May 31, 2016

Two myths about the IPL

Is Kohli as good at T20 as he is cracked up to be? And did the better bowling side win this year's title?

Who's the better T20 batsman? © BCCI

The 2016 IPL was a record-breaking tournament for India's Test captain. Two tropes about the season have become apparent. First, that Virat Kohli is a great T20 player. Second, that the best bowling attack in the tournament defeated the best batting line-up. Both claims are arguably false.

Let's consider the first. Kohli was not even the best player in the Royal Challengers Bangalore side, let alone in the IPL. He seemed to attempt to make the transition from batting to hitting during the course of the tournament, but he never quite shed his classical training as a batsman. In the final, for example, he made 54 off 35 balls. He used up more than 25% of the deliveries available to his team and scored at less than the overall asking rate. He made only 15 off his first 18 balls. A true-blue T20 hitter would not have waited 18 balls to tee off when faced with an asking rate of ten runs per over. The problem with waiting that long is that one has to then pull off slogs for much longer in order to break even.

When the asking rate is ten runs per over over 20 overs, there simply aren't many options. This is not a problem of the imagination, it is a compulsion of arithmetic. The great difference between Kohli and AB de Villiers (the best player in the 2016 RCB side) is that de Villiers can tee off from the start if need be. De Villiers found the boundary once every 4.3 balls. Kohli was a whole delivery slower, at 5.3 balls per boundary. Each hit almost the same number of sixes through the competition (37 for de Villiers, 38 for Kohli), and Kohli faced 233 extra deliveries.

By all conventional measures of batting, Kohli on current form must be considered ahead of de Villiers. He made more runs at a better average and was dismissed less often. But in a T20 game circa now, 11 de Villierses would beat 11 Kohlis, on average by 20 runs.

It is clear that Kohli is an exceptional batsman, but batting - the art of building innings and navigating different types of bowling on different types of pitches reliably and scoring big runs - is a distraction in T20. In 16 matches, Kohli brought RCB 973 runs for 12 dismissals at 152 runs per 100 balls. Would RCB not have been better off getting those 973 runs spread over two or three batsmen at a strike rate of closer to 180-190? They had the wickets to spare.

It is a basic rule in cricket that attempting to score quickly results in more wickets falling. The charts below demonstrate this. The first is a plot of the scoring rate in an over in all T20 matches against the number of wickets lost in that over. As the scoring rate increases, so does the number of wickets.

© Kartikeya Date

The chart below shows how the scoring rate and the rate at which wickets fall develops during the course of a T20 innings. The spike from over one to over six is to do with the batting Powerplay.

© Kartikeya Date

Kohli is a batsman of phenomenal ability. It would not be a surprise if, as he approaches the age of 30, with T20 becoming an ever more prominent part of the calendar for elite international cricketers, he develops into a fearsome hitter in T20. But he's not there yet, and the professional observers (and fans) who think he's the best T20 player in the world are misleading the public about T20, and not helping Kohli's cause.

Another major talking point about the IPL is the idea that the best bowling side defeated the best batting side in the final, seeming to suggest that, contrary to observations by sceptics (like me), bowling is not irrelevant in T20.

Mustafizur Rahman can keep his wits about him. This is a remarkable gift, especially in the high-octane cauldron of T20. An economy rate of 6.9 runs per over through the tournament speaks volumes. His team-mate Bhuvneshwar Kumar managed an economy rate of 7.42. Bhuvneshwar (23) and Mustafizur (17) took 40 wickets between them. These are phenomenal numbers, given that the average scoring rate in the tournament was 8.28 runs per over.

However, compare the record of these two bowlers to that of Shane Watson and Yuzvendra Chahal of RCB, who took 41 wickets between them and conceded 8.58 and 8.18 runs per over respectively. Assuming that each pair gets through their full quota, Bhuvneshwar and Mustafizur would concede ten runs fewer in eight overs compared to Watson and Chahal.

The four-over quota means that a bowler who is one run per over better is worth only four runs to his side over a full T20 innings. The quota limits the value of a bowler like Mustafizur. If you look at bowling attacks on the whole over the course of the tournament, RCB were about 20 runs worse with the ball but 33 runs better with the bat.

