Ball-by-ball Data Sharing
The number one asset required for my analysis is data. More specifically, the number one asset is ball-by-ball data that describe the outcomes from each delivery in T20 matches. When I first started, all that data was sourced from Cricsheet who, at the time, had ball-by-ball data for all T20 Internationals and IPL games. It was brilliant, easily accessible, and extremely reliable
Unfortunately, I needed (or wanted) more than Cricsheet could provide. Most importantly, they stopped updating their database in July 2017 and I wanted more recent matches. So I created my own database (helped by the guy who started Cricsheet). At the bottom of this post is a link that downloads slightly more than half that database
Whilst a lack of alternative was the only real factor driving me to construct my own database, there were other benefits too. One improvement is that I can cover a lot more tournaments using my own code. My new database covers 4,170 T20 matches across 30 different competitions. From the first international in 2005 to the Champions Twenty20 in 2009, the Karnataka Premier League in 2017 and the most recent IPL season
The downside is that I now have sole responsibility for the accuracy of my data. The old Cricsheet data was probably more reliable but I find my data robust enough to give me the insights that I need
My aim here is to facilitate others who are interested in sports analysis or who want to use their maths and coding skills to impress cricket clubs. I want to enable the creativity of others and I'd love to hear back from anybody who uses this data to do something cool. Email me if you want to discuss your ideas
The download below contains four files - ball info, match info, player info, team lineups. They are all in CSV format. There is no data dictionary and no further explanation. It should all be fairly obvious. Included are all the matches that I have from the IPL, Big Bash, CPL, PSL, BPL, Super Smash, Blast, and Champions League. The download contains information from 2,257 matches and 530k individual deliveries. I have no plans to regularly update this