Regression Is The Secret Of CSK's Energy

India ka tyohaar, the Indian Premier League is ready to kick off preparations for its 12^th instalment. The auction is set to come about in Jaipur this December. Teams have released as many as 12 players each after the reshuffle auction last year which only means that we’re in for some more exciting bidding this year. If you’re a fan of the sport, you must have watched the auction like a maniac and millions of questions must have popped into your mind regarding its strategic perspective. You must’ve commended a team for buying a proficient player for the price of peanuts or bawled at them for spending a fortune over someone who isn’t even worth it (Yes Delhi Daredevils, I’m looking at you).  Ever wondered what led to these decisions of theirs?

Team strategists and analysts use the regression model as one of their primary techniques of picking up players. Regression, that one statistical tool every statistics student has loved, hated and tolerated through multiple semesters. It helps in determining the outcome of the dependent variable based on the values of the given independent variables, analysing the direction and strength between the X-Y duo.

The team analysts start off by selecting few variables from a prolonged list including the age of the player, number of runs scored, batting average, batting strike rate, number of half-centuries, number of centuries, number of wickets taken, bowling average bowling strike rate, bowling economy, fielding, catches and stumpings etc. These aforementioned independent variables vary from team to team whereas the final bidding price i.e. the dependent variable (Y) is fixed. This meticulous selection of X variables differentiates an efficient strategy from a weak one. These variables can be divided into three categories that are to say cricketing characteristics (batting and bowling performances), personal characteristics (age, country, home ground, positions of responsibilities etc) and off-field characteristics (glamour, fan following etc). An aggressive Indian opener who has 7m followers on Instagram is definitely a great choice. Great job there on the Rs. 15,000,000 Rohit Sharma retain, Mumbai Indians!
      
Final bidding price (Y)= β₁ + β₂*Batting strike rate + β₃*Batting average – β₄*Bowling strike rate
This sample regression model determines the final bidding price (dependent variable) based on the given values of batting strike rate, batting average and bowling strike rate. This connotes that the team’s game plan focuses on collecting good batsmen who possess the capability to bowl a decent over once in a while.

Choice of explanatory variables depends on a handful of factors identical to the requirement of number of batsmen and/or bowlers in the team. Probably the dire necessity of a fast bowler compelled Rajasthan Royals to buy Jaydev Unadkat for a whopping 11.5 crores. The selection of X variables is also affected by the team’s desire to include emerging talents or experienced players. The 2010 Rajasthan Royals or the 2018 Delhi Daredevils, both known for their promotion of young talent, perhaps had a regression model which focussed on the age of a player and their first-class cricket statistics the most. Selection of local players is also an important element in this process as 50% of a team’s matches are played on the home ground which gives the native contestants an upper edge as they are accustomed to the pitch conditions. The first edition of the IPL made sure that captains of most of the teams were locals. 

Looking at last year’s huge shake-up auction, out of the top 10 highest-paid players most of them were the big names of their country’s national teams. Only a few heroics by Chris Lynn in the Big Bash League proved enough for him to bag a contract of 9,600,000 Rupees. Another man who shared that price was Manish Pandey who joined the Sunrisers Hyderabad (SRH), surprising to many. SRH were probably looking for a middle-order batsman to strengthen their middle order owing to the fact that David Warner was absent in 2018. On the lower side of the payments, the largest share was uncapped Indian players like Sachin Baby and Shreyas Gopal. However, West Indian star Chris Gayle was sold for a mere Rs. 2,000,000 to KXIP perchance due to his age or lack of opportunities because of the West Indian cricketers’ feud with their cricket board. By merely eyeing at the auction results, we can try to deduce the explanatory variables the teams must have used, justifying the price paid to the players.

Obviously, there go many months of collecting and studying the data of sundry players, local and foreign. Statistics of the player’s performance in the year before being given more importance while analysing the data. Players who have had a good run in the previous year are likely to procure more than the rest. Ambiguity still hovers over studying all five formats of the game (T20Is, ODIs, Tests, First class and club cricket) or just the shorter versions. However, this is untrue in the case of the IPL XI aka the IPL models (eg: Mahendra Singh Dhoni, AB de Villiers) who have made their mark for many years in the past and are exempted from one year of poor performance.

So, here’s what you can do before the upcoming IPL auction for this season- become an expert yourself and make a few predictions about how much money Glenn Maxwell or Muztafizur Rahman will pocket this year!