Couch Talk

'My job is to give the player the best chance of success in a match'

The Indian team's performance analyst, CKM Dhananjai, talks about the different levels of analytics used in quantifying a player's performance, how players react to data they are provided, and how he helps them make better decisions

SJ: You are the performance analyst for the Indian cricket team as well as the Mumbai Indians IPL franchise. What are your duties and responsibilities as a performance analyst?
DJ: As a performance analyst, your main role is to create an environment in terms of setting up the right technology and also bringing in the analytics to the table. When I say "analytics", it has to do with helping teams, players and other stakeholders make better decisions in terms of taking a set of takeaways into the game.
Overall, I would say that it is for me to create an environment to enhance performance and can help high performers make better decisions and take crucial insights going into every game, series or season.
In layman terms, I would associate it with any other business process that happens in the day-to-day world.
The performance analyst's role is not a one-dimensional role. There is a team that is involved in the whole process. So right from the preparation towards a team and a player in terms of putting together stuff about the opposition analysis - knowing your opponent, putting together the strengths and weaknesses, the opportunities that are involved and the threats. Doing a SWOT analysis on the opposition as well as our own team in terms of individuals is a key element in performance analysis.
Overall, it is about monitoring, tracking and providing insights to the players or teams that we work with and helping them take that into the game to get the desired results.
SJ: If India is facing Australia, do you put together the packages of the last five series? What level of detail are we talking about here?
DJ: The level of detail is quite high.
To start with, the fact that preparation as a process is quite new in cricket. It is as new as the last ten years. Earlier, there were a lot of qualitative assessments and opinions that obviously were around the team when going into the game. Slowly it has changed to using quantitative methods, or in layman's terms, "data-to-the-table approach". It is more a necessity than a luxury. If you are playing Australia or any other team, or even an IPL team, the process that is involved is of going through each and every element of the opposition, every batsman or bowler - what he does, what he does best, what are his release shots, and stuff like that.
The entire process is about analysing a player, identifying his weakness and strengths and working a plan around it and executing that on the field. There are four stages that go into that. Every cricketer learns in a very different way. We have designed delivery models where this information can be digested by cricketers at every level.
SJ: Let us say you are breaking down an opposition. For example, in preparation for a big Test series, would you be putting together data on the strengths and weaknesses of all the Indian players, perhaps based on what the opposition is expected to do, and similarly from the opposition based on what India is expected to do?
DJ: We have a highly skilled and dedicated team that we have put together that is involved in information collection across the world. We are involved in tracking about 400-500 cricketers.
Before a series me and my team put together information on the opponent and that involves video footage. and more data that we have collected about a player over a period of time. That data gets converted into intelligence through software platforms that we have built over the years.
That gets into the analytics mode, where it throws up a lot of insights on every player that we are going to face. This is taken into the preparatory process of a team, which means that even when they train, they take these insights into play.
For example, if I am going to play a Morne Morkel, a batsman is already given information about what he does, his instances of bowling a bouncer every three or four balls, and if he is hit for a boundary in a particular ball what is his follow-up ball, and all that stuff. It gives the batsman the best chance of success in terms of preparation, going into the game. Whether he does it or not is secondary, but all these insights power their instincts, in terms of expecting something that can happen and pre-empting that and giving them a competitive edge.
SJ: How is all the information delivered to the player? Do you sit together first with the coaches and discuss? Is there a long planning session, and then piece- wise deliver it to the players? How does it work?
DJ: I think you have answered the question quite a bit. There are quite a lot of meetings that happen in terms of planning and preparation. Obviously there is a deck that is taken to the coaches and the coaches take some things along to the player in their training sessions.
In terms of the strategy sessions, we do that with the players. We go into the specifics of batsmen and bowlers. In terms of fielders, we do it in one full team meeting, where some insights about who is bad at running between the wickets, number of instances of a player getting run out - we pick all those stats and keep it in front of them, so that when a chance arises the guys know that they have a chance here. The insights are as intricate as the fact that someone plays an on-drive in the first five balls.
We also have reports on "hit-and-run" case. There are many cricketers in the world today who like to hit and run, and we have analytics on that, so you know that if they hit and run, there is an opportunity for a run-out.
With regard to the delivery to the players, today, delivery in terms of data and video is getting better. You have smartphones now. Cricketers are embracing technology, they are getting comfortable with it.
We have an Athlete Management System that houses all the information, and every player is notified if there is any observation or any report on anything that is being given and he can access it on the go. It is quite simple in terms of what we put together, but it is quite a powerful process that is being set over ten years and deployed since 2003.
SJ: How did you become the performance analyst for the Indian team?
DJ: Well, to be really honest, I don't have a superlative story to tell. It is interesting because I share the same passion for the game that millions of people across the world have for cricket in India. Yes, like your picture on the blog, I started playing cricket exactly in that same spot. I hail from Chennai, a place called Besant Nagar. My cricket started on the beach. I was this ridiculously focused cricketer till I was 22. I played a lot of league cricket in Chennai, aspiring to play for the state and to play for the country. I always thought that I had the attitude but maybe I lacked some skills.
I was ridiculously focused on staying connected to the game. So I decided to do an MBA. I worked as a management trainee with Cricinfo for two years and studied at the same time.
