Mock IPL auction.
Most of my friends, including me, are interested in cricket. The corona crisis has made most of us bored during the weekends and we wanted a way to spend our time away. Few of them came up with an idea of holding a mock auction where they give a rating to each player based on stats in IPL( Indian Premiere League). This rating won't be revealed to anyone till the end. The idea is to have 8 teams (comprised of 3 people per team) put up a lineup of 11 players and the one with maximum rating wins.
- Each team can only select 11 players.
- Batsman - Min 3 Max 5
- Bowlers - Min 3 Max 5
- All Rounders - Min 1 Max 3
- Wicketkeepers - 1
- Max Spending limit - 30 crores. 7 . Overseas Players - 4
Around 240 players , throughout the history of IPL, were chosen by them and a list was released indicating their roles(BAT/BOWL/ALLROUNDER/WK). The only clue about the rating that was given was that it was derived from stats in their own way
Being a data scientist , it was right up my ally and i wanted to build a system that could help me in choosing the best players .
- Scraping -Scraped through the ipl website for stats of all the players from 2008-2019.
- Pre-Processing - Created a heuristic rating from all the stats
- Optimization - Used PuLP package to solve an Binary Integer Programming optimization problem.