Who are the greatest T20 players of them all?

Among bowlers who have bowled at least 2000 T20 deliveries, Sunil Narine has the best economy rate of all, 22.7% above average BCCI

The 11 greatest performers over the first 15 or so years of T20 are Chris Gayle, David Warner, AB de Villiers, Kevin Pietersen, Shahid Afridi, Kieron Pollard, Shane Watson, Sunil Narine, Lasith Malinga, Saeed Ajmal and Dale Steyn. Statistically, Andre Russell would be a certainty in any such list but I've set him aside due to his current ban for breaking anti-doping whereabouts regulations.

This list has been built using two quantitative measures as guides (these measures have not determined the composition of the list). It has also been compiled with studious disregard for the popularity of individual players or for individual performances in high-profile games. There is no logical or cricketing merit to assigning arbitrarily higher significance to specific games. All competitive games are precisely that - competitive games. Players want to win all the games they play in.

The general profile of T20 form is as shown in the chart above, which displays the average wicket rate and scoring rate per over for each over in each innings across all games. Each player's over-wise performance can be summarised from ball-by-ball data.

For example, at the time of writing, based on available data, Gayle had scored 386 runs from 278 balls in the fourth overs of the first innings of T20 games. He was dismissed 16 times in these 278 deliveries. He scored 8.33 runs per over in the fourth over of the first innings, and had a dismissal rate of 0.34 wickets per over. The average scoring rate in the fourth over of the first innings is 7.08 runs per over, while the average dismissal rate is 0.25 wickets per over. From this, it follows that for the fourth over of the first innings, Gayle scored 17.7% faster than average, and was dismissed 36.9% more often than average. He faced 4.4% of his total deliveries in the fourth over of the first innings.

A career summary for Gayle is calculated as the weighted average of this relative performance per over. The weight assigned to each over is the share of the total balls faced by the player in that over. This career summary in the case of a batsman is what I call the misbah measure. It is a two-dimensional measure consisting of the scoring rate and the dismissal rate relative to the average T20 performer. An identical measure is calculated for bowlers. This is the jogi measure. For example, Narine has conceded 215 runs in 291 balls in the fifth overs of the first innings of his T20 games, and taken ten wickets. This gives him an economy rate of 4.43 and a dismissal rate of 0.21 wickets per over. Compared to the average bowler in the fifth over of the 1st innings, Narine's economy rate is 38.2% better, while his dismissal rate is 19.8% worse.

These measures are named after, and in tribute to, Misbah-ul-Haq and Joginder Sharma, the two main protagonists in arguably the single most significant over in the history of T20. It was the last over of the inaugural World T20 tournament. India were defending 13 runs, Pakistan had one wicket in hand. Joginder bowled four medium-paced deliveries. One was a wide, another a full toss that went for six. There was a good-length ball well outside off stump, which Misbah missed because he backed away to leg before it was bowled. And finally, there was Misbah's unsuccessful attempt to paddle one over short-fine leg. India won, and within six months, the BCCI, which only a year earlier had declared at the ICC that "India will never play T20", had set up the Indian Premier League.

From the observer's point of view, it is impossible to determine cricketing merit for individual innings or bowling performances in T20. In a Test match, it is possible to look at a field setting and a persistent line of attack, against which the batsman is forced to play to certain parts of the field. It is up to the batsman to force his away out of this trap. If the bowling is persistent, the batsman is forced to try things he would normally not prefer to attempt. If the batsman executes this perfectly and forces the bowling side, in turn, to try something different, then some coherent discussion of the contest between bat and ball is possible, irrespective of the outcome of individual deliveries.

In T20, the abbreviated format makes such conclusions impossible for individual contests. Batsmen premeditate frequently and fields are set almost exclusively to minimise the damage. But over a large number of games, if one player ends up performing better than average, while another ends up performing worse than average, then it is possible to say with some certainty that the former must be doing something that is systematically superior to the latter's play. It is possible to conclude that a player has an effective range of cultivated attacking shots, or that a bowler has worked out some way to keep players quiet. Gayle fails quite often, and Narine occasionally goes for 40 runs, but this has happened systematically less often for Gayle and Narine than it has for other players.

