Freddie Wilde introduces CricViz’s new Batting Profile analysis based on three new metrics: Attack, Timing and Power.
What makes a batsman the player he is? It is a broad and expansive question – one that incorporates multifarious traits and characteristics. However, we can get closer to answering it by using detailed data analysis which breaks a batsman’s game down into its component parts.
At CricViz we have introduced attacking shot percentage – which has helped evaluate the intent of the batsman – and false shot percentage – which has helped evaluate their control.
These measures are based on data recorded by Opta analysts for shot-type and shot-connection. We then group these shot-types into the attacking, rotating and defensive categories and into the played, missed and edged categories, from which we can produce attacking shot and false shot percentages.
The scatter graph below plots these two measures against one another. Players in the bottom left quadrant don’t attack much but have high levels of control; players in the top left quadrant don’t attack much and aren’t in control often; players in the top right attack a lot but are rarely in control and players in the bottom right attack a lot and have high levels of control. The scatter shows how using attacking and false shot percentages we can begin to get an idea about a batsman’s method and effectiveness with that method.
Now, we have taken these metrics and elevated them further – producing three more advanced measures – Attack, Timing and Power – built by CricViz Lead Data Scientist Sam Green. Taken together these three figures help paint a more detailed profile of a batsman.
The attack metric enhances basic attacking shot percentage by distinguishing between all shots – not simply sorting them into groups. A slog shot is clearly a more attacking shot than a drive for example – but under the basic categorisation they are both simply considered as ‘attacking’.
So instead we have created a more advanced measure which is based on evaluating their attacking-ness according to their overall strike rate. So for example a pull shot – with a strike rate of 194.22 is more attacking than a drive shot with a strike rate of 161.22 which in turn is more attacking than a cut shot with a strike rate of 123.45.
The attack metric is therefore the overall strike rate we would expect the batsman to score at based on the shots they have played and if they had made a good connection with those shots. It is important to assume that they have made a good connection because the measure is only seeking to evaluate intent which remains independent of execution. As a result the Attack rating – or expected strike rate – of batsmen will almost always be higher than the strike rate of that player because not even the world’s best batsmen make a good connection with every shot they play.
A comparison of CricViz’s basic attacking shot percentage and the new Attack rating shows how there is, unsurprisingly, a strong correlation between the two figures with a high attacking shot percentage generally translating into a high Attack rating.
However, a closer look at the outliers reveals more about the measure. For example Andre Russell’s Attack rating is considerably higher than his attacking shot percentage might suggest based on the regression line. This is because when Russell attacks he plays exceptionally attacking shots – for example slogs, hooks and pulls – and when he doesn’t attack he plays very few defensive shots. This elevates his Attack rating higher than we might expect based on his attacking shot percentage. In contrast Shakib Al Hasan’s Attack rating is lower than his attacking shot percentage suggests it should be – this is because the attacking shots he plays are less attacking than most players and he will play more defensive shots. Sam Billings has a similar Attack rating to Shakib despite having a significantly lower attacking shot percentage – this is because Billings plays a lot of slog, conventional and reverse sweeps, shots that are considered exceptionally attacking therefore elevating his attack rating significantly higher; the same can be said of Sarfaraz Khan.
Attack rating is a more advanced measure for evaluating batting intent than attacking shot percentage. The list of the batsmen with the highest Attack rating since the start of the 2017 season is below.
Timing is a more advanced measure of the quality of contact that the batsman has made than false shot percentage. The basic false shot percentage is simply the percentage of balls edged or missed by the batsman. Timing elevates this measure by contextualising the contact made by the batsman according to the format, the shot-type and the bowling type.
We do this by comparing the overall shot average for all players with that shot – for example a drive shot against a pace bowler in T20 has an average of 25.90; with the average for that shot given the contact made by the player – for example a well-timed drive against a pace bowler in T20 has an average of 323.17. The player’s Timing rating is the ratio of their contact average with the shot average; multiplied by 100 – so in this example it would be 1247 (12.47*100). These figures are aggregated across all shots played by the batsman to produce an overall Timing rating.
