On Data and Dogma - Conversation with Himanish Ganjoo

Sometimes cricket acts its age

In October 1973, Dynamo Kyiv - USSR’s hottest football team - signed 34-year-old Valeriy Lobanovskiy as head coach.

It had been five years since Apollo 8 went for a round trip to the moon, four since Neil Armstrong took giant steps on the moon’s surface. The Soviet Space Programme was yet to deliver its first hit. Their clandestine lunar projects had invariably ended up in debris. And so, they spent. Millions for research and development in-house, and millions more for KGB agents in Western Europe to smuggle embargoed technology back to Moscow.

Lobanovskiy’s Dynamo Kyiv won the league in his first season in charge, but he wasn’t quite satisfied. To be spoken of in the same paragraph as Real Madrid and Bayern Munich, they needed to become mechanically successful, especially in continental competitions. So Lobanovskiy leaned on the one thing he knew he could access through the USSR government and KGB: technology. He used high-end software to analyse football games going back decades, and found that ‘a team that makes errors in no more than 15 to 18 percent of its actions is unbeatable’.

He then armed his players with statistical reports of their past performances, and set them targets for sprints, shots, tackles, and everything else. And even that wasn't enough for him. If technology can be used to help with outcomes, why can’t it be used to create better footballers?

Lobanovskiy’s next hire was Anatoly Zelentsov, a scientist from the department of physical education theory of Kyiv State Institute of Physical Education. They created a mathematical model of the training process and physical load for the team. Zelentsov was later put in charge of Dynamo Kyiv’s scientific laboratory, which was popularly called the Zelentsov Center. He devised a program that broke the pitch into zones, analysing details as fine as how many touches should players take in each one, and how fast they needed to be running while they were there.

In 1975, Dynamo Kyiv won the European Cup Winners’ Cup and then-highly regarded European Super Cup, becoming the first Soviet club to win a major European trophy. Lobanovskiy would go on to collect eight league titles with Kyiv and transform the middling USSR national team into European Championship finalists in 1988.

And yet, at every turn, an accusation followed him like a shadow: he had “sucked the soul out of football.” He had mechanised magic. Every time a list of football’s most influential coaches is drawn, Lobanovskiy is forgotten, added only after someone remembers that he fundamentally changed how information was used to gain edges.

Imagine you’re attending a magic show. Now what if, just before the magician is about to enter the stage, someone from the production crew leaned over and explained how he has treated certain cards so that unsuspecting guests pick them by default; that the pigeon in the box is always there, just concealed behind frosted glass? You’ll feel a frothing rage, won’t you?

We prefer our sports wrapped in that kind of magic and mystery. Sachin’s drive and Roger’s backhand are objects of great beauty, creations of the moment when art takes over from athletics, sacred stones beyond microscopes and mathematics. Imagine being told that Federer used his backhand as a solution to problems that his opponent posed. Solution, what a filthy word. And if someone tells me that Federer used to calculate the success-rate of his backhand and used it sparingly against certain opponents, I will need a restroom.

**

Data is everywhere. It is the central nervous system, guiding our decision-making for everything between the restaurant you will order from, to the route you should take to work, to promoting Axar Patel above Hardik Pandya in a World Cup final. And yet, barring professions that survive on mathematics, most of us continue to have a cold relationship with numbers.

Organised cricket is nearly 150 years old, and its awareness about data and statistics seems stuck in the grainy monochrome of 1877. Even with terrabytes of new information available every second, we are rarely ever pushed to understand the game better, beyond the fact that Virat Kohli averages 47 runs every time he bats.

I am on a Zoom call with Himanish Ganjoo, and that is my first question to him. How is it that this sport - watched, followed, and played by enough people for it to be called global - is so archaic about statistics? Why are the numbers on my 4K monitor the same as the ones I remember seeing on my CRT television?

HG: In cricket, statistics are very format-dependent. For example, in Test and ODI cricket, the traditional stats that are shown on screen - average, strike-rate etc - are sufficient to understand the value of a batter. Bowling average and economy is good enough to understand how good a bowler is. T20, however, needs derived, calculated metrics, but we aren’t quite there yet.

Himanish is a post-doctoral researcher in cosmology at the Laboratoire Univers et Théories at the Paris Observatory in Meudon.