Along the way I met Ramky, who was my cricket coach when I was in juniors. Ramky was the analyst of the Indian team. I have never looked back since. To join the Indian team was quite a journey. I was mentored by Ramky for a long time now, for more than ten years. At some stage I made myself eligible for this role, if I can use that word. Obviously I had a very strong cricketing domain, but I had to top it up with analytical skills as well as man-management skills.
SJ: There are some listener questions as well. Saurabh wants to know if you are required to provide real-time data as the match is happening or is it more of a pre-match-preparation role?
DJ: Yes, real-time feedback happens. It is passed on through the hierarchy, if I can say that. As we have gone along, the real-time feedback and decision support systems have become more and more key in a game like T20, where rapid decision-making and quick changes are required.
We have developed a lot of predictive tools that can help a team align itself to a score during a T20 game. We use a lot of statistical modelling - right from linear programming to the fact that we also use stochastic methods to derive predictive insights that can help team get a competitive advantage in a real-time situation. In a tournament like the IPL, it is possible to add a lot of value being there in the dugout and actually trying to change the course of the game and getting a competitive edge out of it.
Let's say that there are programs and models available where we can predict a lot of things at the end of sixth over so you can align yourself to the best possible target or reduce the opposition to a score depending on the venue and game and the strengths and weaknesses.
SJ: Another question from listener Kartikeya: You spoke of a lot of match situations. Do you also monitor the nets sessions systematically?
DJ: Well, yes. That is non-negotiable. Where we come in, we come in with a lot of strong processes. It generates a lot of information - be it at practice, at the game, at the gym session, or training sessions. Documenting and tracking those is of high importance, because every bit of information is important. So what we do in practice is regularly monitor, track and document for the coaches to keep reviewing a player and make sure that whatever sudden or small changes creep into technique are identified and corrected - as against the cricketer realising after a string of failures.
All these video-analysis and performance-analysis tools help your accelerate performance or help you accelerate a particular skill set. If you were working on a drill 20 years ago, it would take six to seven days to get on top of it. With the latest technique today, you can accelerate to it in about 24 hours. That is a straight increase in productivity. Everything around a cricketer today is getting smarter and smarter and I think even we are moving in that direction. The term "performance management" has taken over, rather than us calling ourselves analysts, because we are involved in tracking their entire performance, even in terms of tracking their lifestyle changes.
SJ: For a particular bowler or a batsman, having delivered a particular ball or played a shot - are you getting the information based on what is displayed on the TV in terms of Hawk-Eye and stuff, or do you log it using your own set of tools?
DJ: We have our own set of tools and processes, to the extent of which, if you bring in the case of baseball, you have good analytics to count the number of pitches and the velocity of the throws and things like that. Similarly, in a cricketing environment you can replace them with similar insights. What is shown on Hawk-Eye and the television-based analysis is more targeted towards the fans and the commentators.
But for gaining a competitive advantage over another team or a player, our analysis is completely different. Stuff like x batsman is uncomfortable against a left-arm spinner, and the fact that he gets out to a left-arm spinner every six balls - that is the kind of stuff that we come up with. Our engines are smart with regard to that.
SJ: There is another one from our listener Aditya Baliga. He wants to know how players react to these insights. Do they react to the insights better when they are doing well or not doing well? Or vice-versa? Or they are neutral to it?
DJ: Everyone is different. I think our processes are such that it allows a player to analyse success as well as failure. The default mechanism in all walks of life is that we analyse a lot of failure. We flipped the process and focused on analysing success, which means that the player becomes better and better in terms of that particular skill or thought in terms of what he can do. So, yes, there is a good equilibrium of players who come by default and analyse failures, and some players who come and build on their success and analyse their success more and more.
SJ: There is a question from listener Ravikiran. Have there been players who have been mistrustful of the analytics and didn't want to take whatever it says on board?
DJ: Well, interesting question, because I think "mistrustful" is a harsh word to use. I think we are there because people trust us and what we do and what we bring to the table. I would like to use the word "resistance" here. There has been some resistance in taking some of these things on board, because of the fact that somebody has been successful doing something for years together, and to come and say that they need to do something different is difficult. To take anything new on board is always met with resistance. Obviously, we have been persistent enough in education and creating awareness about analytics, be it amongst the coaches or the players. We have found a lot of success in terms of people taking this as a part of their preparation.
What we want to do is strengthen the process of the preparation model that we created. It is a very strong indicator of players taking it into their preparation process. I would say, yes, there is always an element where, let's say, an insight or an analysis has gone wrong here and there, but, in terms of the hit rate and success, it is definitely more than 75% in most cases.
SJ: Let's take these two situations where MS Dhoni took matches in the recent tri-series in the West Indies all the way to the 50th over. Is he just playing on his instincts or are his instincts being augmented by what has been fed through your analytics? Have there been messages saying, "Listen, you have better chances of success against so and so"?
DJ: I think I can only power their instincts and they experience it on the field what a bowler does. When Malinga is bowling, everyone knows that Malinga bowls yorkers at the death. What we bring to the table is options, in terms of how to play him, or how to play him best, and how to get the maximum out of playing him.
MS Dhoni's approach to the game - he has his own options. What I bring to the table is obviously giving him more options in his decision-making. If he makes better decisions, I would say that I would have contributed 1% of that. But unless that 1% is there, that would not have added to 100.
We have seen a lot of success, we have seen a lot of instances where people have recognised that this insight actually helped me make a play. That is up to the player to say that.