Coming into this season, Chennai Super Kings were the older team, picked for a form of T20 that no longer exists - Dhoniball. But they weren't just old in terms of average age: their batsmen start slow, which is not the way the game is headed.
The 2016-17 Adelaide Strikers had what looked on paper a fairy-tale middle order. Travis Head, Kieron Pollard and Brad Hodge were all there. Even with Hodge over 40 and Pollard not quite the GOAT he'd been earlier in his career, it was a formidable line-up. But it had one problem, all three batsmen started slow.
Players like those three are catch-up players: their overall strike rates aren't horribly low, they just start slow. Head strikes at 89 off his first five balls and 104 off his first ten - placing him firmly on the list of lethargic starters in T20 - but overall he strikes at 138. His jump from 104 to 138 is the fourth highest in all T20 cricket. Most slow starters don't catch up. The player with the third-quickest leap is Hodge.
*(min 200 runs scored off first five and first ten balls)
Of the 215 cricketers who have made 500 runs in top-level T20 since the start of 2015, 35 don't strike at 100 in their first ten balls. Only Hodge goes on to strike at more than 130 overall.
The problem isn't for the players themselves - they often catch up over their careers - it's for the teams, who need to make as many runs as they can. The most balls any players average facing in T20 is the openers, at 20; the No. 6 averages 11.6. If you are in a team of slow starters, and they all get in and out having faced the average number of balls they do, it creates a standstill.
Of course, not all batsmen start in the same part of the game. The average strike rate for the first five balls for batsmen in the middle overs is 91; in the Powerplay that is 99, and at the death it's 120. And among the top 15 quickest starters, there is a healthy mix of openers, death specialists and middle-order players.
Luke Ronchi is the world's fastest starter. He strikes at 161 from his first five balls (which is so quick, it would be the second fastest strike rate for the first ten balls as well). From the first ten he makes a fraction over 17 runs, almost a full run more than second-placed Carlos Brathwaite, and over two more than Chris Morris at three. Somehow Ronchi does all this with a batting average of 34.
Based on his overall strike rate of 150 since 2014, and the fact that he often comes into the match at the death, you'd expect Pollard to be one of the quicker starters, but he isn't. He's a chaser first and hitter second.
His method is far slower and that seems to be because he chases a lot, and two, it is perhaps inspired by Chris Gayle, who is in the slowest 30% of starters in T20, striking at 91 in his first five, and 113 of his first ten. No player has a higher strike rate differential (min 1000 runs) than Gayle; from his first ten balls to his overall strike rate it is a jump of 40. Pollard is a bit quicker than Gayle, at 106 from five and 118 from ten, but he's still in the lower half of T20 starters. He walked to the wicket for his most recent game, against RCB, at 148 for 4 from 15.3 overs, and still scored only five from his seven balls.
It is clearly part of his game plan, and as you will see, that of almost all of T20's best chasers, right towards the top of the list of the slowest starters, which means it is more design than accident. The players who have been not out in the most chases include Hodge, Eoin Morgan, Virat Kohli, Shoaib Malik, Gautam Gambhir, Ravi Bopara, and even Gayle. It makes sense - since Michael Bevan, limited-overs chasing has often been about getting yourself set and taking the game deep.
Since the start of T20, Pollard has been not out in the chase the most times worldwide. But the man who has been not out from a staggering 42% of all his innings (41 times) is MS Dhoni. The king of the chase.
Dhoni starts slow. Not slow for him, slow for everyone. After five balls he strikes at 76 (eighth slowest in all T20, and the slowest in IPL) and moves it up to only 98 after ten. Before 2015, he was up at 106 and 122 (placing him more in the middle of the pack). Since 2015, not even his overall strike rate is as high as 122.
* (min 200 runs scored off first ten and min 1000 runs scored overall)
But his team is similar. There is not a major batsman in the Super Kings line-up who scores faster than 110 off the first five balls, or 120 off the first ten. And it isn't like there aren't well respected T20 players in that list, or well-known hitters. They have Shane Watson, who hits a boundary every 5.3 balls, but he's at 90 and 111. Dwayne Bravo travels the world bowling slower balls and clipping sixes, yet he - the quickest CSK batsman - is only at 110 and 119. And Sam Billings, who is about to explode T20 cricket, is at 98 and 115. That is before you get to obvious slow starters like M Vijay (86 and 111), and the man who can block the ball for days, Faf du Plessis (90 and 104). As far as pace off the mark goes, it's a samey batting line-up.
Until now fast and slow starters were reputations, now they are hard facts, with ESPNcricinfo's smart stats getting off the ground. CSK's players might have different styles - the muscular Watson is nothing like the fluid Ambati Rayudu - but they all start at about the same pace. Teams already try to pick line-ups so they offer combinations that are left-right, good against pace/spin, and perhaps the next thing to look for is how they start.
In Super Kings' last game, Kings XI Punjab made 197, and CSK fell just short. Watson made 11 off nine, Vijay 12 off ten, Billings nine off eight, and from his first ten balls, Ravindra Jadeja was on 12. Those four players started their innings by facing 37 balls (nearly a third of the innings) and making 44 runs; that is, scoring at 7.1 runs an over. That left 83 balls in which CSK needed to score at 11.1 to win.
Dhoni played like his pre-2015 self and made 79 from 44. CSK lost by four runs. In the first game of the year, they were saved by a similar innings from Bravo. In their second game, it was Billings. Super Kings are relying on one player going massively as the rest of the team gets set wasting balls.
We don't know if this will hold them back, since it's a rather new metric: we can't say what effect so many slow starters will have. Perhaps giving all your players a chance to settle will work. But in a game of only 120 chances, slowing it down for 40 of them every game seems like an odd way to go.
This season the tortoise has twice defeated the rabbit, leaving Chennai 2-1, meaning even if they don't start quickly with the bat, they are at least doing so on the points table.
With stats inputs from Shiva Jayaraman