Josh Hobbs looks into the relationship between success and squad ages to see if there is any correlation between age and performance in football.
With speculation swirling around his future, Manchester United manager Ole Gunnar Solskjaer made a very interesting selection decision for their recent game against Tottenham Hotspur. The Norwegian chose to go with his oldest ever starting 11 (averaging 28.27), relying on experience in a must-win game.
In the game, Cristiano Ronaldo (36) and Edinson Cavani (34) came up with two moments of top quality and this saw the Red Devils come out 3-0 victors at the Tottenham Hotspur stadium. Solskjaer said in his post-match press conference that ‘experience had shone through’.
Whilst it must be said that Spurs were extremely poor on the day, it is clear that Solskjaer got what he wanted from his experienced side as they picked up the win. However, it’s worth exploring whether going for experience in the team is something that can be beneficial in the long term rather than just in a one-off game.
In the English Football League, it is often said that it’s necessary to have players who have ‘been there and done it’ in order to win promotion. Older heads are often spoken of as being necessary to ensure that any team has the mentality to make it through a grueling season where dips in form are inevitable.
Whilst this is something which is often mentioned in conjunction with the EFL, there is a prevailing view in football that experienced ‘proven winners’ are vital in delivering success to any level of team. With that in mind, let’s have a look at the relationship between squad age and performance.
Taking Europe’s nine highest-ranked leagues over the last two seasons and plotting each team’s average age (weighted by minutes played) against expected goal difference (xGD), we end up with this scatter graph:
As you can see, there is no relationship between squad age and performance.
In the introduction to this piece, I mentioned that, for some, the thinking is that older players bring experience, and accordingly, leadership. If the received wisdom is true, then we would perhaps expect to see that borne out in defensive performance, where those leadership qualities and intangibles can really shine.
Of course, there is no event data that can measure an experienced centre back’s ability to organise a defensive line. However, our analysis does suggest that there is no correlation between defensive performance and age:
Even when it comes to the elite teams, whose older players are more likely to be world-class players—or at least once were—we still find that squad age and performance have no correlation. In this graph, we’ve taken the top 10% of the first graph in terms of xG difference to represent the elite sides in our data set:
Despite being one of the youngest teams on this graph, Ajax’s dominance in the Eredivisie stands out. It is worth noting that the Eredivisie is a very young league in terms of average age, though, which we will explore further later on.
Manchester City and PSG’s 19/20 seasons are the next most dominant and they sit closer to the far end of the x-axis. However, it should be pointed out that their squad average is still younger than 28 which is considerably younger than the top average ages on the previous graph.
This suggests that an ‘old’ squad in elite terms isn’t the same as an ‘old’ squad amongst the whole of the top 9 European leagues. In this case, we’re talking at least two years younger on average than the oldest squads.
So why not skew young?
As we’ve seen from the analysis above, squad age skewed young or old has no impact on long-term performance. With that in mind, it’s worth thinking about other factors which could impact choices around whether to skew young or old in terms of squad building.
The obvious conclusion regarding those squads which are skewed on the older side is that they are going to be a greater drain on funds without bringing greater performance. This is due to the fact that older players are likely to cost more in wages, whilst those beyond peak age are likely to have a rapidly decreasing re-sale value. Therefore, skewing younger is cheaper in terms of wages and fees, and brings with it a greater likelihood of profit generation in the transfer market.
There are a number of clubs who have recognised this and have incorporated it into their squad-building policy. FC Nordsjaelland in Denmark stand out as a club built on youth. They finished 6th in the Superligaen last season with an average age of 20.29, almost 4 years younger than the average age of the league.
Nordsjaelland’s young players are developed through two different pathways. Primarily, the club brings in youngsters through the Right to Dream academy which owns the club and brings players from Ghana into the country. They do, however, bring in young Danish players too.
There are other examples of other teams which skew young in their first-team signings. In the Premier League, clubs like Norwich and Brentford have made reputations on unearthing high-potential young signings which they have gone on to benefit from their on-field performances before making big profits from them.
James Maddison and Emiliano Buendia both joined Norwich in their early 20s before being sold for over ten times their original fees. More recently, Brentford made huge profit on Ollie Watkins and Said Benrahma who both went on to sign for Premier League sides. This season’s transfer window is the first since 15/16 that Brentford haven’t sold a player for at least £10m.
This summer saw clubs like Arsenal and Crystal Palace focus heavily on acquiring ‘pre-peak’ players as they looked to build squads for the future as aging stars moved on. Both of these clubs are in transitional seasons: the Gunners are trying to re-establish themselves as a top club, whilst Palace are looking to move away from an old squad playing defensive football into a more front-footed team featuring young, exciting players.
