Andre Russell's left foot pointing to midwicket, his right knee bent, and his arms swinging through a ball that's about to be called a wide is not what you think of as a traditional cricket shot. Test batting has actually slowed down in recent years, meaning T20 cricket has never been less like it. The runs-per-over figure in the last three years for Tests is 3.2, for ODIs 5.3, and in all T20s it is 8.2.

T20 has lit a fire under attacking batting, and for someone who believes Test batting is proper cricket, it's hard to look at T20 hitting and not think it's crazy. We've not only perfected the cow-corner hoick but the inside-out hockey-slap, the back-of-point slice, and a host of other shots where someone will lose their Adam's apple. Players now understand how to hit the ball hard.

But the actual gains in run rates across formats are pretty low. Players aren't adding an extra run a year in T20, and ODI cricket hasn't become as much like T20 as some would think. The highest run rate in ODIs over a year was in 2015 - 5.5 per over; and that included extras. Viv Richards scored at 5.4. In ODIs last year the economy rate was 5.3 (mocking the whole "300 is par" notion), though it had risen from 5.12 in 2009. Even if you look at the lowest yearly run rate in the last decade, 5.05 in 2012, against the highest, in 2015, it's only an 8% rise.

The run rate in T20 has grown slightly more quickly. In the last decade, the lowest run rate was 7.48, in 2013; last year it was 8.4, which is an 11% rise. Unlike a good T20 innings, it has been steady rather than spectacular.

And that is because T20 cricket is still anchored in what we might call "normal cricket". In 2018 the T20 batting average for all batsmen was, at 25.06, the highest it has ever been, suggesting that batsmen are putting more value on their wickets now than in previous years. They might hit a lot of sixes, and they might get quicker with their scoring, but they still play T20 like normal cricket, just with more urgency. It has not truly become a sport in its own right just yet, although it is well on its way.

In computing and science, people talk about theoretical limits. In nature, the theoretical limit to how fast something can travel appears to be the speed of light (about 300,000km/s). The fastest observed human is Usain Bolt, clocked at 44.72kph. So there appear to be natural limits depending on mass, physiology and other factors. For T20 cricket we have a theoretical limit of how fast batsmen could score if they did not worry about losing wickets - the scoring rate off free hits.

Since 2008, the scoring rate off free hits stands at 12.54 (runs per over); so batsmen can't score at much faster than two runs a ball. As T20 currently stands, 12.5 is our theoretical limit, which translates to an innings score of 250.

At the moment the rise of the run rate off normal deliveries is fairly in sync with the run rate off free hits. A couple of years one has increased while the other has fallen - or vice versa - but overall they are rising fairly equally. (The free-hit run rate has swung more but that is mainly because there are so few free hits every year, amounting to a fraction of a percentage point of all runs scored.)

Of course, batsmen cannot score at 12.5 for sustained periods because of the finite resource of ten wickets. Additionally, as so often happens in limited-overs cricket, there will be playing-condition changes that make that number redundant. So as long as five bowlers are needed, four fielders need to be inside the circle, there are six overs of Powerplay, and turf wickets continue to be used, it's hard to see players crossing the 12.5 mark. If they keep improving at the current rate - in the last decade the scoring rate in T20 has gone from 7.57 to 8.4 - it will take them over 30 years to make it.

But batsmen are still playing a game that is rooted in conventional cricket-think. If and when that changes, run rates could take a jump.

***

When I speak to T20 players and coaches, many of them point to batting averages or overall runs tallies. Batsmen still don't want to get out, and coaches place a high value on consistency over explosiveness. T20 has changed that equation slightly, but it has not eradicated that thinking.

In 2018 the average of all batsmen in T20 was the highest since the birth of the format. And this year is currently the second highest. Batsmen are batting longer, meaning more players get set - so if you think about it from a conventional standpoint, it would make sense that they score quicker than newer batsmen.

T20 batting line-ups still rely on the top order to a ridiculous degree: 496 players opened the batting last year in T20 in 717 matches (that is for both spots). At Nos. 5 and 6, over 500 players batted in each. Which tells you that we chop and change other batting positions because we struggle to work out what success is in those positions, while for openers a steady-as-you-go start is enough to get you multiple opportunities in that position.

And that is because openers seem at first glance to be the most successful batsmen in T20. So they should be, given they have the field up for almost all of their innings, and even with mediocre years, can find themselves towards the top of the run tally or average list. Opening is the quickest-scoring batting position in T20, so even slower-scoring openers often have better strike rates than Nos. 4 and 5.

If you throw in the No. 3, who gets some of the benefits that apply to openers, the first three batsmen face 57.6 balls an innings, or 48%. That does not leave the rest much of a chance to get started. Openers could go far quicker, but teams still fear the "lose three wickets in the Powerplay and lose 65% of the match" rule.