The chart below shows the misbah measure for all 84 T20 players who have faced at least 2000 balls in their career. The players who are located below the diagonal in the top right quadrant have better than average scoring rates and dismissal rates (i.e. they score faster than average and are dismissed less frequently than average), and the extent to which they exceed the average scoring rate is more than the extent to which they exceed the average dismissal rate. However, as readers will notice, there are players above the diagonal whose scoring rates exceed those of a number of players below the diagonal.

In all, 47 out of 84 players who have faced at least 2000 deliveries are located in the top right quadrant. If the qualification is relaxed to 1000 deliveries, 85 out of 227 players are in the top right quadrant. If it is further relaxed to 500 deliveries, 131 out of 464 players are located in the top right quadrant. A group of batsmen in the bottom-right quadrant show scoring rates that are least 25% above average and dismissal rates that are 20% percent better than average. These are Virender Sehwag, Glenn Maxwell, Jason Roy, Paul Stirling, Richard Levi and Shahid Afridi.

The chart above shows all 68 bowlers who have bowled at least 2000 deliveries. Twenty of these are in the top right quadrant. Note that in the case of the bowlers, the dismissal rates and economy rates lie generally closer to the average. The bowler with the best combination of economy rate and dismissal rate is Ajmal. Narine has the best economy rate of all, 22.7% above average. His dismissal rate is marginally below average.

The misbah and jogi measures can be worked out for different parts of T20 innings as well. Three groups are shown here - the Powerplay overs (1-6), the middle overs (7-15) and the end overs (16-20). For the bowlers, the qualification in each case is a minimum of 1000 balls bowled. For the batsmen, the qualification for the Powerplay and middle overs is 1000 balls faced, while for the end overs it is 700 balls faced. The players with the ten best run/economy rates are listed in each case, regardless of their propensity to take wickets or be dismissed.

Before concluding this article, here is a comparison between the misbah measures for Virat Kohli and de Villiers. As the comparative chart below shows, Kohli is significantly more conservative than de Villiers in the Powerplay overs. Kohli's record suggests that he looks to keep his wicket intact until the end overs, when he cuts loose and his scoring rate soars. His dismissal rate is about average. He and de Villiers make an excellent combination in the end overs. De Villiers is far more aggressive overall. A peculiarity in his record is the vast disparity between his performance in the first and second innings. He is outstanding in the first innings, and very close to average in the second. Kohli is conservative in both innings but marginally better in run chases. It should be kept in mind that while 31% of Kohli's deliveries are in the Powerplay overs, only 21% of de Villiers' deliveries come in that period.

In conclusion, here is some justification for my choice of the 11 greatest performers over the first 15-odd years of T20. They are, again: Chris Gayle, David Warner, AB de Villiers, Kevin Pietersen, Shahid Afridi, Kieron Pollard, Shane Watson, Sunil Narine, Lasith Malinga, Saeed Ajmal and Dale Steyn. My preference has been for players who have had longer careers. Mustafizur Rahman, Quinton de Kock, Mitchell Starc and a few other players may well match or surpass the ones listed here and command a place in the 2025 edition of this list. I have also preferred players who are involved in all phases of a T20 game as far as possible. Samuel Badree has a phenomenal record in the Powerplay and middle overs, but he rarely bowls in the end overs. Finally, I have preferred players who play in many different leagues over players who have built phenomenal records in one league. This leads me, for example, to prefer Pietersen over Suresh Raina. The choice between Afridi and Maxwell was a close one. Maxwell has a slight edge on the batting side, but Afridi has a clearer edge on the bowling side. By the time Maxwell finishes he career, he may well surpass Afridi. Finally, I have preferred the actual record over potential or quality or similar perception-based factors. On this basis, Gayle and Narine are the two outstanding T20 players over the period in consideration.

Data correct until May 17, 2017. It includes all matches (about 80% of total matches) for which ball-by-ball data is available