that Belgium was defeated while England may have painfully let one slip through their hands — taking nothing away from the brilliant effort by Croatia in the latter stages of that match.

Perhaps the most compelling element to this particular third place match is the race for the Golden Boot Award. England’s Harry Kane currently leads with six World Cup goals with Belgium’s Romelu Lukaku in second place and four goals. Watching the two players meet in the third place match could produce a shootout of sorts between the two respective stars. Lukaku might be looking to finish strong after being panned by critics after the semi-final loss to France.

The third place match sometimes offers a rematch from earlier in the tournament, which happens to be the case here, as Belgium and England met in Group G play a few weeks back. But extrapolating anything from that match is tricky. Many people feel England, who had already qualified, tanked on purpose in order to get a better draw in the bottom half of the bracket for the round of 16. The Three Lions put reserves in for the majority of the match, lost 1–0 to Belgium, and didn’t seem the least bit concerned about a second place group finish. In fact, the whole world seemed to know ahead of time the tank job was coming.

Predictions

Our models are a bit split on this one — the majority favor the Belgians, but our player centric model likes England.

Player based

  • England: 54.5% (1.23 xG)
  • Belgium: 45.5% (1.20 xG)

Team based

  • England: 29.6% (1.42 xG)
  • Belgium: 70.4% (2.00 xG)

Both

  • England: 32.4% (1.48 xG)
  • Belgium: 67.6% (2.08 xG)

ELO

  • England: 37.7%
  • Belgium: 62.3%

Our team based projection is on the other end of the spectrum, giving the English the lowest chance of any of our four models (and lower than any of the models mentioned below). From a feature perspective, this is because many of our team level features are xG based and England’s xG numbers (particularly on the offensive side of the ball) pale in comparison to those of the Belgians. Our original model finds a middle ground, though it’s much closer to the team based projection as those features have generally proven to be more powerful. Finally, our ELO based model also gives Belgium a healthy advantage.

Google Search favors Belgium — with at 44% probability of a draw in 90 mins.

538 shows the match as a near pick ‘em

Bing also likes close odds — but the slight lean goes toward Belgium

As my daddy said, “If you ain’t first you’re last”. Enjoy the match!

--

--