Roshan Prizak - How have different IPL players contributed to their team’s winning chances?

Introduction

ESPNCricinfo has published win probabilities after every ball bowled in each match of the IPL. This information, together with ball-by-ball information on the batter identity, bowler identity, ball events (which over, which ball, runs scored, wicket taken?) and other relevant information, opens up the possibility of assessing each batter’s and each bowler’s contribution to their team’s winning chances in each IPL match. This project starts to dig in that direction.

Some questions we’ll probe here are:

Which batters and bowlers (players) have contributed the most to their team’s winning chances over the years?
Which players have contributed significantly higher in wins compared to losses? Which players have contributed equally in all matches independent of the outome?
How does the batters’ and bowlers’ contribution change with match scenarios - over number, ball number, wickets fell, bowler type, setting vs chasing, wagon zone?
What does a typical innings look like for each batter?

But before that, a short note about ESPNCricinfo’s win probability index, or any other win probability index for that matter. First, this is computed from some model that is presumably trained on a lot of cricket data, and should take match conditions, pressure, context and such things into account. Second, one can think of the underlying model as a super smart cricket pundit who knows their stuff. Having said that, there would be many instances where individual people will differ in opinion from what the model suggests. Third, my guess is that as a match progresses, the “accuracy” of the model (supposing there is a super model that knows the True win probability every ball) on average increases. It is with all these assumptions, caveats and understanding that we’ll use this win probability index to understand the contributions of batters and bowlers in the IPL.