INTERVIEW
Follow UsThe 2022 T20 World Cup was one of the biggest seismic events in Indian cricket in the last 15 years. Not only were India out of yet another semifinal, but the manner of their defeat indicated that India needed to drastically adopt a more meticulous approach while planning for big matches.
The fact that India hadn’t won a World Cup or Champions Trophy in the last decade compounded the imminence of those changes.
Rahul Dravid and his support staff wanted to cover every base tactically. The team already had Hari Prasad Mohan, a data analyst with Sports Mechanics, working with them. However, to leave no stone unturned, before the 2022 T20 World Cup itself, Dravid had gotten in touch with someone you’d not generally associate with a cricket team.
Cosmologist Himanish Ganjoo is a well-known figure on cricket Twitter (now X) for his cutting-edge data science work on the game. As it turned out, data analysis was at the forefront of the team's transformation after the tournament defeat.
After going through his work, Dravid enlisted him to help the team with two World Cups in the next two years. Now that Ganjoo’s stint is over with a victorious culmination, he had an exclusive chat with Cricket.com. He discussed his work for the team, how it helped them, and the culture of data analysis the team inculcated through it, which eventually contributed to the Indian team finally getting their hands on a T20 World Cup trophy after 17 years.
Here is the excerpt of our conversation with him -
I remember reading a really in-depth article about Axar Patel from you. Was that the work that made the Indian team management reach out to you?
Yeah, so the Axar Patel article I wrote, I published it on my blog, and I think I sent it to Jarrod Kimber. Jarrod made a video on that thing, and I think that video reached Rahul Bhai. And then he reached out to people at ESPNCricinfo [who Himanish had written for extensively], I think, and he reached out to me after that.
Then, we had a couple of Zoom calls. He [Dravid] was very nice and extremely polite. He was very open to sort of exploring what I did, and he wanted to know where I was in the world and whether I was working with a team. Then we chatted a bit more and the process of bringing me on began in the middle of 2022. In September 2022, we sort of formally started.
So it was just before the T20 World Cup of that year, right? Were you part of the planning for the 2022 World Cup as well?
Yeah, I began before the T20 World Cup 2022, and we started having meetings before it. The tricky part at that time was when I was just starting out, I had to write a lot of code. I had to write a lot of pipelines to analyse the data I just got. Then, I began writing and analyisng on that. Even before the World Cup (a month or so), we had already started having meetings and analysing things about our bowlers and opposition batters.
The most important thing on Australian grounds is where you bowl. What lengths do you bowl? Whether you bowl slower balls more, bowl the hard length more, or bowl a yorker. But also, where a good yorker is pitched, for example, we didn't know that. So, what is the definition of a good yorker? We had to find that out. And we found that out using tracking data. Who has the best yorker in the world? What are we doing versus them? So, a lot of detailed, meticulous analysis was used to research everything.
Also, if I remember correctly, the World Cup started a bit early for Australia in October. The pitches were behaving differently, and pacers were getting a lot of movement. So how did you and the team management tackle this issue that came up?
Oh yeah, this was a factor, and I'd like to commend Rahul Bhai and the coaches for this because they understood it and were prepared for it. We used data from Australia, including Big Bash League data and other sources.
But you had to recognise that this slightly differs from what you'd expect in early October. So we thought about that as well. And every question was researched. And together with Hari [the team analyst] and the coaches, we made sure to research every question we had. And it was everything from as simple as how should you sequence a slower ball after a faster ball. If you do that, where should you bowl?
Can you explain the process of how you approached your preparation for every match? What was the process the team was following from a data analytics point of view?
Before every match, Hari analysed every batter. So, for every batter, we used to get a wagon wheel. We used to get their shot-type data. What are their strong shots? What are the filters you need to pick? Now, when I joined after that, we sort of supplemented that with ball-tracking data. So, one of the first things I did was to match ball tracking data with shot-type data, which earlier hadn't been matched. And then, before every match, I sent about four to five slides for every batter.
Let's say you take Babar Azam, for example. Now, if you bowl pace to him, like, what lengths are you bowling? What lines should you not bowl? If you bowl slower balls, should you bowl them? If you bowl off-spin, should you bowl that? Where should you bowl that?
