As we get to the downslope of another Champions League campaign, and start the second round of SPFL fixtures, Liam Scales is the immovable object at the heart of the Celtic defence.
Seemingly heading out the door in the summer, he is now first pick on merit despite the likes of Lars Lagerbielke, Stephen Welsh and marquee signing Maik Nawrocki being available to partner the un-droppable Cameron Carter-Vickers.
Where Scales sits within the population of Celtic centre halves is a much bigger piece of analysis.
Today I will focus on the question of the extent to which he has replaced previous incumbent Carl Starfelt.
This compares Scales over 12.1 SPFL matches with the 29.1 Starfelt appeared in during the 2022-23 season.
Celtic Way: Scales vs Starfelt
Scales vs Starfelt (Image: Statsbomb)
Once again, a warning that in general Celtic defenders will be at the top of many percentiles due to the team dominance. However, comparing two Celtic players in the same position is largely fair game.
Firstly, Scales completes more defensive actions – 9.96 to 7.05. This is not necessarily a “positive” and indeed likely reflects the higher degree of defending a slightly weaker Celtic have undertaken this season especially given the adverse fixture distribution for the initial 11 matches.
Scales has blocked very slightly more shots, but again that can be a volume due to increased defending thing.
His aerial win percentile is much better at 88 percent versus 79 percent. I would say this was a Starfelt weakness, and it is good to see that Scales is more aerially dominant.
Scales also applies more pressures on the opposition. I suspect this is also a team strategy feature rather than a personal attribute. Whilst Ange Postecoglou wanted his central defenders to be defenders first and foremost given the numbers committed forwards, Brendan Rodgers is slightly more flexible. And Rodgers plays an even higher line that Postecoglou meaning the centre backs are nearer the opposition to engage in pressures. Scales’ average defensive distance is three metres higher than Starfelt’s.
Otherwise, Starfelt is rating ahead of Scales in most other metrics.
Most noticeably, Starfelt was better facing opposition player dribbles given his relative agility and pace.
The Swede also had a higher volume of tackles and interceptions. Whilst Scales is a “front foot” defender as regards aerial duels, Starfelt was very aggressive in closing players on the ground as well. This was possible given his superior ground speed.
Scales is slightly more successful in regaining the ball from defensive actions (both around the 50th percentile). Starfelt’s decision making tended to be over optimistic.
It is to Scales’s credit that as far as defensive performance is concerned, he is largely matching the two year older and more internationally experienced Swede.
However, what you cannot compensate for is speed.
Whilst I felt Starfelt was too aggressive and optimistic in his defending, and Scales is more circumspect especially on the ground, the Swede had the recovery pace to make up for his not infrequent errors.
I will revert to may own packing data for this.
Starting with pack passing overall score. This rewards the player for the number and position (more points for defenders) of players taken out the game by forward passes.
Scales is by far the more effective breaker of lines through passing. His pack passing score is 36 percent higher than the current Real Betis defender.
The volume of such passes is lower in terms of differential. Scales completes 17.6 pack passes and Starfelt was 14.5 last season.
This means that Scales does not necessarily attempt Glenn Hoddle-eque 50 yarders over the defence, but rather consistently works the ball through the initial press.
One of the criticisms was the time he took on the ball and how slow he was to recycle. The increased volume of passes completed relative to Starfelt suggest he may be improving in that aspect.
Remember, however, that Starfelt was not a great passer of the ball overall and tended to “keeping it simple”.
Hence not the toughest benchmark.
Turnovers and Recoveries
Another aspect of packing is the impact of both turning the ball over and recovering it in terms of teammates or opponents taken out the game by either action.
A higher pack recovery score is better. Scales edges this but in general centre backs do not recover the ball under packing conditions as they are generally the last line of defence and so do not have attackers behind them when they win the ball.
More pertinent is the pack turnover data. You want the centre backs to turn over as little as possible for the same reasons as above.
Starfelt had a higher pack turnover rate than Scales which is good for the Irishman. However, Carter-Vickers is on 9.31 whilst Scales is on 12.79.
So, an improvement in relation to Starfelt but short of the Carter-Vickers benchmark and standard.
Scales is preferred for his dribbling and carrying and passing – therefore ball distribution.
Neither score very highly on defensive OBV – Scales 0.02 and Starfelt -0.02. Scales is only on the 48th percentile. Starfelt was 2nd percentile which is not good!
Scales must be commended for keeping his place through a difficult set of fixtures both domestically and in Europe.
As far as performance levels go, he is barely a drop off from Starfelt and it is only really recovery pace where he is lacking.
As regards ball progression, it can be argued Scales is an improvement.
