David Willey is a pinch hitter. I think. It is almost impossible to tell anymore. His batting innings map - a breakdown of what positions in the order he has batted in - is all over the place. He batted between Nos. 7 and 9 in the first five innings of his career, but he was up at No. 3 in his sixth. Two innings later, he was batting at 10.

I know much of that sounds like numberwang, but Willey's batting map really is all over the place. Three years later, he had opened, been a tailender, batted one down, and then opened again. It's a mess. Willey has a different role every time he turns up for a different team. What is different is Willey's range: he's been in the top three over 50% of the time, but also nearly 35% of the time No. 7 or lower - generally you might expect such a spread to be 50% in the top order and 45% from Nos. 4 to 7.

Willey either bats at the top or at the bottom. He's rarely batting at Nos. 4, 5 or 6. He started doing this for Northamptonshire. Before Willey, Northants had a very similar approach with Graeme Swann, who is now commentating on Willey at this IPL.

Weirdly though it is at two wickets down that he has entered for Royal Challengers Bangalore in his last two games, as a pinch hitter, but a similar role to what he has done in the past: a powerplay maximiser. It didn't go well for Willey (even if Royal Challengers won), as he made 18 from 28 balls.

Batting up the order has been the way to use Willey through his career because he struggles to score at the same speeds with the field out. He's at his best when he's at the top - he averages 27 and hits at 145 in the top three - but he's not always at the top because, despite that record, he's clearly a limited batter. The kind who can end up with 18 off 28. And if he wasn't limited, he wouldn't be a pinch hitter, he'd be a top-order player.

And so, despite the fact that you think we'd see more pinch hitters these days, we see fewer pinch hitters now.

If you look at the fastest run-scorers in ODI cricket in the 90s (from the top five in the batting order), you won't find many of the best players of that era. You will find a lot of bowlers, wicketkeepers and a few allrounders.

That was the peak pinch-hitter era. ODI sides threw all sorts of players into their top order. Australia were all over this craze, using Peter Taylor, Brendon Julian, Simon O'Donnell, Craig McDermott, Ian Healy, Shane Warne and even noted slowpoke Jason Gillespie as a pinch hitter. That is some dedication to the cause. They tried literally everyone in this weird floating position that was very much the rage.

But it wasn't just an Australian thing. The list of random players tried as a pinch hitter in the 90s is quite something. Nicky Boje, Rashid Latif, Ravi Ratnayeke, Mark Ealham, Chaminda Vaas, Dominic Cork, Anil Kumble, Phil DeFreitas, Pat Symcox, Javagal Srinath and Junior Murray; Wasim Akram averaged 14.5 and struck at 136 (batting in the top five in the 90s). He was basically a T20 player before we knew it was a thing.

There are many reasons the list of players is the way it is. One is that back then batters weren't as big as they are now. You may have seen the photo of Barry Richards holding a bat from the old days alongside the modern willow. You could do the same kind of picture pitting Corey Anderson's biceps next to Don Bradman's. Batters of the 90s hadn't optimised their hitting prowess. They hadn't grown into the massive beasts that baseball's designated hitters had.

So in the 90s, fast bowlers were usually the stronger players and many had the reputation of being the biggest hitters, which is why they went in. Mind you, so did tiny spinners and random wicketkeepers so the idea was quite clearly just to throw someone up the order to cause chaos. And almost all of them failed to do it.

They just didn't make many runs. O'Donnell played one of the great early innings in ODI cricket, scoring 74 from 29 balls in Sharjah against Sri Lanka after coming in at No. 4. Sharma made a better-than-a-run-a-ball hundred against England at No. 4 in the Nehru Cup in 1989. But more often, someone would come in, swing hard, and miss a bit.

As a general tactic the bowlers and part-timers didn't work at all. Shahid Afridi (averaging 25 at a 105 strike rate in the top five in that decade) and Lance Klusener (averaging 36 with a strike rate of 85.5) were exceptions but moving specialist batters up the order changed ODI cricket completely. At the 1992 World Cup, Mark Greatbatch and Ian Botham were promoted and both their sides prospered. But it was really the following World Cup where it all changed.

Sachin Tendulkar and Mark Waugh had moved up to open by that point. Still, it really was the pairing of Sanath Jayasuriya and Romesh Kaluwitharana that changed things. The two of them were so interesting. Jayasauriya was a proper batter, who was unleashed against the new ball with the field up and he went on to be a legend. Suddenly, everyone wanted a dazzling middle-order player to start off their innings. But it was Kaluwitharana, the more limited lower-order pinch hitter, whose kind all but disappeared.

Pinch hitter is a baseball term, a player who comes on for someone substituted out of the line-up. We don't have substitutions in cricket, so our version is to move someone up the order who doesn't belong that high. And then they're given a licence.

A pinch hitter could be a middle-order player used at the top. It could be an allrounder promoted a few spots earlier. Or it could be a tailender given the role of a lifetime, with nothing but free hits.

The middle-order batter who opens is now so common we don't even call them pinch hitters anymore. Even allrounders going up the order happens often enough, but perhaps not as common as you'd assume in T20s. The most famous recent case is New Zealand's Daryl Mitchell who had never opened before the last T20 World Cup, but seemed to go ok when he did.

