It is very easy to speak in the past bull, as I am going to do it. But what I mean is how easy it is to make mistakes, not only as a human, but also as a machine that has been fed millions of data and used state-of-the-art predictive platforms and systems is capable of not getting it right no matter how much information it contains. , no matter how relevant its computing capacity is, and no matter how ‘neutral’ the algorithms it uses to process the data are.
On June 14, at an IBM event held in Madrid, Javier Valencia, vice president of IBM Cloud for Spain, Portugal, Greece and Israel, and María Borbones, specialist of the artificial intelligence (AI) platform IBM Watson, began a dialogue with Watson asking how Spain was going to be in the first phase of the Soccer World Cup that is being held these days in Russia.
In order to be able to say what was going to happen, they entered into the machine the data of all the previous soccer World Cups, of all the players who played it, of everything that was said about them on social networks, of the weather. that occurred in all the matches, from the weather forecasts of the current competition, …
And all this with the help of analysts and data scientists who have used to feed Watson both structured and unstructured data, data sources of different types, always reliable, image analysis and even the emotions dictated by these same images of the players. Javier Valencia himself pointed out that “we had to homogenize, prepare and refine all the data to create a predictive model that was capable of dictating a result”.
And what were IBM Watson’s predictions for Spain’s first phase of the World Cup? Watson said that Spain was going to draw with Portugal, that he was going to beat Iran, and that he was going to beat Morocco as well. It did not shoe much. We all know that Spain drew with Portugal at 3 goals and beat Iran by one goal. But they did not win against Morocco, and we were in a hurry to lose.
What IBM Watson, nor any predictive platform will ever be able to guess, and here we get into the opinions chapter, is that a goalkeeper like De Gea who has been the best in the Premier with Manchester United has spent all three World Cup matches making the statue, making a single save in 270 minutes, and conceding five goals when the opponents have fired seven shots between the three posts of the Spanish goal. Surely Watson from IBM saves all this information for the next World Cup, and for the games that remain for Spain in this one, for its future predictions.
And what IBM Watson was not able to predict either, and no other would be able to do so, is that between Sergio Ramos and Andrés Iniesta they gave a ball to the Moroccan striker to take the lead. And even less, since in the World Cup in Russia it was the first time that the Video Arbitration system (VAR) was used, Watson introduced among its variables that a virtual referee was going to validate Spain a goal against Morocco that avoided the defeat and annul another to Iran that supposed the tie, with which Watson would have struck in another result if Iran and Spain had finally tied.
Yes, prediction systems that use millions of data, always reliable, and that use algorithms without bias, can help to guess trends, patterns, soccer results or who will win the World Cup in Russia. But they can also be wrong. Will Watson be able to guess the outcome of the round of 16 match between Spain and Russia on Sunday, July 1? And who is going to win the World Cup? Germany will no longer win it, for sure none of the many predictive systems that have wanted to score the goal of guessing the champion was able to risk his elimination in the first phase, something that had never happened to him in the history of the World Cup. But those data, those of this World Cup, will be entered into the intelligent machines that will want to guess who wins the Soccer World Cup in 2022. And Germany will already have fewer chances according to data, algorithms and artificial intelligence.
