Steve Lawrence is a consultant to Cruyff Football and Ajax in the Netherlands. With a Masters in International Sports Management, Steve founded the Football Analytics Lab and is at the forefront of research into the topic of Relative Age Effects. In this article, Steve looks at the age profiles of the teams involved in Euro 2016 and how these may have impacted the competition.

In Issue #10 of Player Development Project Magazine I wrote an article on Relative Age Effect, which introduced the idea of optimum age profiling for football and other sports teams. I had identified a tipping point age in my paper ‘The Age Advantage in Association Football’ – that is to say, an average team age at which (all other things being equal) the competitive advantage passes from the older team to the younger team.

I had looked at 6,389 matches involving 12,778 teams across 24 competitions in 2013–14 and with Simon Gleave I hypothesised that the tipping point age correlated with a notional ‘optimum’. We subsequently developed a method for plotting the actual age profile of any team against a notional ‘optimum’.

For the UEFA Euro 2016 competition I have been publishing the age profiles for the majority of matches on Twitter under the hashtag #euro2016agecurves.

The charts, recreated in this article, also include an analysis of Relative Age Effect bias in the teams shown as columnar graphs. The two metrics I’ve been using to quantify bias are:

• RAEi = Relative Age Effect index = % of players born January to June expressed as an integer.
• ADfN = Average Deviation from Norm = a measure of how much the team month of birth profile deviates from the norm evident in the general population.

Along with the age curves I’ve also included the ATA to 2 decimal places; it denotes the ‘Average Team Age’ which is the mean age of the players involved in any given match or squad.

The comparative curves for the quarter-final matches are shown below to give an idea of the spectrum of age curves evident amongst the teams at the tournament.

Poland v Portugal

Wales v Belgium

Germany v Italy

France v Iceland

The ATA curves seem to fall into 5 broad families:

1. Close to the normal optimum curve.
2. Peaky curve on the older side.
3. Peaky curve on the younger side.
4. Flat top profile with a degree of symmetry around optimum age.
5. Twin peaks either side of the optimum.

There’s still work to be done in establishing whether there’s a correlation between a particular kind of curve and performance outcomes, but what seems clear is that there is a relationship between performance outcome and the absolute average team age.

This was evident in the results for the 2014 FIFA World Cup where it was noticeable that as the tournament progressed the range of average team ages of the various teams narrowed, tending towards an optimum. Something similar has happened in Euro 2016 where the average age for teams in the group stage was 28.29, but dropped slightly in the last 16 to 28.28. It then dropped further in the quarter-finals to 27.90 and in the semi-finals to 27.71.

The moral of this story is that if you want to stand a chance of getting to the later stages of such a tournament, then turn up with a squad which allows you to field teams close to the optimum age.

And so to England, who turned up with a squad aged 25.83, the youngest at the tournament.

It’s my view that international football is now so competitive that the tiniest margins mark the difference between success and failure. So no nation can ignore characteristics which correlate with competitive advantage. In respect of ‘average team age’ some talent pools will simply be too small to allow nuanced age profiling, some will be significantly influenced by one or two players being prodigiously talented, thus skewing the profile because at a young or old age they are too talented to leave out.

But with England, it’s clear that the team profile was too young – perhaps that was driven by a belief that the available talent trumped any need to ‘age profile’ the squad?

The results, however, speak for themselves and it’s clear to me that The FA has to reassess its attitude to selecting the national talent cohort. I would argue that any nation worth its salt has to engineer a development programme delivering talented players into the national squad so that the national squad can hit an age profile close to the optimum at any given moment in time with players of sufficient technical ability and with sufficient international experience. That, in my opinion, means bigger squads, more training camps, more teams and more international matches. I expect the Premier League may not agree with my point of view.

Here are the comparative age curves for the fateful match with Iceland which speak for themselves.

England v Iceland

Another story emerged at the tournament with Wales, whose heroic journey to the semi-finals will be remembered for years to come.

The age profile for Wales is the opposite of England and it shows how hitting an optimum profile counterbalanced a lack of technical quality in the squad. It couldn’t trump the superior skill of Portugal but it did make them heroes.

Portugal v Wales

Portugal v France

Portugal subsequently went on to meet France in the final of the competition with a profile slightly changed from the Wales match, whilst France were unchanged. The Portugal curve had a flat top profile with a degree of symmetry around the optimum ATA whilst France was peaky on the older side. Portugal had the advantage in being closer to the peak average age but home advantage definitely accrued to the hosts. In the end, proximity to peak ATA triumphed and whilst I know that many are sceptical about the importance of this characteristic, perhaps national team organisers will pay it a little attention in the future.