The 2010s, the first full decade in the history of T20 cricket, have been, arguably, what the 1980s were to ODI cricket. The biggest shifts in ODI cricket since the 1980s have involved the opening position. ODI cricket has gone from the staid era of Sunil Gavaskar, Desmond Haynes, Gordon Greenidge and Geoff Marsh, to the best-bat-at-the-top era of Sachin Tendulkar, Brian Lara, Richie Richardson, and Mark Waugh, to the pinch-hitting era of Saeed Anwar, Sanath Jayasuriya, Adam Gilchrist and Virender Sehwag before settling in the contemporary era of David Warner, Rohit Sharma, Quinton de Kock, Jason Roy and Jonny Bairstow, where openers are expected to have power-hitting capability. Where does contemporary T20 cricket stand when compared to the historical trajectory of ODI cricket? What role does the T20 opener play in this story?

In the 1980s, the average first-innings score in an ODI was 223 and the team batting first was bowled out in 20% of innings. In the 1990s, the average score rose to 233 and teams were bowled out in 24% of innings. In the 2010s, these figures rose further to 267 and 33%. Even if some of this growth is caused by changes to the playing conditions, the figures suggest that batting sides in ODI cricket have learnt to spend their wickets more aggressively as time has gone on.

Teams have ten wickets to spend over 300 balls in a 50-over match, and successful specialist batsmen typically tend to be dismissed less frequently than once every 30 balls. Currently, 18 batsmen have made more than 1000 career ODI runs at a rate better than a run a ball. Nine of these players are dismissed more frequently than once every 30 balls.

In ODI cricket now, there are players who approach batting in such a way that if a team was made up of them exclusively, it would be bowled out in most innings. Consider the table below, which tells you how often each team has been bowled out after batting first in an ODI since the 2015 World Cup. England have been the best team in the world in this period, and their record demonstrates a greater willingness to risk dismissal in the (successful) pursuit of more runs.

When compared to the ODI cricket of the 2010s, the T20 game is extremely conservative. As of September 10, 2020, of the 603 T20 players who have scored at least 1000 career runs, only three of the 50 fastest run scorers are dismissed more frequently than once every 12 balls. The fastest - Andre Russell - is dismissed once every 16 balls. In ODI terms, that's equivalent to being dismissed once every 40 balls. A player like Glenn Maxwell, who is considered cavalier and inconsistent, is dismissed once every 17.7 balls. The average T20 first innings ends in a team being bowled out somewhere between 9 and 13% of the time.

The tone for the conservatism in T20 is set at the top of the innings. In the 1980s - an era in which fast bowling is commonly considered to have been superb - the average opening stand in the first innings of an ODI made 59 off 101 balls. In the 1990s this advanced to 64 (95), in the 2000s to 63 (83), and in the 2010s to 70 (85).

In the early years of T20, before the franchise leagues came along, the average T20 opening pair scored 47 in 37 balls in the first innings. In the first half of the franchise-league era (2008-14), this became 49 in 40 balls. Since 2014, the average has advanced to 51 in 40 balls.

The key figure here is 40 balls. T20 openers use up, on average, a third of their team's innings. This is congruous to the ODI game in the 1980s, when openers used up a third of a 50-over innings. If contemporary T20 were as top-heavy as contemporary ODI cricket, where openers use up 85 balls, the T20 openers would use up 34 balls on average.

It can be persuasively argued that, given T20 cricket is two and a half times shorter than ODI cricket, the fact that T20 openers use up a third of a T20 innings makes the approach in T20 cricket significantly more conservative than even the approach in ODI cricket in the 1980s. In a contemporary ODI, after the openers are done, the rest of the batting has to navigate 215 balls. In a contemporary T20, after the openers are done, the rest of the batting has to navigate 80 balls.

The distribution of deliveries available by batting position suggests that T20 tends to be played more or less as a scaled-down ODI. This miniaturisation of the system has to break down somewhere, and it breaks down by reducing the chances of the batting team being bowled out from 36% in ODIs to about 15% in T20s (considering both innings; the figures for the first innings alone are 31% and 11%). Chasing teams are bowled out more often than teams batting first, as one would expect.

As the table below shows, the seventh wicket falls in a T20 innings roughly as often as the ninth wicket in an ODI in the 2010s, and as often as the eighth wicket in an ODI in the 1980s.

Openers use up a disproportionately large number of deliveries. More generally, batsmen batting in the top four in T20s face, on average, 71 out of the 120 deliveries.

Consider the 95 T20 batsmen who have at least 2000 career T20 runs while opening as a proxy for what is considered successful T20 opening, from Chris Gayle to Sharjeel Khan. The chart below gives their average scores (runs - balls faced). The average score among the top 95 T20 openers is 28 (off 21 balls) for a scoring rate of 133.3. Of the 95, 50 score slower than this rate, 45 score quicker. Among those who score slower than 133.3, a clear majority, 31, play more than 21 balls per innings, while 19 face at most 21 balls on average. Among those who score quicker, a clear majority, 31 again, play at most 21 balls, while 14 play more than 21 balls per innings on average.

We can set aside the 50 openers who score slower than 133.3. The most significant thing about this set is how large it is. They constitute a narrow majority (50 of 95) of the top T20 run getters and account for almost exactly half the total number of caps won by these 95 players. It will not be particularly controversial to observe that among the openers who score slowly, the ones who are dismissed earlier on average are less damaging to their sides than the ones who stay in longer.

