Did Ange Postecoglou improve the Celtic players he inherited?

This is not as straightforward as saying X or Y metric is better. Players are at different age points, different places on the maturity curve, have different injury issues and it also depends how the player was utilised. It is also more complex than stating “Postecoglou inspires me because of his pithy epithets”.

Let’s have those aspects in mind when considering whether the following pre-Postecoglou Celts have improved over the 2021-22 season: Anthony Ralston, Greg Taylor, Stephen Welsh, Nir Bitton, Callum McGregor, Tom Rogic, David Turnbull and James Forrest. These are players who were at Celtic the previous season and played more than 900 minutes in this one.

For the defenders and defensive midfielders, the focus is on two aggregated metrics:

1. Defensive action success rate – an aggregated measure of percentage success in executing defensive actions

2. Pack pass score – a measure of ball progression, the extent to which the ball is moved through the opposition lines

For the attacking midfielders and wingers the focused metrics will be:

1. Expected scoring contribution (xSC) – which is expected goals + expected assists

2. Chances created – the number of passes per 90 minutes that lead to shots

Defenders

Anthony Ralston

There are probably few who would subjectively state Ralston did not improve massively this season.  In six previous seasons, he appeared in the equivalent of 14.2 matches. This season he contributed 3,862 minutes which equates to 42.9 full matches.

In that sense, there is no comparison to be made. In no prior season has there been sufficient minutes to make any comparison meaningful.

So, with Ralston, let’s compare his data with the right-backs of the three prior seasons: Mikael Lustig, Jeremie Frimpong, Jonjoe Kenny, Hatem El Hamed, Mortiz Baur and Jeremy Toljan (hands up if you do not even remember some of them?).

Firstly, apologies to Frimpong as these metrics do not show him in best light. He is/was elite at world level in bringing the ball forward by dribbling, which isn’t captured here.

Secondly, it is doubtful you’ll see a full-back as defensively strong again as Lustig.

Thirdly, caveats or no, there is no denying Ralston has made a huge impact. The pack passing data can be viewed as reflective of team style compared to Neil Lennon’s Celtic but, even when adding Josip Juranovic, Ralston still has a higher pack pass score.

Has Ralston improved? It is difficult to say. He may always have been capable of this level of performance but hasn’t previously been given the chance to show it.

What we can say is that he is probably more effective than most of the full-backs Celtic have tried since Lustig and is therefore well worth keeping hold of.

Greg Taylor

A bit more straightforward this one. Taylor completed significant minutes in each of the last three seasons for comparison.

Taylor's DASR has improved by one percentage point over his 2019-20 season. What has increased dramatically, as with Ralston, is the pack passing score.

Again, it is worth emphasising that this is as much a consequence of how the players are asked to play as any individual ability improvements. However, it is still indicative of a player able to adapt to the manager's demands.

Also, one perc cent improvement in DASR may not sound much but more importantly it illustrates the tiny margins between perceptions of having a ‘good’ versus ‘great’ season.  Overall we can see tangible improvement as might be hoped for in a young player.

Stephen Welsh

The young centre-back made his breakthrough in the 2020-21 season, playing the equivalent of 26.28 full matches. That means we can make a direct comparison with this campaign.

This is another good news story: young player’s stats improve season over season.

There are, however, two ‘buts’. Firstly, 76 per cent DASR is quite low for a centre-back. Top-end players like Cameron Carter-Vickers, Christopher Jullien, Kristoffer Ajer, Virgial van Dijk would be 80 per cent and over.

Secondly, Welsh’s pack passing didn’t improve to the same extent as Ralston’s or Taylor’s. That is, he does not seem to have adjusted to Postecoglou’s demand for aggressive forward passing allied to maintenance of possession and control (though it is a tough balance to be fair).

So, a qualified success. Welsh is young for a centre-back and, given there is tangible improvement, he is worth persevering with.

Defensive Midfielders

Nir Bitton

Bitton has now left but the hypothesis here is ‘Postecoglou improved the existing players’ and he was at the club the previous season.