In the spring of 2021, England’s Test team arrived in India with their usual cargo of hope and technical inadequacies against spin. Ravindra Jadeja was injured, which opened a door for Axar Patel at Chennai in the second Test. A left-arm spinner with reliable lower-order batting, Axar seemed like Jadeja-lite, a natural replacement.

What followed was barely believable. Twenty-seven wickets in three Tests. Twenty-seven. The English batters were like a pack of deers stuck in a Gurgaon highway on a weekday evening. There were 10 kilowatt headlights whichever way they looked.

Amidst the hysteria, Himanish wanted to understand the mechanics behind the magic. So he turned to ball-tracking data, to find what Axar was doing differently.

Analysing ‘lift’. Source: Himanish’s article on Axar Patel.

Months later, this analysis found its way to Rahul Dravid – then-coach of the Indian team. Conversations began, questions were asked and answered. Eventually, Himanish was employed by the Indian men’s team from 2022 until the conclusion of the 2024 T20 World Cup.

SD: Do you think there’s a sophisticated understanding of data and statistics within cricket? As audiences, we might be exposed to easily-digestible metrics, but even the commentators - most of whom are ex-cricketers - seem to be talking in glib cliches and surface-level numbers.

HG: You’ll notice that most of our conversations about numbers are still stuck in the mid-20th century. T20, of course, is extremely young and very tactical, so there is some movement in terms of analytics, but otherwise, the sport hasn’t moved on from the conventions of how Bradman and Sobers were judged. We are judging a Liam Livingstone T20 innings the same way as we judged a Jacques Kallis Test innings.

SD: Is there an inertia, then, within the sport to accept data?

HG: It’s a bit complicated because of the three different formats. I mean, yeah, the sport is unserious about data. But, if you consider Test and ODI cricket, they are slow formats and not very tactical. So, teams eventually have to rely a lot on technical skill. Thing is, professional cricketers are a) very, very skilled and b) hyper-aware of their game. They can figure out most of the stuff by themselves. So you don’t necessarily need deep statistics to unlock new information. Secondly, every team has video analysts anyway. Whatever information players need about themselves or the opposition, they watch videos and figure it out themselves. It is understandably tougher to make a case for cutting-edge data analysis in these two formats.

The difference comes in T20 cricket, where one good matchup can win you a game. The format’s a lot quicker, a lot more volatile. So, yes, it is easier to see the impact of data. But it is also very nascent. The problem is - to reach the level of using data to make decisions, one needs to first reach a point where the dynamics of the format are understood well. You see actual teams playing batters with 120 strike-rates in the powerplay. Data can’t help you see what you aren’t ready to see.

SD: What does it take for a team to become data-savvy?

HG: You need people to think five steps ahead. More importantly, they should be willing to ask questions and be aware that, for all their experience, they might not have all the answers. Rahul bhai and Rohit [Sharma] were extremely progressive with their thinking. We were committed to find everything we could, and then figure out what to use. For example, in the T20 World Cup in Australia, we knew that length would be key, so we put markers on the right lengths to bowl.

See, you aren’t going to tell a Rohit or a Kohli how to bat, of course. They’ve been playing the game for 15 years. They know what works for them and what doesn’t. But data can tell Bumrah that Phil Salt has a weakness if you move the ball into him. [smiles]

At a team-level, it takes a very forward-looking coach and captain to be willing to use a non-cricket object - a computer or a data-analyst - and trust it to help them. Trust is important, and that’s a barrier most analysts are finding tough to breach.

SD: Fair. As an analyst, how do you ensure you earn the trust of people with a million runs in their career?

HG: You have to contextualise information. You have to work on your presentation. For that, you have to know the game very deeply. That’s a non-negotiable. Throwing raw numbers at a team management serves no purpose. One has to create information even for the coach and captain. Whether they use it or not is a different thing, but when you’re serving data, you need to add to their options on the field, not give them a quadratic equation.

SD: Right, so am I correct in assuming that an analyst has to find non-obvious data to begin a conversation that the coaches/captain haven’t yet had?

HG: Yes, totally.

SD: How does one do that? Is granular cricket data as easily available as, say, in football or baseball?

