The Google cloud platform won the semi-finals of the 2014 World Cup in Argentina, Germany!

As I am a football fan, some days ago, Google used its cloud platform to predict the World Cup eight to four games and achieved a 75% accuracy rate, which caused me a lot of vibration. Although I have been hearing about how powerful big data is to predict and view trends over the years, this time it is even more shocking because many people are watching and trying to predict the World Cup, at this time, the Google cloud platform was very promising.

Of course, there is one thing that must be said: The World Cup has always been the most popular in group stage matches, and fewer people will enter the knockout round, therefore, it is a good practice for Google to start making predictions in October 4. However, in any case, it is difficult to predict small probability events based on big data, and it is very difficult to predict the majority of non-unpopular events.

There will be a while before Brazil's match against Germany, so I just took the time to read a Google cloud platform blog and found their latest article, explaining why the German-French war has not been accurately predicted, by the way, I gave a forecast for the semi-finals. Will it be accurate? I will soon know!

I have translated some text, the content is as follows:

 

==================================== Lili's line ====================

 

Google's predictions for the 8-to-4 shopping spree were only false in France's predictions for Germany.

• Brazil vs. Colombia: Brazil (71%)
• France vs. Germany: France (69%)
• Netherlands vs. Costa Rica: Netherlands (68%)
• Argentina vs. Belgium: Argentina (81%)

 

Why did we get Germany-France wrong?

For the Germany-France WarWrong prediction?

World Cup teams are especially difficult to model because they play so few games together. USA coach Jurgen Klinsmann recently told the New York Times that he sees his players about as often as he sees his barber. if data is the lifeblood of a good model, we suffered for want of more information.

Modeling is difficult to predict the performance of the World Cup team because they had little chance to play together before. American coach Klinsmann told the New York Times that he saw his players more than he saw the hairdresser. For good modeling, data is crucial, and we always suffer from insufficient information.

 
But, we know that in the same environment, others fared better in their predictions (H/T Cortana; their model relies more on what betting markets are saying, whereas ours is an inductive model derived from game-play data ).

But we also know that, in the same case, there are still many other predictions better than we do (for example, the H/T Cortana web site is http://mashable.com/2014/07/02/cortana-world-cup ), their modeling is more dependent on the odds given by the gambling market, and we are more based on competition data.

So, why did we get Germany-France wrong? In the first four games of the World Cup, France took more shots than Germany, had more shots on target, and their shots were from a more "dangerous location" (that is, closer to the goal ). this information complements actual goals to form an 'expected goals 'statistic in our model.

So why is the prediction of the German and French competitions wrong? That's because in the previous four matches (three group matches + 16 to 8), the French team scored more shots than the German team, the number of shots at the main door (the so-called main door refers to shooting within the door frame range) is also more, in addition, they shot from a more threatening area (the so-called threat area refers to the closer to the goal ). We combined this information with the actual goal data and obtained the "expected goal" statistical value in our model.

Moreover, in the first four games, Germany allowed their opponents to take more dangerous shots, and thus the expected goal statistic was higher for their opponents. and, it allowed their opponents to pass better in their third of the field. in the Germany-France game, France actually outshot Germany with 13 shots. 8 for Germany, and 9. 6 on-target. with a little more luck on their side, they may have pulled ahead.

In addition, in the first four games, the German team gave the opponents more threatening shots, resulting in the German team's "expected goal" value below its opponent. In addition, the German team gave its opponents a better chance to pass the ball (that is to say, they were not strict enough) in their 1/3 venues ). In the German match against France, the French team scored 13 shots, while the German team scored 8 shots. The number of hits in the door frame was 9 times for the French team and 6 times for the German team. If the German team is not lucky, the French team may have been ahead.

What about the semi-finals?

What about the semi-finals prediction?
Here's our predictions for the next round:
The predictions are as follows:

• Brazil vs. Germany: Germany (59%)
• Brazil vs Germany: Germany wins 59%
• Netherlands vs. Argentina: Argentina (61%)
• Netherlands vs Argentina: Argentina wins 61%