Cricmetric (http://www.cricmetric.com/) provides detailed insight into the performance of Cricket players, and the game of Cricket in general. Cricmetric was conceptualized and founded at the Massachusetts Institute of Technology. The aim of Cricmetric is to become the “Moneyball” of Cricket.
The work of Cricmetric has been along three dimensions: Data, Advanced Metrics and Analysis. We have collected ball-by-ball data of all the International Cricket matches that have been played in the last ten years. In addition to that, we also have match data for popular T20 tournaments, such as the Indian Premier League and the Champions T20 tournament. Having such an exhaustive database allows us to answer many interesting questions, for example, “How does Sachin Tendulkar perform against left handed spinners outside the sub-continent?”
We are also collecting ball-by-ball field placement information for each match which is not available very easily. We are using this data to build a model which would predict the optimal field placement for a particular player in the context of the game and based on historical performance. This model would be very valuable for teams and coaches to build a strategy (both real time as well as before a match starts) as well as for team selection under different playing conditions.
We have also developed several advanced metrics to analyze limited over Cricket matches, and the performance of Cricket players. We have a model that, in real time, calculates the win probability of a team in a limited over Cricket match, based on the historical data of previously played Cricket matches. Using the win probability statistic, we are able to measure the contribution of each player in the match. This helps us evaluate player performances beyond the traditional statistics used in Cricket, such as batting average or strike rate, for example. We have many more such advanced metrics for ranking players and teams on our website.
Using our database and metrics, we are able to provide in-depth analysis such as:
– How do different venues affect the scoring rate in Cricket?
– How can we mathematically define “choking” in Cricket?
– Which bowlers have been most successful against a particular player?
– Given two Cricket players, who is the better player of the two?
Try deep learning using MATLAB