Why limited-overs cricket needs to measure boundaries attempted

The things that are measured in cricket are typically self-evident. Interpretation is not required to count runs, wickets, deliveries, extras, boundaries or sixes. Other sports have embraced measurements that require interpretation. In tennis a distinction is made between a "forced error" and an "unforced error". In baseball, "earned runs" and "unearned runs" are distinguished by judging whether or not a safe hit was the result of a fielding error or not. Baseball also has the concept of "defensive indifference". In American football, things like "decisive passes" and "assists" and "tackles" are counted systematically.

ESPNcricinfo records a "control" measurement for every delivery added to its database. This measure is based on the answer to the question "Was the batsman in control of the ball?" Two answers are possible - yes (in control) and no (not in control), but it is not always self-evident whether or not the batsman was in control or not in control. This is probably the very first attempt in cricket to develop a measure that involves the systematic use of judgement.

If the bowler beats the bat, finds the edge, or induces a miscue, then even if no dismissal results, the batsman is said to be not in control. A lofted drive off the middle of the bat that results in a catch at the boundary is not in control. As readers will notice, these conclusions require the exercise of judgment. It follows that some judgements will be easily arrived at, others not so much. A classic example is when a batsman is tested just outside off stump and starts out playing at the ball but withdraws the bat at the last minute. Was the bat withdrawn in time? Whether or not the batsman is judged to be in control of the delivery will depend on the answer to this question.

Currently this measure is applied to all types of cricket. This post is about what the control measurement tells us, and what it doesn't. At the end, I propose that T20, and perhaps the limited-overs game generally, requires a second measure to go with control.

Here is a summary of the control measurement for the 2016 IPL, based on ball-by-ball details provided by ESPNcricinfo. The summary is prepared as a proportion of 120 balls faced. For instance, the average losing team in the 2016 IPL was in control of 88 out of 120 balls and scored 132 off those deliveries.

An earlier analysis shows that on average, the batsman is not in control of one in every four deliveries faced in T20, and this figure holds for the 2016 IPL. The scoring rate off these not-in-control deliveries in the 2016 IPL is about three runs per over. Most dismissals occur when the batsman is not in control (this is due to the way the control measurement is defined).

In the 2016 IPL, of the 90 or so deliveries in a T20 innings in which the batsman is in control, winning teams that set targets scored 34 runs more (165) than defeated teams that set targets (131). Winning teams scored, on average, two runs per over faster off those deliveries.

For chasing teams, the scoring rate in control was about the same (about 9.5 runs per over) in wins and defeats, but winners remained in control for ten balls more than defeated teams.

Put another way, when a team bats first, success depends on their ability to hit as relentlessly as possible. On the other hand, for a chasing team, success depends on hitting just enough for the longest possible time.

Consider a player who makes 50 (off 30 balls) against a spread-out field while being in control for 24 out of those 30 balls faced. This player is in control 80% of the time. By T20 standards, this is above average control. What if the total came with five sixes, one four, 16 singles and eight dots? Against a spread-out field, singles are uncontested. Even the fielders inside the circle stand on the edge, happy to concede one in the hope of saving a couple of extra boundaries. It would be useful to know whether the singles were the result of hard hits that failed to find the boundary, or whether they were simply pushed to the fielders in the deep. While this type of innings may make sense in a chase, it is less valuable when batting first.

The control measurement does not tell us this part of the story. In a Test match, this is not particularly significant. Risks are taken far less frequently in Test cricket. Wickets are the single most significant resource by a large distance. Since most Test matches that yield an outright result end before the full 15 sessions of play are completed, it could be reasonably argued that wickets are the only significant resource in a Test. Consequently, the ability to score runs reliably is paramount because it suggests that the player can be relied upon to make runs consistently.

In T20, higher control is not necessarily better. Most teams do not use up their full quota of wickets in 20 overs. They could afford to lose a wicket or two chasing a few extra boundaries. In order to measure this effort, the control measurement needs to be accompanied by a hitting measurement. One approach to this measurement would be to provide an answer to the question "Did the batsman attempt to hit a boundary?" The answer ought to take into account the field setting among other things. Perhaps the measurement could even be termed "boundary attempts".

Instead of averages and strike rates, T20 batsmen could be measured by how much control they are able to enforce (or in the case of bowlers, how little they are able to concede) and how many boundaries they attempt to hit. For example, take a batsman who attempts a boundary once every three deliveries, or a bowler who forces batsmen to be not in control on 11 out of 24 deliveries. This could then be extended to consider the boundary attempts per dismissal, and the boundary attempts per balls faced. Much like hitters in baseball go through hot streaks and slumps, the form of T20 hitters could be measured on the basis of these measurements. A derivative measure could be "control on boundary attempts" (since not every not-in-control delivery produces a wicket).

Together, control and boundary attempts (or hitting) represent the trade-off in the T20 game and could provide a language to describe T20. If T20 is to emerge out of the mother game's vast shadow, such a language, which permits the measurement of merit in T20 on its own terms, is vital for T20 as well as for cricket in general.