How ISL teams are turning to tech
From player movement to predicting the future, data crunching is an evolving game in Indian football
After every match day, Indian Super League (ISL) debutant Jamshedpur FC’s analyst Gaurav Khilari spends 6 hours watching football videos, studying the players using Sportscode (a video analysis software, also used by US National Basketball Association team Denver Nuggets) and editing it according to the needs of his coach, Steve Coppell.
To make sure the clips he cuts provide insight, he has to watch their entire game twice, at times thrice.
He then gives Coppell videos of an individual player’s, or the entire team’s, movements, as required. Then, he and the coaching staff conduct meetings with players, each of which may go on for an hour.
Video sessions before a game—meetings where the analyst and coach sit with players and attempt to explain mistakes and positives from the previous game—usually last 90 minutes. But given the ISL’s uneven fixtures—Kerala Blasters played their last three matches within eight days—some team sittings wrap up in 5 minutes.
“Our coach is defence-oriented,” says Khilari. “The first thing we see before every game is how the opposition team attacks.”
Having conceded only nine goals in 11 games, Jamshedpur have the best defensive record this season so far.
“I also create motivation videos of all the good things the players did during pre-season and in the previous matches, I show it to them with music to give some flavour. I add some funny clips as well,” says Khilari.
Like Jamshedpur FC, other ISL clubs too have begun to analyse player data. Currently, most sides have just one analyst. According to a 2014 article in The Guardian, Manchester City had 12.
How much does analysis matter? And how deep do teams dig? Mumbai City FC’s assistant coach Juliano Fontana says a third of his team’s strategy comes from analytics. FC Pune City head coach Ranko Popovic prefers shorter analysis videos so players are not overloaded with information. These simple examples confirm that data and technological tools do matter to Indian football clubs.
Crunching numbers and arriving at conclusions is, however, just one aspect. How long before they begin to predict the future?
The level of technological dependence, varies across clubs. While footballers wearing vests with GPS devices are common even in Indian football these days, Jamshedpur’s players use these solely for practice sessions, not during matches. The tracking devices they had procured from Australia were detained by customs officials in Thailand, the location of their pre-season camp.
“We received them just recently in Bengaluru. We didn’t have it for the first six games, so the coach decided not to use it since the players are not used to it and it’s also uncomfortable,” says Khilari.
Lokesh Bherwani, the performance analyst at FC Goa with a master’s degree from the Madras School of Economics, has a mathematical model to find out which player’s pass carries more weightage. Who has greater probability of creating a goal-scoring opportunity? “I came up with this model,” he claims. At Goa, players wear the Catapult Optimeye S5, tracking technology used by Dutch outfit AFC Ajax and the Brazilian football team.
After every match and practice session, Bherwani feeds data from this GPS device on to his computer to check the distance players covered and the number of sprints. “It helps us to know if we’re putting too much load on a player,” he says. Like Khilari, Bherwani spends 4 hours after every game analysing videos with the help of Video Observer, a software that generates player and game reports.
GPS systems like Catapult produce scores of data sets (total distance covered, maximum sprint speed, number of times a player changes direction, heart rate, number of dives made by a goalkeeper, etc.) but Fontana says he can, at most, use 25 of them. “It’s not easy to use, you need to study it,” he says.
Apart from producing statistics about an athlete’s physical activity during a game, trackers help coaches understand whether players have successfully carried out game-specific tasks.
Mumbai’s central defender Gerson Vieira has often operated as a defensive midfielder this season. On an average, the Brazilian covers 11km during a game.
But covering this kind of distance could prove counterproductive for the team, as positioning is key. Mumbai’s captain and centre back Lucian Goian once ran in excess of 10km. The 34-year-old Romanian ventured out of his designated area far too often, thus leaving open spaces at the back. “He needs to cover only 8-9km,” explains Fontana. The team looks far more solid when he sticks to covering the stipulated distance.
With 68 tackles in 11 outings, Goian is currently the league’s most effective defender.
Dissecting an opponent may not guarantee success but it throws up interesting narratives. An analyst, who did not wish to be named, says that after studying a rival adept in dead-ball situations, his team replicated one of its successful set-piece routines and scored a similar goal against the same rival.
Popovic says: “I don’t give too much information to my players. If you make the players’ minds too heavy, they forget what they actually have to do. I don’t like to make longer videos or watch the game with my players again because it is boring and this can put them to sleep.”
Instead, he asks his analyst to make 12- to 15-minute videos before every game, encapsulating defensive patterns and how their opponents build their attack.
At the other end of the spectrum are those obsessed with numbers. Studying past events and correcting them is just one aspect of football analytics. To truly harness its power, one has to accurately predict possible future outcomes and trends.
Bengaluru-based Amrit Murali, a product manager with management consulting firm Zinnov, says ISL teams haven’t explored this aspect. The 24-year-old is a fan of the hardly understood concept of expected goals (xG)—which essentially involves trying to evaluate the quality of scoring chances. He got into it to improve his team in the Fantasy Premier League, an online game.
After running a statistical model 10,000 times, Murali makes a bold claim. “There’s a 65% chance that Chennaiyin will finish on the top of the table and a 20% chance they will finish third or fourth. Bengaluru, Pune, Goa and Chennai will finish in the top four. But the position of every team within the top four is hard to predict.”
According to data provided by Fontana, Mumbai City FC’s attacking midfielder Thiago Santos doesn’t shoot much but when he does, there is a good probability he will score. In nine matches, he has taken only five shots —and scored four goals. His colleague, striker Balwant Singh (five goals), takes far more shots but doesn’t score as often.
Most fans are yet to warm up to the idea. When C.K. Vineeth scores a 90th-minute winner, courtesy a well-taken header from a cross that he received from the right wing, do you credit the manager or the analyst—besides, of course, the player?
If you’re a Kerala Blasters fan, you’re simply living in the moment and experiencing ecstasy. For now, the analysts are essential yet invisible cogs. As Khilari puts it, “The fans think I am some camera guy.”
Editor's Picks »
- Policy rethink and higher volumes to aid container shippers
- DCB Bank delivers a strong Q2 but pressure on margins foreseen
- Havells India: Rising costs give a jolt to profitability in September quarter
- All’s well at Mindtree, except for high client concentration risk
- India’s rising steel demand is making companies starry-eyed