As the Serie A season start progressed this past fall, I had been thinking about examining how it compared to past seasons and how I might assess Max Allegri’s performance. I also wanted to look into the influence of coach turnover on team performance given Juve has changed coaches 3 times in the past five years.
Full disclosure: In my eyes, Andrea Pirlo is magical and can do no wrong, and should never have been fired, but I promise to present the below data without any subjective bias.
Season start win-draw-loss records and points
To compare the start of the 2022-23 season to past seasons, I looked at the win-draw-loss records and total points of the first 15 games from 2018-19 to 2022-23. (Data is from fbref.com in Figure 1.)
Note, two things:
- Since Serie A only played 15 games before the World Cup break happened this season, I truncated the previous seasons starts to this number, and
- These are the first 15 games played, and not necessarily the same as “matchday/week” because in 2020-21 matchday 3 was played in the spring (vs. Napoli).
Clearly the biggest takeaway from these data seems to be that for the past five seasons Juventus has been incapable of winning their ninth game of the season. Kidding — but seriously what an interesting coincidence!
The comparison is not perfect because they didn’t play the same teams in those 15 games each season, there was variable time between games across seasons, and it is hard to compare pre-COVID to COVID times, but even with those confounding variables, Allegri’s start to the 2021-22 season could be viewed as the worst using these metrics. (For context, the number of top teams played — which I view as Inter, Milan, Lazio, Roma and Atalanta — in the first 15 games starting with 2018-19 were 3, 5, 3, 5, 4.)
Comparing this season to last season alone it is a nice improvement, but comparing Allegri’s current performance to his previous stint at Juve we would expect at least a couple more wins in there, and he is not doing any better than Pirlo did in his first season as a coach.
Key performance indicators
Total shots and shots on target are good indicators for a team making it into the top 3 of Serie A, so I looked to see how these numbers compared across the past 5 seasons.
I present the data both in absolute numbers (Figure 2) and in averages (Figure 3) to account for across game variation, but they demonstrate the same patterns:
- Juventus had significantly more shots and shots on target in Allegri’s 2018-19 season than they did in his 2021-22 or 2022-23 season
- Juventus under Maurizio Sarri had similar shots and shots on target to the 2018-19 season
- Juventus under Pirlo had less shots compared to 2018-19 and 2019-20, but similar shots on target
- Juventus under Pirlo had more shots on target compared to 2021-22
- Juventus in 2022-23 has NOT increased their shots or shots on target from 2021-22
As I reported in August, Juventus has to increase their numbers of shots/shots on target. We have evidence from the start of the 2018-19 season that it is possible to have an average of 19 shots a game under Allegri. Juventus is only averaging 13 shots a game currently.
Impact of manager turnover on team performance
I have always been interested in how changing a coach will impact a team’s performance and — if it does have a positive impact — what caused it. Was it actually tactical brilliance or just some sort of psychological effect on the players?
I reviewed the scientific literature on how coach turnover impacts a team’s performance and the main takeaways seemed to be that there doesn’t seem to be a consensus of what the impact typically is (depending on the data and/or analyses, positive, neutral or negative impacts were found), or how to go about doing the analyses to determine it. In any case, analyses done specially on coach turnover in Serie A found that coach change had no significant impact on team performance (using data from 1997-98 to 2008-09) and that it had a negative effect on the team performance and points earned at the end of season (using data from 2007-08 to 2016-17).
However, both these analyses were based on coach changes during the season and I could not find anything on impact of coach changes between seasons for Serie A, which is what happened in Juventus’ case.
There are some interesting hypotheses to explain the potential impacts a new coach can have (reviewed in this paper), including for when positive impacts occur:
- the “common sense” hypothesis, which indicates that the previous coach was the issue and replacing the coach improves things
- the “mean reversion” hypothesis, which is based on the idea that the new coach actually didn’t do anything to improve things and the team was likely to revert back to a particular average performance anyway
and when neutral or negative impacts occur:
- the “ritual scapegoating” hypothesis, which postulates that the previous coach was dismissed to show that the board was doing something but the new coach doesn’t actually improve things
- the “vicious circle” hypothesis, which states that previous coach was not the issue and replacing them with a new one just makes the situation worse and the team performance decreases.
With regards to changing coaches, Juventus in the past five years is a unique case. Typically, coaches are replaced when the team is under-performing. Allegri won Serie A in 2018-19 and was replaced, and then Sarri won Serie A in 2019-20 and was replaced. The outcome of replacing Allegri with Sarri does not fit with the common sense hypothesis, and to some extent neither does replacing Sarri with Pirlo. But there seems to be evidence for the ritual scapegoating and vicious circle hypotheses for Allegri’s return. Juventus’ performance in 2020-21 was not likely entirely due to Pirlo’s ability as a coach, but rather a number of other factors, and by replacing him and bringing back Allegri, the situation did not get better and to some extent just got worse.
I like Allegri, but I don’t need an analysis to state that, in my opinion, bringing him back was a mistake. It was a mistake before he even coached a single game. There is a term that is used in my field, “solastalgia,” which means missing a place that doesn’t exist anymore due to environmental change. It doesn’t directly fit the situation I am trying to describe here, but it is what came to my mind when Juventus brought back Allegri. They wanted to recreate something they had in the past that simply did not and could not exist anymore.
Or perhaps a better analogy might be that you “cannot go home again” — you cannot relive or replicate the past now.
Something else I had also been pondering, as the start to the season presented disappointment, was perspective. I happened to be reading a history of Juventus and it seemed to help me not get too down on the team. It was like taking a big step back and seeing the team across its 100-plus years and all its successes as well as ups and downs made the start to the 2022-23 season seem like just a small moment. It was sort of like Neil deGrasse Tyson’s cosmic calendar to put into perspective life on earth.
Anyway, I wanted to see the history of Juventus in numbers, so I plotted out their end of the season total points, points by game 15, and their end of season position in the Serie A table across time starting from the 1929-30 season to today (data from transfermarkt).
Several things to note about the data and Figure 4:
- Prior to 1994-95 season, Serie A awarded two points for wins instead of three, so I recalculated out those using 3 points to make it more comparable across time
- The number of games played across the timeline varies from 30-38 due to varying numbers of teams in Serie A (see Table 4): currently Serie A has 20 teams and plays 38 games in a season, but from 1988-89 to 2003-04, there were only 18 teams and 34 games played, and prior to 1988-89 there was some variation between 30 and 34 games. This will effect the total season points over time, but not the points at 15 games
- For the 2006-07 season I put in the points but technically they had 0 points and were in 20 position due to Calciopoli
- In the 2007-08 season Juventus was in Serie B and played 42 games
- I could not make the text any larger without making the figure too large, so I am sorry the text is so hard to see!
Using these Juventus data, I ran a couple of regression analyses and found there are moderately positive relationships between points earned and position held by game 15 and those results at the end of season. This is silly and not something to put any money on, but since I had the data and equations right there, I calculated what we would expect Juventus to end this season with using the previous seasons’ relationships, and calculated that Juventus will end the season with 69.02 points and would be in approximately third place (2.33 using the points they have now, and 2.73 using the position they have now).