Apart from that, in T20s, you need to organise your bowlers. You need to know what a good matchup is, and what a bad matchup is. If you're stuck with a bad matchup, where do you go? So if you're stuck with bowling a slow left-arm bowler to a left-hand bat, there are still areas that you can bowl in. However, how do we determine what those areas are? What are those lengths and lines that you can bowl? What are the angles you can exploit? Where are these people hitting their shots?
Hari and I used to come up with this together, and it used to be some four or five slides for every batter. Before every World Cup match, for example, I worked on every single team. And for every prominent batter, we used to have a clear set of plans. Also, sometimes, you don't have a plan because the data doesn't say anything, but if it said anything, I ensured we had it there.
So yeah, that was a workflow we had before an important match.
That sounds like you had to do a lot of information crunching before every game. But how did you take this to the assistant coaches and Dravid? Even the data that was crushed and visualised into presentation slides would have to be broken down into the simplest understood denominator, right?
I would send, let's say, five tables or five plots, but that might be too much information. So the coaches are very good at this. Coaches are very open to data. So they were very open to taking in all the information I was giving them, and we fine-tuned a process where, you know, they understood, and they made me understand how they wanted the data.
By the end of it, it was just that they trusted me enough to analyse data. Earlier, I used to send them like tables or plots. But later, they trusted me enough to analyse those tables and plots and send them as pointers. So instead of tables, I'd send them many point-based insights about a certain batter. By the end, it worked because we understood our workflow and what they needed regarding data.
The other thing is, you can't confuse players with data, right? So there was a pipeline of communication between the coaches and me and the players and Hari, where he and the coaches would take the data I gave them and distill it into, like, let's say, three or four points that you can tell the players because players don't care about what they are averaging on a good length, they don't care about all that. You can't muddle them with all that. So what you do is you design a method to process that data and simplify it enough to talk to the players, and they are the ones who did that, and I send them all that I could.
So, was it a case of the players also coming to you, or was it only the coaches?
I didn’t talk to the players directly, because the art of communicating with the player is difficult, and I was working remotely, and I didn't know them, right? So that was a little difficult. Hence we had a pipeline where I talked to the coaches, and the coaches decided what to say to the players and how to process the information.
Would it be right to say that no Indian team had used data in such a rigorous and meticulous manner?
I think so, and the credit for that goes to Rahul Bhai because he is the one who is forward-thinking and radical enough to begin this culture. His idea is always to design a culture within the team, and he was the one who started this. I can say, surely, that we were right up there with the best teams in the world in what we were doing in terms of the amount of data we were exploiting, how we were analysing it, and how deep we were going.
So, I went very deep into these things. Ultimately, only 1% of this makes it to the ground. But the important thing is that you do 100% of the research, and we were doing that.
I'm very proud to say that we (with Hari) and the coaches placed this culture, and used all the data we could. We had ball tracking data, we had shot type data, and we were merging everything together. We were getting these insights. So, yeah, I can't say for sure, but I don't think any Indian team before us has worked with data to this extent.
Before this, you were mainly working on cricket data on your own and putting out the work on social media. It was still very deep and diligent work that teams could derive many insights from. So how different was it doing the same kind of stuff but for the Indian men’s team?
There's a big difference between doing things for media or doing things academically versus helping a team. Many of the things I've done for various websites, like ESPNCricinfo and my Twitter or blog (sample below), were academic. They were nice tools to understand the game, but how to remodel is not often important to teams.
Now, some aspects of them are really useful. I used to have a model for impact scores, for example. So those things are good guidelines to tell teams that you know you are scoring but you're not scoring fast enough. So you have to remodel how you package information for the coaches and for the team, because they don't care about academics. They don't care about all that.
You want to do directly applicable things that you can tell a player, or you can tell to a whole team. Those are useful, and only those things you tell them. You don't have to show off to them. They don't care about all that. They are focused on many things that are useful to them. So, for example, things like the RAA (runs above average), I don't think we ever showed it to any player because they don't care about this. But the coaches were fairly on board with the idea of it. We showed them that how even if you've scored some 100 runs, your RAA can still be negative.