HG: Hahaha, absolutely not. Cricket data is heavily-guarded by the handful of companies that produce it. Home boards bring their broadcasters, who own the optical and ball-by-ball data. They don’t publish it; they have no incentive to. The companies who log tracking data guard it because they will make some money by selling it to data analysis companies. So data beyond the obvious stuff is very, very hard to get.

And now, even some of these companies are moving away from assisting teams because there isn’t proportional revenue for the effort they have to put in. Which is worsened by the fact that the sport itself isn’t interested in creating an ecosystem of data analysis.

SD: So where does a rookie analyst go?

HG: They can start with Cricinfo and Cricmetric. They won’t give you granular data, but they’ll get you started. After that, look up Reddit and other places for people maintaining databases. That’s how I started. These days, I publish a lot of data myself, so hopefully that helps.

SD: That seems like a narrow world. Do you see the sport changing its ways in the future? Do you see cricket becoming as tactical and data-aware as some of the other sports?

HG: Yes and no. Right now, cricket is being managed as a TV show. And like the footage of a show, cricket doesn’t want anything to get out of its hold. The moment they make data accessible, it would mean that a lot of people from outside the sport will start to influence it. Anyone with an eye for statistics will have the tools to create something of value and potentially break the existing hegemony. Which they most definitely do not want. Besides, a lot of people who are paid to talk about cricket like this status quo because they get to cruise on their existing, archaic knowledge without being challenged by statistics. But there is also the contradiction to that coming from T20. Data is the present and future of creating an edge for yourself in that format.

Probably, a new generation will have to come in and change the narrative. The hope is that they are more comfortable and accepting of data than what we have today. Some commentators, like Abhinav Mukund and Shane Watson, are excellent with data.

Hopefully, with time, there is a greater understanding of how data can help with tactics. For that, you need coaches, captains, commentators to be on board with the idea that the skills you are taught in your childhood coaching sessions may not be the only route to success. Someone like Andy Flower uses data very well. We are a long way away from that kind of awareness, where such coaches are the norm instead of the exception.

***

Surviving a promotion to the Premier League is one of football’s great challenges. Promoted teams often have to spend the entire year just catching up with the physical and technical demands of the Premier League. And then there is the money that the top teams are armed with - oligarchs, oil barons, sovereign wealth funds, what have you. Most teams perish, understandably so.

It was always going to be tough for Brentford. Actually, not just tough, near impossible for a club whose previous taste of top-flight football came when Neville Chamberlain was still telling everyone peace was coming.

Yet here they stand. Since their 2021 promotion, they’ve beaten every Premier League opponent at least once, never finished a weekend below 16th, and have outscored Manchester United in the last four years. All this while operating on the second-lowest wage bill in Europe’s richest league.

A lot of this improbable rise can be pinned squarely on Matthew Benham, a boyhood Brenford fan who grew into a mathematical maverick. At the sports gambling company Premier Bet, his job was to help develop predictive models for gambling. Benham made millions. And in 2012, he bought Brentford FC.

But, before he could rinse his own club inside-out, he needed a laboratory to test the material. Benham spent £10 million on Denmark’s FC Midtjylland. Out went football men with football thoughts, in came analysts with algorithms. In February 2016, Midtjylland beat Manchester United in a European game, with a squad assembled entirely through data-driven scouting.

Today, Brentford lean heavily on science and information. In the last couple of years, they have spent £16 million on research alone, publishing academic papers on everything from sleep therapy to training loads to positional play - all of it used heavily by head-coach Thomas Frank.

If Lobanovskiy and Benham’s stories sound familiar - like that of a book you expected me to quote in an article about data in sports - that’s because lightning, when carefully engineered, can strike as many times as you want.

Sports has come a long way from scouts sitting in trench coats and berets noting every action by hand. These days, the Head of Research at a club of Liverpool’s stature can be an astrophysicist from Cambridge.

Cricket, stubbornly, arrogantly, lags behind, yet to have its Lobanovskiy moment, nevermind a Matthew Benham moment. Talent isn’t an issue. It has a Himanish, a Nathan Leamon, a Freddie Wilde, but it needs more people like Andy Flower and Rahul Dravid to truly open the doors for a culture-shift. And it needs to come out of the dogma that critical analysis is an intrusion in an atmosphere that can operate on muscle-memory alone.