Of course, there is an element of risk to this kind of approach. Interestingly, where there does appear to be some correlation between squad age and performance is amongst the worst-performing teams:
If you take the bottom 10% of the teams in our original graphic, you can see that there does seem to be a significant trend away from younger squads. In this sense, if you are a club who are likely to be in a relegation battle going into a season, then a preference for older rather than younger players might be sensible.
Average Ages of European Leagues
As we touched on earlier, Ajax were the most dominant team in our analysis, despite being one of the youngest in the top performers graph. However, as we noted, the Eredivisie is one of the youngest top-tier leagues in Europe. As you can see from this table—which uses data taken from Transfermarkt to show the 25 most valuable top-tier leagues in Europe—the Dutch top tier is the 3rd youngest by average age:
It’s worth mentioning that, as we could see in the earlier graphs, Ajax’s 19/20 and 20/21 squads sat just above the average age for the league but not enough to be considered ‘old’.
The other thing to say is that the Eredivisie has some very valuable young players. This is illustrated by the fact that, although the league is the third youngest, it’s also the 9th most valuable in terms of cumulative player value according to Transfermarkt.
At the other end of the scale, three of Europe’s ‘Big 5’ leagues sit in the top five for the oldest average age: the Bundesliga and Ligue 1 buck the trend and skew significantly younger than La Liga, the Premier League and Serie A. Whilst teams in La Liga, the Premier League and Serie A can for the most part, probably afford the financial cost of skewing older, it seems likely that the other leagues in the top ten for average age are overpaying for their talent.
When you consider that Turkey, Greece, Russia, Scotland, Hungary, Cyprus and Romania’s top tiers are the other leagues that make up the top ten, this reinforces that view. The Russian Premier League is probably the only one of those leagues which sees major sales come out of it with any regularity, so most of the clubs in those leagues are sinking the majority of their budget into aging players which they won’t get any future return from other than if they are able to lead their side to success on the field.
Amongst these leagues, there are some teams that are attempting to do things differently to the rest, recognising that they struggle to compete financially.
In Greece, the average age of the Super League 1 is 26.8 but Asteras Tripoli have an average age of 24: 2 years younger than the next youngest team. They are currently midtable in the division, despite having the lowest average market value for players in their squad:
In the Russian Premier Liga, only three teams have an average age below 25 with all of them averaging 24.9. One of those teams is Lokomotiv Moscow, who currently sit in 2nd place after 13 games, 5 points behind leaders Zenit St Petersburg. However, Zenit are the highest-ranked team for average market value in the league with Lokomotiv the fifth most valuable on average:
As Zenit have one of the oldest squads in the league, they are unlikely to be able to make money back through player sales, However, the Lokomotiv’s value is likely to grow as their players mature.
Another example comes from Hungary. Zalaegerszegi are the team with the youngest average age—24.3—and sit in mid-table. Again, their squad is very cheap in terms of value as they have the second-lowest average value in the division.
All of these clubs are performing very well in the table, relative to their budgets, and are set well for the future due to being younger than their rivals. Considering the leagues they compete in have such high average ages, these teams seem to have made a conscious choice to operate differently and it seems to be paying off. It will be interesting to see if they find themselves in stronger positions over their rivals in the next few seasons.
One thing that I haven’t addressed throughout this piece is that the performance vs age graphs obviously use underlying numbers, rather than other metrics like points per game or actual goals scored and conceded.
Particularly when it comes to points per game, it might be worth cross-referencing the analysis we have already done with expected goals difference against league performance to see whether there is any weight that should be given towards the idea that the intangibles that experience can bring are crucial in a relegation battle or to a team fighting for titles and promotion. It may be the case that more experienced teams are more consistent in being able to ‘grind out’ results despite not performing well. However, expected goals were used in this case as this metric is widely regarded as the best for gauging long-term performance.
It would certainly seem from anecdotal evidence from players themselves that experienced leaders can play their part in driving a team onwards. However, these could well be a smattering of older heads in a younger-skewing squad, rather than needing to be a squad pushing towards the high end of the age range in their particular competition.
Also, it’s clear that managers in situations like Solskjaer found himself in before the Tottenham game are often likely to go with players they feel they can trust and have records of success. Individual games are a different thing to long-term performance though, as the analysis has shown.
In that case, whilst there’s still an important role for experience within a squad, it’s better to aim for a lower average age in order to avoid the situation that Crystal Palace found themselves in this season where 9 first team players moved on for free at the end of their contracts and almost £70m was spent on replacements. As was mentioned earlier, this has looked successful so far, with the side firmly in mid-table. However, even just a couple of the new players failing to settle could have put Patrick Vieira’s side in real danger of relegation this season.
Should Palace be able to make a good profit on one of their new signings from this summer—which given their early performances in the league there seems a good chance of—this could be the first fruits of a new transfer model for them and lead them onto steadily increasing the overall quality of the squad. This should allow them to compete higher up the table in the future, which is of course the ultimate goal in all squad building.