Living rooms these days are buzzing with football fans cheering for their favourite teams. While this is the scene at the front, there is an invisible Brobdingnagian – Artificial Intelligence, around this world cup.
While the surprise of what will turn out every match is thrilling for the fans, there is a lot of investment and pride that is at stake for each of the teams and the players. Every coach wants his team to win. A view of the football game from a coach’s standpoint is interesting. Football analytics started with Charles Reep who in the 1950s, used pencil and paper to note down his observations. He inferred with his observations that most goals were scored with fewer than 3 passes and hence to improve the probability of a goal, he recommended 3 or fewer passes. It is called “The Long Ball“. Since then, football analytics has come a long way.
These days, algorithms are predicting which of the teams will move forward based on the performance data of various teams. What do these new techniques predict as the likely outcome of the 2018 World Cup? An answer comes from the work of Andreas Groll at the Technical University of Dortmund in Germany and a few colleagues. These guys use a combination of machine learning and conventional statistics, a method called a random-forest approach, to identify a different most likely winner. One of their model outcomes looks like this,
There is another prediction by Goldman Sachs which says Brazil will be the winner after simulating “1 million possible variations of the tournament in order to calculate the probability of advancement for each squad.”
It has also been proven that expert predictions of winners in the game of football have been around 31%. Machines are able to significantly improve the winning probabilities.
A reflection on the Champion of 2014 FIFA world cup winner “Germany” tells us another story.
All of the final 32 teams competing for the World Cup in Brazil had a dedicated performance and video analyst, but Germany appears to be the only one that had a specially built database to measure and analyze individual and team performance and strategies. Not only did the German team collect and analyze a vast amount of data on its own and opposing players, but it delivered the data in a visual and easily understandable manner to its players, trainers and coaches, via a custom-built app, so they could use it on their mobile phone or tablet, as one German coach said, “whenever and wherever they want.”
Germany’s big win over Brazil was attributed also to the data and insights Germany had, which The Wall Street Journal called “Germany’s 12th Man at the World Cup“.
Bringing the context back to Fashion, this is an important period wherein many brands are placing their bets for Spring Summer 2019. Whether you are doing bookings from your customers or placing bets for your own stores, there is a significant bet that goes in every style, colour option, and any error in these bets is bound to be expensive.
With today’s technological advances, every brand or retailer can now bring the power of prediction algorithms in the form of machine learning and artificial intelligence. As Germany did, it is not only important to get insights but also to have it available in a simple visual form with all key decision makers/stakeholders as and when they need it. This is impractical if you are working with centralised intelligence units in traditional computation formats.
This calls for data (visual, textual and signals) gathering at internet scale, relevant to your brand’s context. Further machine learning models discern the unknowns from the knowns.
Fashion prediction has been a non-trivial problem to solve. Traditional statistical models have not shown promise. We have adopted a unique modeling using thousands of aesthetic and semantic parameters with past, present, and future signals. This is what is cooking as Stylumia APOLLO.
Using machines along with our intuition is all the more critical considering that we all have a cognitive bias. In our recent panel discussion at ReTechCon2018in Mumbai, we demonstrated the cognitive bias humans have. It is easy to try yourself. For how to, watch the video (1 min 23 s) below…
Would you like to be the champion in the game you play in Fashion and win your fans (consumers)?