Rodger’s Celtic were moulded to play in a high pressing style out of possession. He is heavily influenced by Guardiola whose Barcelona team were famously assertive in the press out of possession. Has Lennon changed Celtic’s defensive style?
PPDA
For a number of years Passes Per Defensive Action (PPDA) has been accepted as a useful indicator of press effectiveness. Simply, it is the number of opposition passes attempted divided by the number of Celtic defensive actions (tackles, interceptions, fouls). The higher the number the more passes are being allowed (low effective pressing) the lower the number the less passes are being allowed (highly effective pressing).
Lennon v Rodgers PPDA
Any individual game is dependent on many factors so I have taken a 6-game rolling average across their respective reigns.
The orange verticals indicate the end of the respective season. The red vertical indicates when Lennon took over.
The trend line (black horizontal) shows that Celtic have been increasing less effective at pressing over time (i.e. they allow more passes per defensive action). Bear in mind this is all within quite a narrow range (between 2 and 5 passes allowed per defensive action).
The peak, especially in the early parts of the season, likely correlates with playing European ties – much harder opposition and different styles to counter.
Another trend is that Celtic tend to refocus after the new year break and play with higher pressing intensity at the sharp end of the season. This has been true over all four seasons.
The pressing intensity increased (a dip in the graph) markedly after the mid-season break under Rodgers then increased markedly just as he departed (included the tough Valencia games). Lennon was mindful not to change too much and pressing intensity returned when the Irishman took over and got the two trophies over the line.
Last season, pressing intensity decreased steadily over the seasons until post new year again when Celtic played with renewed rigour.
As mentioned, the European games add some skews so here is the data with SPFL matches only, again 6 match rolling average.
The trend remains the same – pressing intensity was decreasing under Rodgers and that has accelerated under Lennon.
When analysing league games only, the PPDA spiked up as soon as Lennon took over and has been at a higher level (indicating less pressing) ever since.
Conclusion
This isn’t a “Rodgers good, Lennon bad” article. Pressing is certainly a popular defensive strategy deployed by the likes of Liverpool, Bayern Munich, Manchester City. But it isn’t the only way to defend. We know that Lennon sought less possession for its own sake and more direct attacking in offense. He seems to have a different view of defensive strategy also.
This is an exercise in assessing one particular stylistic approach. As long as the wins and the trophies keep coming it won’t really matter if pressing is seen as less effective (or being executed less vigorously).
Francis Fallan says
I feel that both Managers did employ a pressing defensive policy, but it also at times left out defence totally wide open on both flanks, as the two back players apparently went down both wings at the same timd, and the evidence has shown left out goalkeeper on his own with the outcome of losing goals.
So whilst pressing defensively has it’s good parts it also shows it can be dangerous. My own observations show that when the team keep a regular defensive player option then such a risk is minimum and hence can be controlled as the players involved have a well established awareness of his they react in a given set up.
SFTB says
How does this nuanced change in pressing affect Celtic?
Are we losing more goals under NFL? Are wescoring more to counteract this trend?
You have probably produced an article on this already but it might help to have some hperlinks at the end for related comparisons
Finn McCool says
Hi Allan,
I’ve just been catching up on your June articles.
This one in particular caught my eye. I moved from Frequentist statistics to Bayesian about a year ago. This was a lot to do with other ‘sciency’ analyses I was involved in and I was never convinced about xG numbers in football stats. You don’t get 1.13 goals per game. You get 1 or 0. A binomial distribution.
So with your analysis of PPDA would it not be better to have a more granular look at it by:
Splitting into home and away games
Ranking the opposition on league position/ strength of opponent
Weather conditions
Referee 🙂
And, of course, the great leveler, Fortuna.
It may lead to the counter-intuitive conclusion that PPDA is not a god metric for winning games or scoring goals. Of course, I may be talking out my arse 🙂