Apart from that, you just try to give them actionable insight that you know, like, if this bowler is not bowling in the right line, you tell them so they can change it. So, there's a vast difference between what you do for media and what you do for coaches.
Can you recall any particular instance where some insight from you worked on the ground that we can recall from this World Cup?
I didn't get any feedback on whether this or that worked. It was a team effort. I used to send things sometimes that I would see happening. There are multiple instances, but it's hard to pinpoint one.
I think it was evident in the Phil Salt dismissal in the semi-final. Now, I don't know if this was because of me because I haven't got any feedback on it, but in the slides that I did send, we saw that Salt has a weakness on the leg stumps. You needed to angle the ball into the middle and leg to him. And that was what [Jasprit] Bumrah did. He brought all his balls angling into the leg stump. And Salt couldn't do much with it, and then he got out to a very good ball. But this is one of the examples.
However, for me, the larger thing was what impact this work is having on changing the culture of thinking about the game with data. These coaches are very smart. They have played a lot of cricket. They know cricket, but just remodelling your perspective using numbers adds a lot to it. For example, what is a good length? What are its measurements and location on a particular pitch?
Just knowing this and having this knowledge in the team that this is what an actual good length is and where you need to focus is crucial. These things change the culture. We all started thinking about the game in terms of numbers and quantifying it better, so these things are very important.
Yeah, the cultural aspect is huge. However, since your work was still based on a contract and not a full-time thing, do you think that the next management will also continue this culture you set up along with Dravid and his coaching staff under Gautam Gambhir? It’s almost an entirely new backroom staff of coaches, so how does that affect the assimilation of such practices in a team’s psyche?
I mean, this is a very important question. I think Hari is still with the team. So Hari is part of that setup and culture that we had. I'm sure he's doing great work with them. And I'm sure Gautam Gambhir is also doing great work, and he has his own way of beginning.
However, I haven't continued this job, so I don't know what will happen to it [his work]. It's still there. So, whatever you know, code we wrote, visualisations we made, player tracking we were doing, it's all there. But the team doesn't have it anymore, so I don't know their plan. They must have some plan, but they haven't contacted me, so I’m unaware of that.
There’s also a case to be made of things working in one setup but not in another. What are your thoughts on that?
Yeah, definitely. The fact that I was working for the team, which sort of worked for us, is largely down to the coaches and Rahul Bhai and how they were open enough to take my input. And they were big enough and great enough to understand that this guy can bring some value, which I am very grateful for.
That's very, very lucky because, within cricket, the sort of acceptance for this is not much. Still, I was very fortunate that all the coaches, and Hari and Rahul Bhai, were very accepting and very open, and they sought my work. So that part is very important. You know your knowledge ultimately has to go through the coaches to the players, and if the coaches understand it and appreciate it, that's very useful. However, it's not like that one thing that worked in one setup would also work in another.
Now that you are done with your stint and India has won the World Cup, I would just like to know how you view your work after two years of working with the team. What does it mean to you? Also any future plans already in work in cricket for you?
We were committed as a management team and a data team. I was committed to knowing every bit of information about our own players and opposition players and to being at the forefront of data analysis.
I think we can safely say that we've done that along with the team. And the kudos again go to the team again, the team that executed. But at least we should know everything in the back room. We should have all the information we need. So I think we did that very well.
As for the future, there are no plans at all. I had a full vision for the Indian cricket setup, but it seems like there's no interest. So, I’m just getting back to writing and general cricket research. I built many pipelines, methods, and metrics that I want to keep up with and improve upon. There's no incentive to do much right now, though, so I'm just taking a break and switching off from actively following cricket.
Before I let you go, I have to ask you about the state of data analysis in cricket today. Where do you think we lack? Show a way forward here.
Analysis in cricket mostly refers to video analysis. But video needs to be used in conjunction with data to glean insights that large numbers can give you and see things that aren't obvious to the eye. The process should be to understand and model the underlying theory of the game using a combination of cricketing insight and models.
Cricket needs to go a long way to build a culture of getting intelligence from data.
If you’ve not downloaded the Cricket.com app yet, you’re missing out — big time. Play Fantasy on Cricket.com NOW! Download the App here.
