At times, Mohammad Abbas bowled beautifully in the first Test against New Zealand in Mount Manganui, notably on that first morning with the new ball but also to start the second innings. He ghosted past both edges, hit them too, and more generally, gave Pakistan control whenever he was on - much as we have come to expect.

At times, Mohammad Abbas bowled beautifully in the second Test in Christchurch, notably on the second morning and afternoon with the new ball. He ghosted past both edges, hit them too, and more generally, gave Pakistan control whenever he was on - much as we have come to expect.

The sum total at the end of all this, though, was four wickets in two Tests, from 76 overs. Only one of those wickets could be considered meaningful - when Ross Taylor fell in Christchurch to leave New Zealand 71 for 3.

Taken in isolation, this wouldn't warrant deep attention. Bowlers bowling well for scant reward is a simple - and occasional - fact of the trade. Except with Abbas, this is becoming a pattern, at least ever since those heady times two years ago when, after a ten-wicket haul against Australia in Abu Dhabi, he became the No. 3-ranked Test bowler in the world.

Two Tests later, after his 12th, he had 61 wickets, at an average of less than 17 and no wonder Dale Steyn - and the world - was getting so excited. But he injured his right shoulder during that Test against New Zealand in Dubai and missed Pakistan's next two Tests.

And, since his return, there has been a lot of Abbas bowling well but very little of Abbas getting wickets. England last summer was a good illustration, where Abbas bowled some fine spells. Remember not just that Ben Stokes wicket, but the trouble he caused Dom Sibley, or a mini afternoon spell to Zak Crawley and Joe Root in the second Test. Yet he ended the series with five wickets (even though it is legitimate to wonder how many more he might have had in that heavily curtailed second Test).

Since the injury, his strike rate has more than doubled, from 42.4 in his first 12 to 93.3 in 11 Tests since. The economy rate has not budged much (2.34 to 2.52) so he's maintained control, but this young pace attack needs his wickets to feed off, and he's only got 23 in that time.

The injury itself is more a marker than a reason, even though there was a sense he returned too quickly - and given the PCB medical department's track record with injured fast bowlers, it will come as little surprise if his rehabilitation was mismanaged.

His pace since has not changed drastically. According to analysis from Cricket21, that covers around 75% of the total balls he's bowled, his average pace has dropped from 128.6kph to 126.6kph. Small dip, though it could be argued that at his already slow pace, perhaps it makes a slightly bigger difference.

The noticeable difference is in his lengths. Before the injury, nearly a quarter of all deliveries Abbas bowled (24.6%) were classified as full by ESPNcricinfo's ball-by-ball data* (includes yorkers and full tosses). Since the start of January 2019 - when he returned in Cape Town - that has dropped to 8.4%.

What defines Abbas, of course, is not so much the full ball as the one he lands on a length - or actually, that in-between length, where batsmen are not sure whether to go forward or hang back. Ball after ball, over after over, match after match.

After his return, there has been an increase in both the percentage of his length and back-of-length deliveries, but at the cost of his fuller deliveries. And that's significant because of his strike rates by length. Before the injury, his strike rate for fuller balls was 31.9 (20 of his 61 wickets). That strike rate has ballooned to 162 since, though the sample size of full deliveries is so small now that it's difficult to conclude whether it's become a less effective weapon. Compounding it is that his strike rate for length balls has doubled, from 41.6 to 95.6.

A fair bit of this will be down to where Abbas has played. In his first 12 Tests, he played mostly on slower, lower surfaces in the West Indies, UAE and England. He thrived, allowing him to go full more often knowing that driving him - with lateral movement around or lower bounce - was loaded with risk. His last 11 Tests include matches in South Africa, Australia and now New Zealand where, generally, surfaces come truer and with more bounce. Driving, especially to Abbas's pace, carries greater reward than risk. Pulling back lengths could simply be Abbas adapting to conditions.

Except that when he played in Pakistan last winter, where he might be expected to bowl fuller more often, his full-ball percentages were even lower, at 4.6%. That suggests something else at play, something out of Abbas' control.

Increasingly, over the last couple of years, batsmen with more exposure now have realised they can defang Abbas simply by taking guard well outside the crease. His pace doesn't just allow it, it actively invites it, messing up his natural lengths and cutting down space for the ball to do something. Cricviz data confirms that the average interception point - where batsmen meet an Abbas delivery - has gone up from 1.83m (from the stumps) in his first year to 2.2m last year, and was 2.55m during the first New Zealand Test.

Recall how prominent a feature it was when England's top order played him last summer, even if the most memorable outcome was the Stokes dismissal. In Pakistan last winter, both Niroshan Dickwella and Kusal Silva tried to take him on similarly, but found, as Aaron Finch did in the UAE with Abbas at his peak, that on those low-bounce pitches, pulling back lengths doesn't necessarily take lbws or bowleds.

The tactic has taken lbw out but it has possibly also impacted the edges Abbas is getting, a number of which have been falling short of slips. For good length and full balls outside off-stump, Abbas took 32 wickets before his injury, eight of which were caught by the wicketkeeper or in the slips region, and 13 of which were lbw. Since then, he's taken eight (two lbws and two caught). Are edges not carrying because batsmen are so far down the pitch, where they also won't be adjudged lbw? Off that line and length there is a very slight dip of 1% across the two periods in the false shots Abbas is inducing, so he's still deceiving batsmen. Yet the final result - in terms of not getting wickets - is significant.

The counter has been to have the wicketkeeper come up. In Karachi, in fact, Dickwella was bowled the very first ball after Mohammad Rizwan came up. But it's not something that has ever looked like developing into a more sustainable strategy and, ultimately, the trade-offs from a keeper standing up to fast bowling over a longer term are unknown.

It's not as if he's going to - or can, at no cost to something else - suddenly increase his pace. It's also not as if he can recreate that Stokes delivery - not least on surfaces with greater bounce - on demand. Maybe there's not much to do for now, except to accept that this is simply the kind of statistical regression to a mean that happens in every career which begins with such freakish numbers (barring one), and that, in conditions more suited, he will remain a wicket-taking threat. Not every batsman will take guard outside and it was no surprise that in Christchurch, Abbas looked at his best - and picked up the wicket of Taylor - when he was bowling to batsmen who didn't. It just so happened that Kane Williamson was one of those, who saw off the best of Abbas where lesser batsman may not have.

Otherwise, it might be worth focusing on developing the other fast bowlers in the attack - wicket-taking in nature - around him, because the control Abbas brings is still gold for Pakistan.

*ESPNcricinfo captures length from where the batsman takes guard

Osman Samiuddin is a senior editor at ESPNcricinfo