Over the course of the season, Sunrisers beat RCB in two out of three games. All three were high-scoring games, well above the average for the league. This belies the idea that it was a win for the Sunrisers bowlers. After all, David Warner and his colleagues did put 208 runs on the board in the final. The margin of victory - eight runs - was a matter of a couple of mishits. More narrowly, the winning side conceded four runs from wides, while the losing side conceded 11 - the number of wides is pretty close; the number of runs from wides is a bigger difference.

Finally, there is the fact that batsmen are in control of about 75% of deliveries in T20. The individual players who produce higher control percentages (for example, Kohli) tend to use up a large number of deliveries (if a batsman defends off the middle of the bat, the batsman is in control, but nothing accrues to the score). So out of 24 balls available to each bowler, the batsman is in control about 18 times. Given that a top-edged boundary produces four or six, just like one off the middle of the bat does, those six mishits can amount to wickets, or on other days, a substantial number of runs.

TEAM RUNS FOR RUNS AGAINST DIFFERENCE
Gujarat Lions 161 169 -7.5
Royal Challengers Bangalore 193 175 18.6
Sunrisers Hyderabad 160 155 4.9
Kolkata Knight Riders 168 165 2.1
Mumbai Indians 161 164 -2.9
Delhi Daredevils 157 161 -3.1
Rising Pune Supergiants 165 165 0.3
Kings XI Punjab 158 171 -12.9

In the long run, it is arguable that these things even out, especially since the majority of IPL teams are more or less equally good (five out of eight teams this season won either seven or eight games each; the best team won nine), but on a given day they probably don't. When the total number of runs is so small (on average 166; see table above) and the units of scoring comparatively so large - 6% of the total can be produced off only two of the 120 available deliveries - the potential consequence of a mishit is commensurately larger.

So what should a team do? Should it pay top dollar for a good bowler or add hitting depth? The answer seems obvious. Given the over quota, and the marginal value of having deeper hitting, teams are better off looking for deeper hitting. RCB ought to shore up their hitting depth, not their bowling attack. Unless the IPL decides to get rid of the over quota per bowler, or relax it, the value of the high-quality individual bowler is limited. The same goes for the high-quality individual batsman. Kohli's runs could easily come from a couple of power hitters who will get out more often but also tee off quicker, and perhaps bowl a couple of overs from time to time.

The evidence and arguments presented here echo observations made by R Ashwin in a recent interview, where he said that T20 is not cricket but a different sport and must be understood as such. Ashwin predicted that a new par score will emerge in the near future, and said that teams that understand these points earlier will be more successful. Teams have an incentive to take evidence seriously; if they don't, they are likely to lose games to teams that do.

The pundits have no such incentive. It is one of the reasons why, 13 years into its life, no systematic account of the new sport that distinguishes its trade-offs from those of cricket exists. In part, this is probably due to reasons of marketing, which dictate that drawing a clear contrast with cricket and cricketing measures of merit is likely to alienate cricket's fan base, which T20 continues to draw on. In part this is also because pundits have no real incentive to be critical and examine evidence too closely. They cannot lose as long as the ratings remain high (the ratings are guaranteed by the Kohlis and de Villierses). In the process, they continue to discuss and describe a game that does not exist.

T20 is nothing like cricket. There isn't a "good bowling side" or a "good batting side". There is, arguably, hitting depth, which can be considered alongside the twin dimensions of power and reliability. Describing bowling requires far more than merely counting dot balls instead of maiden overs, or describing the length and line and the clusters thereof.

How about a measure of a bowler's tendency to drag the batsman out of control? If we learn, for example, that against Mustafizur, for a given asking rate, the average batsman is in control only 40% of the time, while against some other bowler this figure is 70%, that would tell us something useful. It doesn't really matter if Mustafizur bowls legcutters or offcutters, because batsmen are not defending the top of off stump, and Mustafizur is not trying to beat the edge of a vertical bat.

All this, however, must thankfully be set aside for the moment. The coming months bring with them Test match cricket, where the pundits are on surer ground. That is something to look forward to.

Kartikeya Date writes at A Cricketing View and tweets here

Comments