How much has Ange Postecoglou improved Celtic this season?

The art of continual improvement is a laudable aim and indeed is firm evidence of a manager adding value to the club.

But what evidence is there in the data that this goal is being achieved?

Note: this analysis covers league football only

Attacking

We will start at the front. Another of Postecoglou’s topics in the presser was that, while clean sheets and defensive solidity were an important by-product of a successful team, in itself this wasn’t the aim.

The Australian has always aimed for his teams to play in a manner that would have made his late father proud – attacking football or, in other words, the Celtic way.

Let us compare StatsBomb’s view of the team’s attacking prowess this season and last:

Here is a summary of the differences:

Both expected goals total and per shot have increased slightly. It’s small margins but, over a 38-game calendar, that would potentially add another 2.28 goals to the total.

A pleasant surprise was that the number of shots following a high press and number of clear shots has increased. The reason for the surprise is hinted at by the flat counter-attacking shots metric. In general, Celtic’s opponents have been sitting ever deeper, even more compact and crowding the middle of the pitch.

Therefore, seeing no increase in counter-attacking shots is probably not a surprise. However, teams are finding it difficult to then break out against Celtic given how deep and compact they sit. Therefore, the champions can press high and aggressively knowing that the opponent has 100 yards to go with few attackers ready to break (hence the improvement in high press shots).

The improvement in clear shots indicates to me that, despite opposition boxes being increasingly congested, Celtic’s forwards (all apart from Oh Hyeon-gyu and Sead Haksabanovic were here last season) are getting ever more comfortable with the expected attacking patterns and movements.

READ MORE: 'Incredible' Ange Postecoglou and Celtic's Lennoxtown battlefield

The total number of shots is down very slightly but still over 19. This is not something to worry about so long as the total expected goals and expected goals per shot are improving; it’s quality over quantity.

The average pass length from Joe Hart has decreased – this is marked as ‘green’ because the keeper being able to play out shorter means it is more likely possession is maintained. It is also indicative of how deep Celtic’s opponents are now, with less emphasis on trying to press the champions high.

The only true blemish is the reduction in cross effectiveness, which is down 15 per cent to a 28 per cent success rate. It is alright to feel fairly ambivalent about crossing and prefer a ‘pass’ to a ‘cross’. In the data world this can get confusing depending on definition and capture protocols. A cross implies a launch it vaguely in the box aspect to the action as opposed to picking out a man. In any event, this may simply be a victim of more congested opposition boxes.

Overall, it looks like Postecoglou’s attacking scorecard is generally positive – definitely so in the more critical key performance indicators.

Defending

A similar methodology for the defensive metrics, again using StatsBomb data.

A drop in both expected goals conceded and average expected goals per shot makes this analyst’s little heart sing.

Do not be underwhelmed by the small differences. Professional sport is a world of tiny margins leading to big outcome changes. These are key metrics so any reduction from what is already a very small starting point, is welcome.

The total number of shots conceded is also down but the standout is a 54 per cent reduction in counter-attack shots conceded. The biggest risk in Postecoglou’s system is being caught in transition with full-backs advanced and the midfielders pushed on attacking the box. This is happening less often, at least in Scotland. Partly this is due to the depth of the opposition and partly due to Celtic players becoming increasing more organised as regards their rest defence.

Rest defence does not mean they are sat down enjoying a cup of Yorkshire Tea (other teas are available). No, it means that, when you attack, what are the other players who are not involved in the attack doing?

Even a Postecoglou Celtic side does not attack with 10 players and the alignment of those guarding the fort is crucial in those transition moments. Don’t underestimate also the dark arts and the Parkhead side’s ability to halt counters at source with a sly foul.  Their relatively high foul count compared to their possession is partly explained by their active transition suppression.

The number of clear shots conceded is up slightly. Given the low base, this is not much of a concern.

Opposition pass completion is down from 68 per cent to 65 per cent. When you look at that in conjunction with the decrease in Celtic aggression (this is termed ‘the proportion of an opponent's pass receipts that are tackled, fouled or pressured within two seconds’) and an increase in passes the Hoops allow before a defensive action, a story emerges.

That is, again, that opponents are sitting ever deeper. Getting out of defensive situations is harder given the distances to travel and the isolation of the opposition forward, hence the lower pass rate exacerbated by Celtic’s ability to press higher as seen above. However, this does reduce the overall aggression and passes per defensive action metrics as, simply, the opposition are not keeping the ball as long for the champions to press.

Finally, the defensive distance increasing is a further sign Celtic are being allowed to push the opposition ever deeper and play higher up the park – all good.

Postecoglou is correct to assert that the more effective their attack, so the easier it is to defend given the lengths to which opponents are now going to stop the champions.

Expected goal difference

I track a six-match rolling average of the difference between expected goals for and against. It is a bit like heart rate or blood pressure in that it is the most basic but revealing indicator of team health.

Here is the profile in league football across the manager’s reign:

Amid the ups and downs of a campaign, the trend is for the difference between expected goals for and against to increase over time as seen by the orange trend line.

The exciting thing here (cue more heart singing) is that the xG difference is now above two expected goals per game on average. If you are generating over two xG more than your opponent each and every game, the chances that variance will lead to misfortune decrease.

That is, the impact of a poor officiating decision (never happens in Scotland, I know), a sending-off, the opposition goalkeeper having a worldy, a poor pitch or poor weather (never happens in Scotland, I know) can largely be mitigated by such a differential.

Steady, incremental improvement is all we can ask.

Summary

With some very minor exceptions, Postecoglou is good to his word and his own personal benchmarking to continually improve this Celtic side – at least in league football.

Furthermore, there is evidence that opposition proclivity to pack defences and play deeper is exacerbating these trends while leading to tiny drop-offs in the champions’ pressing data.

The report card for Europe would be far more complex but that will not matter if league success is not secured. In that regard, Postecoglou is the star pupil.