The Israeli midfielder is difficult to judge as in prior seasons he moved around from defensive midfield to centre back. There will, therefore, be some positional ‘noise’ in this comparative data.

The above is without the 2018-19 season data as he had a bad knee injury and barely played. What is remarkable is the consistency of performance data in the three seasons where he mostly played as a centre-back.

Last season Bitton played predominantly in midfield, hence his defensive numbers went down and his pack passing numbers went up.

Therefore, we could conclude that the data is telling us more about the role switch than any actual improvement per se.

READ MORE: How Ange Postecoglou's key Celtic players stack up against 2019-20 quadruple-treble winners

Callum McGregor

This has been the first season McGregor has had the midfield anchor role to himself after years of Scott Brown. The data will, therefore, reflect that he has in the past performed a more attacking role. For that reason the last three seasons of Brown’s data have been added for comparison.

The graph clearly shows they are quite different players, which is hardly revelatory. Brown is stronger defensively and McGregor on progressive passing.

What we should recognise is that McGregor’s defensive numbers went up last season, as did his progressive passing volume. Again, we are seeing an impact of team style but also a different dynamic generated from playing a ball player at that pivot position.

To the question ‘has McGregor improved?’ The answer is ‘not sure, but he has certainly adapted’.

Attacking Midfielders

Tom Rogic

This has been Rogic’s most ‘present’ season in terms of playing minutes – but he still clocked fewer than 3,000.

Rogic has been pretty consistent in his output as well as in appearing in less than half the available matches. However this season saw him record his highest expected scoring contribution of 0.72. This is despite playing a more hybrid number 8 / number 10 role than the more orthodox 10 of previous seasons.

His chances created was down a little but we are talking about the difference between 2.27 and 2.52 per 90 minutes.

For Rogic, then, not only did Postecoglou improve his performances but he got him playing more minutes than any other manager too. Probably not a coincidence.

David Turnbull

Turnbull’s pre- and latter Celtic careers have been blighted with serious injury but we do have two full seasons to go on.

He created virtually the same number of chances per 90 minutes as the season before (0.01 difference). Indeed, his chances created from open play was identical at 2.31 per 90 minutes.

Turnbull did have an overall higher expected scoring contribution (xSC). This increased 0.08 per 90 minutes. The value of both xG and xA went up, the latter suggesting that he was creating better quality chances even if not more of them.

This, again, speaks a little to style as Postecoglou wants chances created central with low passes into the danger zone as opposed to high speculative crosses into the box.

So it’s a cautious ‘yes’ for Turnbull on the improvement question in terms of adapting his creative passing to the manager’s demands. His underlying data has experienced an uptick as a result.

James Forrest

We have a wealth of data for Forrest, fortunately, although the last two seasons have been hit by injury.

Not to leave the worst to last but, clearly, Forrest’s performance numbers are not as good as the previous two seasons.

While overall expected scoring contribution is commensurate with the 2018-19 and 2017-18 seasons, chances created has regressed markedly.

We cannot say Forrest has improved under Postecoglou and therefore a new three-year deal is surprising for a 30-year-old winger who relies on pace.

Conclusions

This exercise is as much an insight into the need to always contextualise data as well as answer a seemingly simple question like ‘has Postecoglou improved the players?’

Have the younger players improved because they would have anyway, or because of the way they’ve been encouraged and deployed?

Ralston’s progression – if we can say that given the lack of comparative minutes – has been stark but maybe he was always able to be that effective. Taylor and Welsh have improved modestly, as could be hoped younger players might.

McGregor and Bitton, meanwhile, had to play ‘new’ roles compared to previous seasons and both showed evidence of being able to adapt to those demands. Rogic and Turnbull seem to have responded well to the new manager and showed upward movements in their data.

What is consistent is that almost all players have adapted well to Postecoglou’s demands. Progressive passing data, in particular, supports this. Small, incremental improvement is all that can be asked.