Even in our world increasingly awash with data, Cricket stands out as a sport that generates unparalleled and incessant data. Every delivery, in itself, is an entire database. Yet, as the old maxim of big data goes- garbage in, garbage out.
Steve Smith has racked up extraordinary numbers by all but the big Don’s standards. Given the endless shots of team England and their multitude of laptops, one can only assume that data is readily available to the management, and no doubt numerous analysts are at hand to get the best out of this data. So, why have teams been unable to find a way to get him out? Does Smith’s unconventional style transcend data analytics?
Two things in life I love- data analytics and Cricket. So, I took a quick glance at data that was readily available and tried to see if I could come to any conclusions of how to get Smith out in test matches. Here is what I found.
Smith is susceptible to the ball moving away from his outside edge
This may seem an obvious point, however, the emphasis is on bowlers that naturally move the ball from leg to off. Smith has been dismissed 95 times in his career. Of these 58 (61%) have been by right-handed bowlers and 34 (36%) by left-handed or leg-break bowlers.
Given the greater proportion of bowlers in world cricket that are the former rather than the latter it suggests a weakness in his technique to such bowlers.
Closer inspection of the types of dismissal is also telling. To right-arm bowlers 34 (59%) of his dismissal were at or behind the wicket (bowled, LBW, stumped, and caught behind). In comparison, 28 (84%) of his dismissals were behind the stump to left arm and leg-break bowlers (p = 0.02, chi-square test). When left arm bowlers are considered on their own this figure goes up to 88%, suggesting Smith is being beaten by left-handed and leg-break bowlers more often. Interestingly, his average against right-handed bowlers that have dismissed him is 50.5 compared to 46.2 against left-handed bowlers.
Now looking only at bowlers who have got him out 2 times or more. Table 1 is a list of these bowlers and how many times they have dismissed Smith and the number of innings he has faced against them. Left-arm or leg-break bowlers have taken his wicket every 0.54 innings, i.e every other innings he has faced them. Whereas, for right-handed bowlers have taken his wicket every 0.23 innings i.e. every 1 in 5 innings (p < 0.0001, Figure 1). This is all the more surprising given right-handed bowlers outnumber left-handed bowlers by 3 to 1 in most teams, if not more. His average against left-handed bowlers in Table 1 is also lower (37 vs 43), although this fails to reach statistical significance.
Is this a real phenomenon?
In biological sciences, the question we often pose of observational research is whether there is biological plausibility to our findings. So, is there a cricketing plausibility or rationale to these findings? Here is my hypothesis. In test cricket the biggest threat to the batsmen is his or her outside edge. Steven Smith with his unique and quirky batting stance and style, in conjunction with being one of the all-time greats at playing the ball late, has almost completely negated the risk posed to his outside edge to balls coming into him.
He barely scores down the ground at long-off- hardly surprising given before facing a ball his bat is angled towards long-on and with his big stride he probably takes a fourth stump guard. In essence, Smith has incorporated the corridor of uncertainty into his hitting zone. This is more so with bowlers that naturally move the ball from off to leg- top of off stump and 4th stump is effectively bowling at his pads.
Left-handed bowlers naturally drift the from leg to off forcing Smith to, at least consider, if not actually play straight on the off side. This brings his outside edge into play- increasing his susceptibility to being dismissed behind the wicket.
Right-handed bowlers could choose to bowl around the wicket but Smith is too patient and too good to fall for that. In addition, it completely takes out LBW and bowled from the equation unless you have searing pace. In swinging conditions, right-arm away swing would have a similar effect as left-handed bowling, as evidenced by his modest record in swinging conditions in England. His performances have been exceptional only on the relative flat pitches of London.
Slower balls by right handed bowlers (leading edge) are also probably under-used against Smith. I also think leg-cutters (outside edge) may be also be effective against him. This remains mostly speculation. From right-handed bowlers I have watched, I would love to see Venkatesh Prasad and Terry Alderman bowl at him. Of course, Wasim Akram would be the guy you would really want to throw the ball to, but that would stand true against almost any batsmen. I would love to know how he gets on against Starc compared to other bowlers in his national team nets.
The bottom line is, against Steve Smith and his Australia every side needs at least one left-handed or leg-break bowler, preferably two. A left-handed bowler in your side means he will get Smith out every other innings. A side without such a bowler is akin to going to a sword fight with a rhubarb stick (Boycott’s granny notwithstanding). And, if you don’t believe me or the stats, just ask England.