It's something teams still hold on to. A team like Adelaide Strikers seem to have taken this so much to heart that they don't attack much in the Powerplay at all; instead they start attacking once the field goes out. This approach is steeped in traditional cricket thinking. Many of the losses that that 65% rule refers to are because teams slow down after they lose the three wickets and consolidate, and not just because they lost three early wickets. In part that is because they have stacked their batting at the top, but mostly because that is how batting always has been in cricket: your top order makes most of the runs, and if they don't, you have to rebuild to set a new base.

But what if the base is wrong to begin with?

What if T20 batting isn't about big contributions from the top but a series of contributions all the way through? In 2018, No. 8s on average faced 4.07 balls per match, and only appeared in 58% of possible innings. So that means they hardly get in, and when they do, it's not for long.

The top six currently face 100 balls, or 83%. And often they do that by conserving their wickets and then trying to attack towards the end.

To score a lot quicker, you could select a team of players who average 20 at 1.6 runs a ball; a top eight, perhaps. It takes 12.5 balls to make 20 at that speed, and if you had eight players who had the ability or temperament to do it, you would have 160 runs from 100 balls with 20 balls left in the innings. There would be two extra batsmen to consume the 100 balls, but they would be scoring what teams score in a match now, with 20 balls left. That comes down to ability, but you'd also need all-round skills, to not weaken the bowling. It so happens that a lot of the players who smash it loads do have bowling ability, like Russell, Sunil Narine, Colin de Grandhomme and Corey Anderson.

Some of this goes back to traffic management. If you build a highway, the worst thing that can happen is, the highway is seen as the only option. You want to build a highway and still have a large proportion of people use other roads. Otherwise everyone will clog up the highway. If you rely too heavily on your top three and they bat slow for too long, they clog up the innings. If you have a multitude of scoring options, and everyone is trying to score as quickly as they can, the chances of traffic jams are limited.

If your top three face 60 balls every game, unless they are incredibly fast scorers, you're limiting what your other talent can do and allowing the opposition to stack their bowling in match-ups. If you have a collection of players who all can face 12-15 balls per game and go for it, everyone is a threat, and you'll have a team with many diverse strengths and weaknesses.

Even the king of T20 batting, Chris Gayle, gets himself set to ensure he scores big more often, knowing that his power can catch him up and then some. Gayle is one of 26 batsmen who have faced ten or more free-hit balls, and he is the only one with a strike rate of above 300 from them. So if even Gayle is getting set, getting a good average and ensuring his run total is high, what of mortal men? The average of all openers last year was 27 and they struck at 134. Not that long ago, batsmen who averaged 30 and had a strike rate of 130 were in demand, but openers have got so good in that position that they do score 30 at 130 as default.

This system, of players scoring faster and doing so deep into the batting order, might feel riskier at first, but it could lead to more consistent high scoring.

Now while these big data trends often tell us a lot, they can be tricky. In the last three years, among all grounds used more than ten times, Trent Bridge has the highest run rate, 8.82 (which is a strike rate of 147). If you are playing at Colombo Cricket Club Ground or Multan Cricket Stadium, the runs per over are about 6.84. So a strike rate of 130 at Trent Bridge is not sensational, but in Colombo or Multan, it's great.

Factors like this, not to mention collapses, weather, bowling match-ups, all affect what batsmen do. Also, there are teams like Sunrisers Hyderabad, Perth Scorchers and Chennai Super Kings, who have preferred to achieve consistent totals rather than score as many as they can, and that has helped them win titles.

Even as T20 hitting has improved, cricket still finds a way, through bowlers or conditions, to make people go back to what they know.

***

There are ways - not exactly new ones - of extending your batting line-up and making each batsman more efficient.

The pinch-hitter has been around in cricket for a long time. It was perhaps 1992 when the idea took a strong hold in ODIs, and pinch-hitters have been used occasionally since then. Narine has brought the trend back in T20 cricket. As an opener he averages 18 and strikes at 160. You're not losing that much if you throw someone like him up: he won't face many, if any, balls batting at Nos. 9 through 11, and he's less likely to play for his average.

There are plenty of players like Liam Plunkett, Adil Rashid, David Willey and Rashid Khan who have the game to be thrown up the order. If they will only face five or fewer balls anyway, targeting the Powerplay with a relatively unimportant wicket might work. You could also make the argument that batting at the death is harder, with the field out, a soft ball, and death-bowling specialists operating. You might want smarter batsmen there, whereas at the top, hitting is at its most simple. A score of ten off four balls at the top is worth more than ten off four at the death, as death over run rates are higher.

But it is in the middle order that things get trickier.

***

Players who find themselves in the middle order are usually professional batsmen. They are trained a certain way. There are coaches who still get upset when a player hits a boundary and tries to hit another the next ball. Batsmen in the top six want to bat long and sensible, even if the definition of that has changed for T20.