As with attacking shot percentage and Attack rating; false shot percentage and Timing rating are strongly correlated with a lower false shot percentage translating into a higher (better) Timing rating.
However, as with attacking shot percentage and Attack rating the outliers are illustrative of how the measure works. Sunil Narine, for example, has a false shot percentage of 33% but his Timing rating of 106 is slightly above what we might expect and this is because Narine is a very attacking batsman and plays shots that typically have a high false shot rate. At the other end of the spectrum, someone like Hashim Amla has a higher Timing rating than his false shot percentage might suggest, this is likely to be a consequence of the fact that he makes a lot of very good connections and when he edges the ball they are generally thicker edges – in other words he is being beaten to a lesser extent than, say, Axar Patel who has a similar false shot percentage but a far inferior Timing rating.
Timing rating is a more advanced measure for evaluating batting contact quality than false shot percentage. The list of the batsmen with the highest Timing rating since the start of the 2017 IPL season is below.
The third and final new measure is Power which sees CricViz explore beyond intent and contact and look at a player’s ability to find and clear the rope when they make good contact. Power looks specifically at shots when the batsman has made good contact and is a ratio of their well-hit boundaries to the boundary-rate of those shots. For example, when Kieron Pollard makes a good connection with a drive shot against a spinner he hits a boundary every 1.85 balls. This number is then weighted towards sixes which are worth more and then compared to the overall shots per boundary for all players with good connection drives against spinners of 3.49 balls per boundary, which is also then weighted towards sixes. As with Timing figures are aggregated across all shots and multiplied by 100 to produce a Power rating.
It is important to stress that in this form Power is not a measure of how far or how hard the batsman can hit the ball – something we can categorise as Raw Power; but instead it is a measure of how effective the batsman is at hitting boundaries, and more specifically, sixes – something we can categorise as Game Power.
There is clearly a strong correlation between boundary percentage and Power rating. In the scatter below those players above the regression line are typically those who are better at hitting sixes or finding the boundary from their shots.
The list of the batsmen with the highest Power rating since the start of the 2017 season is below.
Attack, Timing and Power – in that order because they are sequential – are interesting measures in their own right. But it is when they are presented as a three that a profile of a batsman can be formed.
Across the last two and a half seasons of IPL the average Attack, Timing and Power ratings are 165, 117 and 116. Based on this the batsman closest to the IPL average according to these three measures is Vijay Shankar.
Russell is an exceptionally dangerous and aggressive batsman – combining the IPL’s highest Attack rating with the second highest Power rating. His Timing rating of 132 ranks him 25th in the league which is not anywhere near as elite but is still very good. Russell’s main weapons are his intent and his effectiveness at hitting boundaries.
Shakib Al Hasan is a player who bats very aggressively – with an Attack ratting of 173 and his Timing rating of 127 is solid. However, he massively falls down in terms of Power with a rating of just 70, which is the 74th best in the league. Shakib’s Attack rating reflects his role in the lower order but his Power rating suggests he is not up to the task.
Where Shakib is effective in two areas, Ajinkya Rahane is effective in just one – Timing. But in terms of Attack and Power he falls well short of the league average and this is often reflected in his T20 innings where he often looks elegant but lacks the aggressive game to score at higher rates.
Shubman Gill’s potential as one of the most exciting players in India is confirmed by his exceptional Timing rating of 175 – the second best in the IPL. His Attack rating is relatively low but a Power rating of 104 is encouraging for a young man who is only going to get stronger.
Rishabh Pant’s Batting Profile shows him to be one of the most complete batsmen in the IPL. Indeed no batsman’s profile is further from the average player than Pant who combines exceptional Attack, Timing and Power ratings.
Combining all three metrics on a scatter allows us to illustrate how batsmen play in greater detail – elevating the attacking shot and false shot percentage scatter at the top of this article.
Freddie Wilde is an analyst at CricViz. @fwildecricket