Amidst the utter frustration of another set of Champions League group stage points being left on the field by Celtic, a timely reminder. In professional sports, we, as fans, fixate on the outcome, the result, and how it makes us feel.

Professional coaches will largely fixate on performance, as that is the part they can control whilst recognising the outcomes dictate their employment prospects. That is not to suggest in any way that Brendan Rodgers’s stock is harmed after the latest home defeat 1-2 to Lazio (no win since October 22nd 2013) but that when we pick the bones out of it, the most important aspect in terms of what happens next, is the performance analysis.

READ MORE: The Celtic numbers: Lazio capitalise on wasteful home side

This loss, and that to Feyenoord in Rotterdam sting more than those of last season under Ange Postecoglou - for me, anyway - due to the relative control and lack of threat the opposition posed.

So, the outcome is the same, but the performances changed, is my hypothesis.

New Campaign, Same Outcomes

This is what the media will focus on, as well as significant parts of the support. Let’s compare the performances last season and this so far (all small samples, remember) in terms of firstly, the attacking metrics:

Last season, Postecoglou took the view that the team were going to attack relentlessly as if at home to Dundee and to hell with the consequences. Die with your boots on, and all that.

His more reasoned analysis is that you may as well discomfort the opposition as much as you can rather than bunker in and wait to lose. Celtic certainly ‘had a go’ in all their matches, even in the Bernabeu against Galactico-fuelled Real Madrid.

At this stage after two games (one heavily undermanned for a significant period), Celtic’s attacking potency at this level has fallen off a cliff. xG per game is almost halved from 1.32 to 0.68. Shot quality is down from an average of 0.1 to 0.07 and shot volume has decreased from the 77th percentile 13.67 to the fifth percentile 9.5.

The only piece of attacking play in StatsBomb’s view that has improved is the cross-completion percentage.

Again, playing for nearly 30 minutes in Rotterdam with only nine players has a profound effect on this but there is no doubt Celtic’s attacking potency at this level has dramatically decreased.

What about defensively?

This tells the opposite story. Last season, Celtic were arguably one of the worst teams in the Champions League defensively (actually, there was one side that was significantly worse). Conceding 1.72 xG per match, giving up over 13 shots and allowing a high proportion of both clear shots and efforts from counterattack opportunities, Celtic were a mess at the back.

Incredibly, despite two defeats and four goals coughed up, Celtic have one of the best xG against figures in the tournament with a measly 0.68. Furthermore, the average opposition shot has an xG value of 0.06, which is probably the best in the competition in terms of limiting good chances.

READ MORE: Celtic instant analysis as late Lazio sickener stuns Rodgers

Celtic are in the 100th percentile for allowing clear and counterattacking shots – another excellent outcome. The team still allows for too many unhindered opposition passes (15.06 per defensive action), and the opponents enjoy an 86 per cent pass success rate – as per last season Celtic are quite easy to pass against – indeed the easiest in the competition.

What is surprising to me is that the average defensive distance has INCREASED to 51.09 metres from their own goal. This, despite lacking pace and experience in the centre-back areas. In short, Rodgers’s side is squeezing the game even further than Postecoglou’s relatively high defensive line. It seems Rodgers is building from the back in his tweaking of Celtic’s approach. This has blunted the attack but made the team far more robust defensively.

Team Shape

What has Rodgers tweaked to effect this dramatic change in attacking and defensive performances?

Let’s compare the average team positions from FotMob (Opta) from the RB Leipzig tie last season to the Lazio loss this week.


Nat Phillips was dropped deeper than Cameron Carter-Vickers last season, which is why I was surprised about the overall defensive line data. What we can see from the above is that the central midfielders are slightly more pushed on, but the forwards are slightly deeper. Such small margins have the effect of making the team more compact overall and this is harder to play through.

The trade-off is that whilst the nominal number 10 versus RM Leipzig (Sead Haksabanovic) pushed right up top with Kyogo Furuhashi, Reo Hatate tended to be much deeper against Lazio. Similarly, Daizen Maeda was much less aggressive in getting forward than Liel Abada had been a year ago.

Overall, Celtic sacrificed both numbers in attack and depth of attacking positioning to make the midfield and forward defensive lines more compact.


The Sofascore momentum bars are a useful indicator of the flow of a match in terms of which side is dominating both possession and field position over the 90 minutes.

Again, here are the comparisons between last night and the home match-up versus RB Leipzig:

Under Postecoglou in Europe, Celtic tended to have bursts of very intense attacking interspersed with long bouts of opposition control. Whereas against Lazio, apart from the period after the Celtic goal, the home side maintained a decent level of overall control.

Bear in mind also that RB Leipzig are a very direct and vertical passing-orientated team, whereas under the cigarette-munching Maurizio Sarri, Lazio are one of the most possession-hungry and ball-dominant sides in Serie A.

Despite this, Celtic exacted admirable control of the ball for long periods in the second half last night. Whereas Postecoglou’s side would have been more direct, Celtic were able to be less expansive but more controlling. This obviously limits the opposition in terms of attacking – Celtic had to reduce the chances being offered up given last season’s 16 goals conceded.


Despite results (two losses) and goals (1-4 against) not reflecting, Rodgers has spun Celtic’s Champions League performances on their heads to become (after only two match days) one of the most miserly defensively whilst not hitting last season’s heights in attack.

But good sides are built from solid defensive foundations. And as certain risks were not addressed – specifically a more agile goalkeeper - arguably two goals have been conceded (the free kick in Rotterdam and the headed corner equaliser last night) that a younger, faster more bendy ‘keeper saves. Add in a chronic centre-back injury list and this compounds defensive risks (bad luck).

READ MORE: Brendan Rodgers rips into referee's 'strange' Celtic decisions

The attacking deficiencies as part bad luck (Abada’s injury) and part the failure to adequately replace the creativity and goal threat of Jota and Aaron Mooy. In the round then, if Celtic can get their best XI on the field and continue to set up as Rodgers has in the first two games, there is no reason they cannot get points in this group.

Celtic need to either dramatically improve the quality of output coming from both the wing positions and the number 10 position, or alternatively, switch to a 4-4-2 diamond to get the best out of Maeda. A little good luck would certainly help as well.