Computers and humans: in search of sports prediction

When it comes to predicting the outcome of a football game, fans turn to all kinds of skills: they analyze each player's skills, look at the history of the matches between both teams or, in the absence of all that, they simply feel that match can end 1-0 or 2-1. What these fans don't have is the ability to analyze statistics in the way a computer does. To know more click on 먹튀폴리스의 안전놀이터인 토토사이트.

To answer that question, it is convenient to take a tournament that is currently being held: the African Cup of Nations, which is the most important football tournament on that continent and that began in Equatorial Guinea and Gabon.

The winner will be Ivory Coast

Or at least that is what the black box of the sports statistician Robert Mastrodomenico predicts. It is a computer model made of complex algorithms.

He enters historical data on the performance of the teams participating in international games and the model calculates the relative strengths in defense and attack of each country.

Based on that information, Mastrodomenico can classify teams.

"The model shows that the strongest team in the competition is Ivory Coast, followed by Ghana. Then it shows the following teams - Tunisia, Morocco, Senegal - with similar ratings," he says.

In the names of these five teams, the computer predictions match those of the Ladbrokes and William Hill bookmakers, who have these teams (in addition to Mali) as favorites.

But the sixth team that mentions the computer is a surprise. It's about Burkina Faso.

Mastrodomenico says his data shows that Burkina Faso has similar results to Tunisia, Morocco, and Senegal, but is usually ranked below these teams, around the eighth place.

So what do football experts think of applying this type of analysis to their sport?

"Drunk computer"

"I think that computer is drunk," says Farayi Mungazi, who is covering the tournament in Gabon for the BBC.

"I mean, Burkina Faso is more likely to send a man to the Moon to win this tournament."

"They are very inconsistent. They have good players here and there, but if a person bets their money on what Burkina Faso will win, I think that person has more money than common sense."

But Mastrodomenico argues that by letting the word have the data and equations on his computer, his prediction has an unbiased or dispassionate profile that the market doesn't have.

Customers pay their company, Global Sports Statistics, for their models to deliver information that can then be used for a variety of purposes, including betting.

Farayi Mungazi points out that a complex computer is not needed to indicate that Ivory Coast and Ghana are strong teams.

"I am sorry to be Rob's math spoiler, but I could have chosen a couple of young people from the streets here in Gabon and they would have chosen Ivory Coast and Ghana."

And he argues that statistics can't tell everything about a team: that's why he doesn't think Ivory Coast is the most likely winner.

Senegalese surprise

"I bet my money that Senegal will win simply because it has the best team, put by position, of the tournament."

"Their strikers are loaded with goals, they are hungry and they showed in recent months that Senegalese football is back."

"Remember that everyone was surprised at the 2002 World Cup, but that has turned out to be a curse rather than a blessing, because, after the 2002 deed, Senegalese football went downhill."

"But in the last two years he has risen again and I think this is the moment when he can shine."

He adds: "I don't think Ivory Coast is good enough to win this."

"He has very good players, but he doesn't have players who play like a good team, and that's his Achilles heel."

"She was the big favorite last time and did nothing."

"Ghana has a good history in the Nations Cup, but it has so much pressure to win a fifth title that it seems to affect players a lot."

Mastrodomenico accepts that to a certain extent, relying solely on historical results can be problematic in a tournament that has a direct elimination structure, in which most teams play in an environment that is unfamiliar to them.

"As for predicting events, models like this work well, but are more appropriate for predicting matches than competitions."

"For a match, we can use the strength of the teams to calculate the odds of victory, draw and defeat."

"And on the odds, no team in the competition will be one hundred percent sure of winning, so everyone has a chance, at least in math!"