One of the most notable exceptions is Russell. He faces far fewer balls than a normal middle-order player - only 13 - and can bat anywhere from three to seven. Russell in the last two years, since he came back from his drugs ban, has averaged 31, but at a strike rate of 182.

Russell has been slowly improving over the years, averaging more. His enforced break from the game seems to have helped his batting. But unlike most middle-order players, his batting is not the sole reason he is in the team. He could afford to risk it all and only face a few balls a game because - like Narine - his bowling is strong. Now if his knees keep getting worse, he has more than enough of a reputation as a batsman to continue on the strength of that skill alone - but the ability to bat the way he does came from the original freedom his bowling allowed.

Though there are few batsmen with Russell's skills, he is not alone. Players like Anderson, Ben Cutting and de Grandhomme can score at inhuman levels. One major difference is that often these guys either aren't bowling at all or they are the fifth or even sixth bowlers. Russell is a genuine front-line bowler. He can start batting in sixth gear without fear; most players cannot, either physically or mentally.

Now imagine you had the ability to bat like Russell or Narine, but no second skill at all. There aren't many teams who take punts on players who average around 20. Most coaches don't see it as a passable average, no matter what the strike rate. Unless you are batting at No. 7 and giving something else in the field. So if you had any batting ability, pushing your average up while limiting your strike rate would be a fairly safe bet.

These kinds of players need to be unlocked in order for batting to be spread more throughout the innings and for players to cut loose with fewer fears.

***

You can't fault the players for taking the position that will help their career. But free hits have shown what is possible, and so has the T10 league. In the last T10, 21 players scored over 100 runs apiece, of whom 15 scored at over two runs a ball, and not one scored at slower than 1.5 per ball. What was interesting was the jump in strike rates for some players between T20 and T10.

Andre Fletcher is the second highest scorer in CPL history. He can smoke fast bowling, including doing things like score 20 runs in four balls against one of T20's most parsimonious bowlers, Mohammad Irfan. But after the Powerplay his strike rate often slows down 30 points or more. In the T10, without a natural slow-down period, and without the worry of having to bat for a long time, he struck at 214. In the last two years his T20 strike rate is 122; that's some leap between the two.

And it wasn't just him. There are so many players, like Nicholas Pooran, Najibullah Zadran and Fletcher, who have so much more to give. So while there aren't many Russells out there, the T10 and free hits have shown us that many batsmen can score at scary levels.

There were only four players of the 21 who made over 100 runs in the last T10 series whose strike rates didn't rise by 40 or more: Alex Hales - whose strike rate in the last two years is 146 outside of Trent Bridge and 209 at his batting-friendly home; Cameron Delport, who struggled to get going in the T10 tournament; Hazratullah Zazai, who has been striking at 189 recently, so it's hard to improve on that too much unless you are Russell (who hit three fours and 17 sixes in the T10); and Narine, who has been at 153 in T20, and went up to 187.

So we can assume that Narine is batting near his natural scoring limit, and it seems like most players are not. It's not a surprise to learn that despite getting into pretty much any side as first or second bowler, he has also spent years working on his striking - at first, hitting sixes over fours as a tail-end cameo man; and then, after Aaron Finch promoted him to try and upset frugal left-arm spinner Michael Beer, he became a semi-frequent short-innings opener.

But this is a low sample size, and for all we know, the wickets were made for batting. Also, I talked to one bowler in the T10 who admitted it was hard to get up for only bowling 12 balls a day knowing you were about to be destroyed. However, like the free hits, it shows what is possible when you embrace the hitting and stop worrying about wickets.

Russell was an aggressive player from the start, and with two first-class hundreds in 17 games, he is a decent batting talent to begin with. But what he did was use all those elements to become the biggest force he could be with the bat. And still, his strike rate jumped 89 in the T10.

It is not simple for a team owner to buy all the Andre Russells in the world, because there's only one. And even cheaper versions, like de Grandhomme, Anderson, Willey and Chris Morris, are still quite pricey.

But what free hits, T10, Narine and Russell have shown is that T20 can get a lot quicker. It's just a question of when it will happen.

If you are a specialist batsman, it's not that easy to make a call when it is your livelihood on the line. T20 tournaments are short. If you tried going all out for one series and ended with an average of 16, would it matter that your strike rate was 180? While individual players will try it, how long before batsmen as a whole go for this approach? It might even be dictated by the people who choose teams and not the batsmen themselves.

There is definitely a limit to T20 scoring at the moment, and considering that in no year have we topped the nine-runs-an-over mark, that seems to be it. Not long ago, it was eight runs an over. These limits are not natural, though; they're man-made. As more specialists come into T20, players try new methods and continue to work out what is actually possible, the sky may not be the limit, but 12.5 will not stay as the theoretical one.

Jarrod Kimber is a writer for ESPNcricinfo. @ajarrodkimber