How do Celtic stack up in the league with On-Ball Value?

StatsBomb is one of the industry leaders in the provision of football data and one of the leading innovators in terms of developing new ideas and models for measuring football performance. Their work on goalkeeping data is particularly leading edge whilst their expected goals model could be considered the most sophisticated around accounting for ball height and speed in calculations.

One of their signature data innovations is On Ball Value (OBV). This can be applied across passing, shooting, dribbling and carrying, defensive actions and goalkeeping at both individual player and team levels. Both for and conceded. Here is an excellent video explainer.

I will admit to being initially dismissive of this possession state model because there was little explanation to go with the launch. Videos such as the above moved me to a state of curiosity and scepticism, my usual default, as the video was compelling. But I still feel other concepts, especially packing, better articulate some of the action captured.

However, StatsBomb are industry-leading boffins whilst I am an amateur numpty. So, I am warming to it. I wanted to share with you the team-level OBV model positions in the SPFL Premiership to show the relative strengths of the two title challengers through this lens.

OBV Team Vales ‘For’

These are the OBV values generated by the teams themselves, i.e. with the ball.

Pass OBV

Passing accounts for approximately 75 per cent of all on-ball actions in a football match. Data-wise, it is the bread and butter of analysis. Here is the passing OBV for teams in the Premiership:

At 1.5 average per 90 minutes, Celtic have the highest pass OBV of any team by 14 per cent over Rangers. Celtic have the highest possession in the league, at 70 per cent, and with 60 per cent of their passes occurring in the opponent's half, the most in enemy territory. They also play the least number of long balls, being only four per cent of their attempted passes.

Dribble & Carry OBV

Dribbling and carrying as seen in the video is moving the ball forward with the ball at their feet. A much less prevalent action, but as we can see from the OBV values, it can be impactful.

For example, one of the reasons that Rangers had such a successful Europa League campaign in the 2021-22 season is that players like Joe Aribo and Ryan Kent were able to carry the ball large distances and thus alleviate pressure – a valuable commodity when up against superior sides.

And it is a feature Celtic are not as strong this season. Their rivals are 10 per cent stronger in this field. Celtic’s main midfielders Matt O’Riley and Paolo Bernardo are not highly effective here while Callum McGregor’s bursts are more fleeting with age. Ross McCausland, Danilo, Todd Cantwell, Tom Lawrence and Dujon Sterling all feature high in the player charts for this whilst Celtic’s Luis Palma and Yang Hyun-jun are Celtic’s main proponents.

Defensive Action OBV

This view rewards some of the teams at the other end of the table who have much defending to do.

This is one metric that would benefit from being possession adjusted. The teams at the bottom of the league simply undertake more defensive actions. Most will be successful because they are professional players. That said, Celtic and Rangers undertake similar levels of defending, at the margins, at their respective defensive action OBVs are both 0.3.

Shot OBV

Here is where StatsBomb starts to hate Scottish football. Shots. I suppose when your top scorer and league poster boy is Lawrence Shankland, then they have a point.

All sides bar Hibernian and Motherwell are in negative territory in that their shots, net, harm their efforts at scoring more than hinder, on average. Rangers are the worst. Call it the Cyriel Dessers effect. -0.37 on shot OBV. Celtic can hardly crow on -0.18.

Although Celtic take fewer shots, they have more on target and take 68 per cent inside the box compared to 65 per cent by the Blues. Both sides work the opposition's keeper about the same (5.03 saves versus Celtic to 5.07). But not a differentiator.

Goalkeeping OBV

I have covered this before and Celtic with career-ending Joe Hart are in a worse place than Rangers with career-rebuilding Jack Butland.

Whilst Butland is not the England saviour the Scottish press would wish you to believe, he is slightly better overall than a regressing Hart, who is wisely retiring before he falls off the form cliff. Props to Kelle Roos, Dimitar Mitov and Will Dennis who are all having stellar seasons. Liam Kelly being in the Scotland squad for the Euros should alarm us all.

But back to the title race – Celtic won’t get ahead of this indicator and must hope Hart is steady enough through the run-in.

Total OBV

If you look at all that together you get the overall team OBV.

This view supports my eye test of the overall quality of the league with one exception – I’d swap Kilmarnock and Heart of Midlothian. It's a powerful view then when you take in all actions by all players in all matches.

Celtic have the edge; can we see this on the points board at the season's end? Well, it’s only half the story – OBV covers what happens against you from the opponents.

OBV Team Value ‘Against’

Pass OBV

This is a nice proxy for how easy or not teams are to play against.

Whilst Aberdeen has a good goalkeeper and is effective on the ball relative to the league, they are simply too easy to pass through and have the highest pass OBV against. The top two sides have the lowest OBV against, but the Ibrox club has the edge. Their OBV against is 13 per cent lower than Celtic’s. Celtic like to deploy a high line and tend to have one sitting midfielder compared to two and sometimes three by Rangers.

Maybe this influences the relative solidity differential.

Dribble & Carry OBV

St Mirren have a particularly athletic midfield, and this is an aspect of Celtic’s play, whilst improved when Bernardo plays, that remains a concern. It can manifest in being run past with the ball, especially in the middle of the pitch.

Again, the Govan club have an advantage over Celtic in terms of operating against ball carriers.

Defensive Action OBV

By this view, we see the stronger teams at the top end of the scoring, i.e. with the higher values of OBV generated by the opposition.

This reflects how well other teams defend against you. A higher score reflects better defending.  Is a high score indicative of you not giving the opposition enough to worry about, or is there randomness and luck at play in terms of who gets the goal-saving block or intercept in or not?

Difficult to say but what this says is other teams are defending more effectively against Celtic than others.

Shot OBV

Only one side is allowing the opposition to have a positive shot OBV against them.

Aberdeen is simply too easy to play against. Again.

Back to the top pair, and Celtic force slightly more negative shots by the opposition than their rivals (-0.2 to -0.11). Some of this may be defensive organisation and some of it may be simple bad shot choices by the opponents. Celtic allow slightly more shots, but Rangers allow a tiny fraction more on target. Celtic allow more shots inside the box, but the Rangers keeper is forced into more saves.

So, it isn’t a straightforward picture, which is fine, that’s often the case in this game.

Goalkeeping OBV

This view is a little more revealing.

Simply, goalkeepers are having a really bad time against Rangers with a league-leading average of -0.26 per 90 minutes.

Last season Celtic benefitted heavily from poor ‘keeping by opponents so those are indeed the breaks. There may be a bit of ‘finishing’ here but again, Dessers. Celtic are average as regards the opposition's keeper performance whilst St Mirren have come across good custodian performances. Shots are relatively small-volume events, especially those on target, so luck plays a huge part here.

Total OBV

Aggregating all the above then.

Rangers are the hardest team to play against according to this, with the lowest opposition OBV of 1.24. As we have seen some of this is certainly team shape and organisation – not as high a line as Celtic, but more protection from defensive midfielders.

Some of it is due to the normal variance of football–keeper performance, and how well opposition forwards finish.

Summary

Celtic appear to have the edge as regards what they do on the ball. But Rangers are more effective at stopping the opposition. Other than the relative goalkeepers where there is a bit of daylight in their performances, all other aspects remain open to the vagaries of fortune and variance inherent in a low-scoring sport, which is simultaneously comforting and worrying.

Strap in.