The 45 openers who score quicker than average are a more interesting bunch. Going by the analogy with ODI cricket above, the 31 openers who score quicker and are dismissed after 21 balls or fewer on average (let's call them the cavaliers) are better than the ones who score quicker and stay in longer (let's call them the anchors). It could be argued that if the anchors stayed in longer and scored quicker, this would be advantageous, but the opportunity cost of not being able to use a pure hitter against those extra deliveries has to be considered as well. The cavaliers score at 142.5 runs per 100 balls faced, while the anchors score at 141.4 runs per 100 balls faced. The anchors are some of the biggest names in T20, and they are also generally better batsmen compared to the cavaliers.

After 21 balls, 11 out of the 14 anchors above are already ahead of the 28-run average (for T20 openers who have made at least 2000 career runs). For example, Chris Gayle's average score after the 21st ball is 29.9, while Jos Buttler's average score is 33.7. Further, the earliest delivery on which Buttler's scoring rate exceeds 133.3 runs per hundred balls faced is the seventh delivery of his innings. For Gayle it is his 16th delivery. Finally, Gayle is dismissed in 46% of his innings before his 16th delivery, while Buttler is dismissed in 15% of his innings before his seventh delivery. The table below adds texture to the idea of the anchor. Not all anchors are created equal.

Virat Kohli is an extremely conservative anchor when he opens the batting. It takes him until his 38th delivery for his average scoring rate to rise above 133.3. He is dismissed before his 38th delivery in three out of four innings as opener.

The point of this table is not to describe who is better at hitting than whom, but to describe each player's approach. That Kohli can hit when he wants to is clear from the fact that he scores, on average, 30 runs from his 36th to his 49th deliveries, compared to Gayle's 27 and Buttler's 18. While Kohli is conservative as an opener compared to other openers, his record as opener looks aggressive when compared to his record at No. 3 (where he has played a majority of his T20 innings).

It is worth reflecting on just how large the figure of 21 balls is in a 120-ball innings. The 50-over equivalent of a 21 ball T20 innings is 52 balls. Consider that Sachin Tendulkar's average ODI innings lasted 47.3 balls.

Today's T20 openers are, in some respects, even more conservative than the ODI openers of the 1980s. Desmond Haynes scored at 64 runs per 100 balls faced; Gordon Greenidge 65, the cavalier Krishnamachari Srikkanth 72, Geoff Marsh 57, John Wright 57, Graham Gooch 65, Sunil Gavaskar 62, Mudassar Nazar 52, David Boon 68, and Ramiz Raja 67. Chris Gayle's T20 record, which looks so fantastical now, may one day look just as dated as Haynes' ODI record looks today.

Compare the batting contributions of openers to Nos. 5, 6 and 7 in the order over our three categories of matches: openers face a third of overs in both T20s and early ODIs. In contemporary T20, the openers and Nos. 5, 6 and 7 bats score at about the same pace. In the early 1980s, when the role of the opener was to see off the new ball, openers scored at about 80% of the speed of those at five, six and seven. In contemporary ODIs, the openers and Nos. 5, 6 and 7 face roughly the same number of deliveries (85 for openers, 80 for Nos. 5-7) in the average innings.

T20 openers score quicker after the sixth over (137 runs per 100 balls faced) than they do during the powerplay (120 runs per 100 balls faced). This suggests that much of the ballast of the opener's high strike rate in contemporary T20 (when compared to those of Nos. 5, 6 and 7) comes towards the end of their longer innings. The scoring distributions for the 14 anchors above illustrate this. In essence, openers in T20 see off the new ball and attack later in the innings if they survive, while openers in 1980s ODIs performed the role of seeing off the new ball, but rarely survived long enough to accelerate.

The difference between David Boon and David Warner is probably not ability but approach. We have seen the effects of shifts in approaches in ODI cricket from time to time. Most recently Eoin Morgan's England decided to adopt the approach used by Brendon McCullum's New Zealand to devastating effect. Today, England's openers use up 13 overs out of 50 on average. India, with a more top-heavy batting line up, more wicket-taking bowlers, and fewer power-hitting allrounders, take a different approach, and their openers use up, on average, 18 overs out of 50.

There is scope for such variations in approach in T20. It is possible to foresee a future where the average opener who is in demand on the global T20 circuit will average 15 balls per innings instead of 21, and contribute somewhere between 25 and 30 runs. The average T20 team score will probably rise to about 180, and T20 sides will get bowled out more often.

A change in approach requires a different understanding of risk. Conventional cricket wisdom today still regards anchoring as the safer option, when in fact, given the facts of T20, it is significantly riskier. Think of it this way. If an anchor pushes singles for 15-20 balls, then those 15-20 balls are conceded to the opposition without contest (since overall scoring rates are greater than one run per ball). Instead, if two or three batsmen try to hit boundaries over those 15-20 balls, there's some chance that some boundaries will be scored.

So Kohli may well approach his innings the way he does because he feels he must carry the burden of making a big score. But under the upside-down logic of the 120-ball contest, this is a self-defeating proposition. The additional work which the batsmen who follow have to do in the relatively fewer deliveries available to them to make up the scoring deficit caused by those 15-20 balls only compounds the problem.