With the surge in Covid-19 cases in India and players from different franchises testing positive, BCCI had very little choice but to postpone the league indefinitely. But Jain couldn’t wait for the resumption of the league and decided to run a simulation program to find the winner.
The model factors taken into account to run the simulation on a ball-by-ball basis are different types of pitches, match-ups and recent form and record of teams and players.
“I made a Python program to simulate an entire season of the IPL (excluding the playoffs) using past data, predictive analysis, and randomization techniques. I mainly wanted to share this because there are some really interesting scenarios in the scorecards or some epic batting collapses, also I think RCB fans will like this one,” he wrote on Reddit.
“I extracted each player’s past five years’ data for batsmen and bowlers. What kind of runs a batsman scored – singles, doubles, fours or sixes – what kind of runs the bowlers leaked, against what bowling did a batsman thrive, when they scored their runs – powerplay, middle overs or death – what mode of dismissal bowlers dismissed batsmen the most with and vice-versa how batsmen got out. This includes detailed data such as frequency of wides or no balls or the average catches taken by a player in a game,” Jain was quoted as saying in a report on Wisden.com.
Here is how the league stage ended:
Qualifier 1 – DC v RCB (RCB won)
Eliminator – CSK v PBKS (PBKS won)
Qualifier 2 – PBKS v DC (DC won)
Final – DC v RCB (RCB won)