Willey is more of a bowling allrounder, but he averages 27 with the bat in first-class cricket. But the tailenders who have a swing have all but gone.

Mitchell McClenaghan was a pinch hitter in the IPL once. R Ashwin, Jofra Archer and Shardul Thakur have done it as well over the last five years but generally, it's not really a part of the IPL or wider T20 culture - which is why Willey doing it stands out.

Part of that is just that teams have specialist batters who hit and have the licence to do so now. Also the business economy of T20 has paid people to hit. The top seven is more flexible than in other formats, so the idea of dropping someone in from outside the hitting group makes less sense. And almost everyone is an opener somewhere now, so it's actually hard to tell who is and isn't a pinch hitter when looking at proper batters.

After working in T20s one of the questions I am asked the most is why don't teams use pinch hitters. And I get it, especially from casual fans. We had them in the 90s, and on the face of it, wickets are more disposable in T20s than in any other format. So if your tail doesn't face many balls, why not throw one or two up the order and let them have fun?

There are a bunch of reasons why it doesn't happen more. Let's start with the most obvious: tailenders don't bat down low by accident. They don't have the batting skills to handle top-level bowling. So throwing them up the order over people who are paid millions of dollars to do exactly that seems like a waste.

Also while using all of your wicket resources in a T20 makes the most sense (as in, you really want your team going hard as possible), there are two sets of resources. One is wickets, but the other is balls. And pinch hitters can waste balls.

The first way is through swinging and missing. Instead of getting their fundamentals right, setting their base, and playing to their strengths, they're usually out slogging really hard. So now you have a non-elite batter swinging before they are set, so obviously they often miss a lot.

If a pinch hitter enters and is out first ball, the commentators often call that a failure. It isn't. The biggest crime is when they come in, eat up balls, and then are out. A first-ball duck is a par result for a T20 pinch hitter; making three off eight is the worse failure.

And not all of those results come from players swinging and missing. Anyone who has ever played cricket will tell you that batting positions mess with the mind. You can shave a rabid monkey and send it out to bat No. 3 and it will suddenly have a high front elbow as it tries to pierce the offside ring. Many pinch hitters just start acting like batters instead of doing their job.

But in the IPL - as well as most T20s - we now have something that wasn't as common in the glory days of the pinch hitter: analysis. A tailender walking out now will be a collection of weaknesses because if they weren't, they'd bat higher. And before, teams barely knew what a tailender could do. It came as a proper surprise because they didn't have access to ball-by-ball or Hawkeye data. Now bowlers and captains know all the weaknesses of each of the XI they are facing.

Scheming for a pinch hitter is really easy. Chances are you know how to dismiss them, or more importantly, keep them to ones or dots, because that is probably the best-case scenario. Get them to chew up their team's resources. A pinch hitter can still do damage. If you have a No. 9 who loves hitting offspin, and a wicket goes in the first ball of one of those overs, you can send them in for five balls of free hits. If they get out, nothing is lost.

It can be tricky for future overs if they get stuck at the non-striker's end. And the fielding captain may remove the offspinner as well, meaning you have upset their plans without even hitting a ball yet. Of course, this is nice, but it means less if your pinch hitter ends up making 7 from 11 balls.

The best way to use your bowlers who can hit is in the powerplay. There are always going to be dot balls in this period. So a couple more from a No. 10 looking to hit out won't matter. Plus, if they do connect, even with a mis-hit, they are likely to go for boundaries because the field is up. But teams worry excessively about losing three wickets in the Powerplay and what that means to their chances of winning the game. Surely losing a No.10 should mean less than a real top-order wicket. There is a lot of data floating around in cricket, but I would assume no one has yet looked at whether losing a tail-end wicket matters as much.

So we see fewer pinch hitters now because there are just so many more downsides to them, especially in a T20. A pinch hitter isn't a free hit; it's usually an unscientific gamble. And the odds are rarely in your favour.

Willey was pinch-hitting in the same game as Sunil Narine was batting in his normal position. Normal is never the right word for Narine, as he has batted in every position in a T20 match, mocking Willey's ten-spots career.

While Willey has batted down the order at Nos. 9 and 10, that's been in powerful batting groups. Narine was a genuine No.10 and even popped in at No.11 a few times. And when he started working on his batting, it wasn't to be a batter but to learn how to hit sixes at the death.

Then Aaron Finch threw him to the top on a whim in a Big Bash game. In the PSL that followed he was used in the middle and before long he was Kolkata Knight Riders' opening batter. Then he wasn't again and ended up back in the middle. He's now not really a pinch hitter, he's a low-scoring, fast-striking middle-overs player sent to maximise the spinners' overs and disrupt opposition bowling plans.

Not as catchy, right?

In truth, he and Willey are semi-permanent pinch hitters, which really doesn't make all that much sense. But they are actually very successful, even if it often doesn't look like it. The proof is that they keep doing it, the ones who fail to do the job never get seen again.

In T20 cricket that is most of them. –Cricinfo