IPL And Big Data Analytics: A Match Made In Heaven



IPL and Big Data Analytics: A Match Made in Heaven

Raise your hands if you were born somewhere in the 2000s. Okay, now raise your hand if you have (voluntarily or non-voluntarily) watched cricket on one of the country’s leading sports channels. Now, clap you were at least mildly intrigued by the IPL hype (at least during the starting years). And, shout like the fake audience of this year’s IPL if you were impressed each and every time by the new run cum ball count board at the bottom of your TV screen for every new edition of the Indian Premier League. 

That vibrant and impressive bar you see every year is nothing short of being called a new generation masterpiece—particularly in the field of data analytics, if not in that of cricket.

So without any further ado, let us dive straight into the AI-backed analysis, different kinds of metrics, and methodologies adopted by different brands, and team administrations.

Historically, the big data analytics finds it’s use only in the last few decades, where it was largely adopted to show off economy rates, strike rates, chase precisions, and whatnot. However, it was only the starting of this decade, particularly with the boom of shorter formats like the IPL that it found its true calling. What started out as a simple scorecard to keep track of runs and balls, is the deciding factor when it comes to choosing a player for the team.

The big data analytics offered to provide a win-win situation for all parties involved—be it the team, the audience, the umpires/on-field decision-making personnel, or even the sponsors for that matter. This is because the analysis provides:

·       Choice in players and team formation.

·       Various approaches (algorithms).

·       Improved viewer engagement.

What are you waiting for? Let’s start the innings.

Choice in players and transformation:

In today’s time and age, the competition in cricket is no longer about how many runs a batsman scored, or how many wickets the bowler claimed. Specifically, in the Indian Premiere League, it is all about how valuable an asset the player can be in the long run. True that the concept boils down to “balls vs runs”, it’s a lot about other factors too. Be it the current reputation of the player, experience, or even other political beliefs (the harmful kind), the selectors always try to make a calculated list preference prior to the official auction of players. The list of potential candidates is often determined by utilizing certain domain-specific metrics. These metrics mainly include the Batting Metrics and Bowling Metrics, among many others.

These metrics in turn constitute the weighted composite of a particular player. This is also known as the Most Valuable Player Index (MVPI).

Batting Metrics:

·       Hard-hitting abilities = (no. of fours + no. of sixes) / total balls played

·       Finishing Abilities = (no. of not out innings) / total innings played

·       Scoring Abilities = Total Runs / Balls Played

·       Consistency = Total runs / no. of times given out

·       Between wickets running ability = (total runs - (no. of fours + no. of sixes)) / (total balls played – boundary balls)

To demonstrate with an example, one can tell how good a player like Chris Gayle is when it comes to hard-hitting abilities. While a person like him might score extremely high in such categories, his position might not be that great in situations when it comes to taking 1’s or 2’s. Thus, the low score in the last category.

Bowling Metrics:

·       Economy = no. of runs conceded / (no. of ball bowled/6)

·       Wicket taking abilities = no. of bowls bowled / wickets claimed

·       Bowling consistency = no. of runs conceded / wickets claimed

·       Ability to take snap crucial wickets = no. of times four or five wickets were taken / no. of innings bowled

·       Short performance index = (no. of wickets taken – no. of times four wickets were claimed - no. of times five wickets were claimed) / (innings played - no. of times four or five wickets were taken)

When it comes to demonstrating with an example in this, we can take a bowler like Amit Mishra. The spinner who is not known to make any recent appearances is often referred to as a mastermind when it comes to changing the fate of the team or game. The spinner thus often scores highly in his Short Performance Index.

Different people show different results when it comes to a particular metric. After all, each player has their own strengths or weaknesses. Thus, the overall score is calculated by using appropriate algorithms and formulas that help in determining the fate of prospective candidates on the players’ list. But in the end, the bigger the better.

Various Methodologies:

With all points conveyed regarding the selection metrics in the previous section, now will be the right time to understand what goes on behind the screen when it comes to the selection procedure. With a lot of money on the line, and budgets always running tight, making calculative, yet efficient decisions are highly crucial. What remains even more important is that at least 50% of the players included in the list are checked off.

The following steps are mainly followed when it comes to creating/choosing potential candidates for the team:

·       In the very first step, the 10 metrics mentioned earlier are used to calculate the effectiveness of a player in the 20 over format. Once that is done, they are immediately tallied with previously benchmarked ranges for different metrics. After all, each metric is unique, and so is each player. Keeping this mind, it is important that the players do not get an unfair advantage over others due to domains of their preference and/or respective batting/bowling styles.

Thus, normalization of scores is highly important. The score for the same is calculated as per the following formula:

IPL and Big Data Analytics: A Match Made in Heaven? - Firstpost

 

Normalized score of a particular feature = (players count – rank in that feature/total count of players)

·       In this step, the weights of each metric is calculated by using the feature selection tool. To be more specific, this is decided by the Recursive Feature Elimination technique, which helps by retaining features and narrowing down to smaller attributes concerning the same.

To train this model, initial attribute sets are fed into the system and is followed by rigorous computation of each attribute’s relative importance. Towards the end, unnecessary and least important attributes are instantly eliminated. The process is further followed until all undesirable attributes are eliminated and only the desired ones remain.

·       Now for the last step, an aggregate score is released by multiplying the calculated wights with the respective value scored. In the end, the scores are arranged in a decreasing order.

One important point to remember is the dynamic environment the auction room shows off, and that no result or list is final. You never know when the tables might turn.

 

On a side note, however, players are often requested to realis the nature of the system that’s been adopted off-late, and that no personal sentiment should find itself attached in this profession. At the end of the day, it’s just a couple of professionals negotiating over a deal and simply doing their job.

A selection is based on the circumstances and not personal vendettas.

Improved Viewer Engagement:

Viewer engagement is what major sporting leagues like the IPL cash-in on. It is only due to the war involving TRPs, views, sponsorship deals etc that the entire league exists. So, it will only be right that the audience is offered just the right interface and environment while experiencing such a highly energetic event.  The main highlights of such services that excite viewers of all ages are the distance covered by the ball while when it goes for the boundary, analysis of all sorts of domain-specific rates, overall top performers of the tournament, and even important replays at the right time.

The high tide, however, was disrupted this season due to the ongoing Corona Virus pandemic. With an empty stand replaced by system produced cheers and applause, the disaster was sort of inevitable. Or was it?

The lack of charisma that was often brought out in the game by the audience created a void in the hearts of many. However, to bounce back from the setback, various brands have started coming up with apps that allow them to engage in various fantasy leagues with real-time players. By keeping track of recent performances and records of each player, fans create their own team and show off to friends and family. This is a win-win situation for both parties, as the interest is maintained and viewership is also gained. Even teams have come up with respective apps that allow the fanbase to grow and help the community to reach out to one another. So, every time you open your app, you are reminded of IPL. The same goes for the situation vice-versa. Isn’t this nice?

Wonder who’s going to win the cup this year though. Wait, let me check.