Celtic v Rangers: Who has actually had the better season?

With Celtic having clinched the league after being an underdog at the start of the season, and Rangers making an impressive run to the Europa League final, the answer is reasonably up for debate.

Of course, ‘reason’ is unlikely to be at the centre of many of the arguments so bringing some analytical heft to the discussion is the goal with this column.

The first two radars below compare the teams within the Premiership and then within the Europa League competition starting in the group stage:

Perhaps unsurprisingly, the radars basically conform with the expectation as Celtic have been superior domestically and Rangers similarly in the Europa League. The Hoops created significantly more xG across the various categories, conceded less to opponents and had higher average quality chances while conceding lower. The outperformance in this regard has been significant and comprehensive.

However, the picture was very different within the Europa League as Celtic’s issues with conceding chances proved a real problem. That manifested in conceding more than twice as many goals per 90 minutes as Rangers (2.50 v 1.07). In addition, Rangers’ proficiency in attacking set-pieces within the domestic league carried over into the Europa League, while Celtic’s did not.

Next, let us have a look at on-ball value for both teams in the league and the Europa League sample:

These show that the sides were relatively close across the two competitions, with a couple of vitally important components standing out. Within the league, Celtic’s shot OBV of 0.26 reflected the relative quality of finishing across the season and the 0.19 of shot OBV conceded versus Rangers reflected the relatively good finishing of their opponents.

READ MORE: Mohanad Jeahze scouted for Celtic

The OBV radar for the Europa League showed the shootout nature of the various Celtic games versus Bayer Leverkusen and Real Betis, with shot OBV high for and against though the Hoops also faced better goalkeeper play. The OBV difference for Celtic was barely positive and significantly negative (not shown) outside of the two Ferencvaros games. In contrast, Rangers’ OBV performance metrics were broadly and consistently solid.

What other variables may help further the debate? How about head-to-head? The following is summarised using the StatsBomb data for the four league games:

This graphic showed the aggregates for each metric across the four games. We have broken out the OBV metric and grouped them into what is labeled ‘core’ and 'shot+goalkeeper'.

The core number is a combination of pass OBV, dribble & carry OBV, and defensive action OBV. Those are relatively high-volume actions within games, whereas shots and keeper actions are very low volume by comparison and subject to far more volatility.

Looking at the metrics this way suggests an interesting delineation in the head-to-head match-ups, as Rangers were extremely poor in both finishing and goalkeeper play, while generally doing quite well in the middle third of the pitch.

The above table is built to combine the Europa League sample with the head-to-head in order to try to obtain a broader view of each team’s performance levels versus the higher standard of opponent.

The aggregate of these metrics was incredibly close, as reflected in an OBV difference which was within 0.03 of one another. I suspect if I had StatsBomb data for the Cup semi game, that difference would likely be close to the same.

Overall, the teams each won two with one draw in five head-to-head games across the whole season. Celtic enjoyed an extremely high level of performance consistency domestically, while Rangers struggled with that at times.

Conversely, the Hoops struggled in Europe and Rangers performed well and consistently.

The analytical conclusion will likely be unsatisfactory to both groups of fans. These were, simply, two very evenly matched teams with relative strengths and weaknesses that suited different competitions. It is, essentially, a 1-1 draw.