Is next season's Chelsea squad too young to win the title?
(De Bruyne cracks me up in this one.)
So, this has taken awhile; this is partly because I'm a little lazy and partly because this took a fair bit of time to complete because I had to open everyone's Wikipedia page individually, get the birthdays, then do the math. Fun times.
But what is it, you ask?
In conversations on the board it has come to my attention that the fans (yes, you) are worried that we, Chelsea Football Club, have not spent well enough to bolster our needs. Need a striker? Lukaku is like a new signing! Need a midfielder? Here's a 20 year old! Are we Arsenal, now? We are too reliant on the youth and thus cannot win the title because the squad is so young. Mourinho has echoed as much.
But does that belief hold true?
I wanted to investigate this phenomenon.
First, people, this is a PSA: please do not use
You can't win anything with kids.
because using that quote would be like using General John Sedgwick's last words to win an argument:
"I’m ashamed of you, dodging that way. They couldn’t hit an elephant at this distance."
He was promptly shot between the eyes.
The former quote is from Alan Henson, long-time Liverpool player and current pundit. He uttered this statement following United's 3-1 defeat to Aston Villa on the opening day of the 1995-96 season; as you who can use Wikipedia could guess, United did, in fact, win something with kids.
But, can we? So I did a study.
To calculate squad construction and such I thought about the best way to present the data. It's not that complicated, but I wanted to figure out which number was the best. So, I got the birthday of every player in the squad and all their appearance dates. I then sorted them by the amount that they played,* and then used Excel's logical functions to generate the number of players that fit into each of my three age categories: Olympic (under-23), Prime (23-30), and Old (over-30).
*I didn't have minutes data for the teams, especially earlier on, so I simply counted starts as one full appearance and subs as 1/3 of an appearance. The methodology is flawed, of course, but I don't really have another method, and I kept it consistent throughout calculations.
I compared several title-winning squads: 93-94 United, 94-95 Blackburn (outlier!), 95-96 United, 97-98 Arsenal, 03-04 Arsenal, 04-05 Chelsea, and 06-07 Chelsea. I also have last year's Chelsea squad, this year's projected squad, the 91-92 Milan team, and the 08-09 Barcelona team* for comparison.
After sorting the players out by age in the total squad I then wanted to see how the 'first-team' squads were composed by age and also the first XI. Note, my first XI calculations have absolutely no knowledge of how the team lined up, I just used the eleven most featured players. The 'squad' designation was given by averaging the Adjusted Appearance figure for every player who made an appearance then using one standard deviation below that value to generate a cutoff.
After that, I then averaged all of those title-winning PL squads plus the Barca team (because the Milan team didn't give me enough data to use my category designations) to find out what the 'average' great team looked like.
I took all of that data and stuck it in Excel and generated a cool table. Awesome. The following is a screenshot of what the average team's data sheet looked like...
All of the data, collated and color-coded by team
So, for comparison we should talk about what I think next year's team will look like. I predicted this starting XI...
(Yes, I have Van Ginkel in it, sue me.) So, this team would break down as four Olympic players (Marco Van Ginkel, Romelu Lukaku, Oscar, Eden Hazard), 5 Prime players (David Luiz, Juan Mata, Branislav Ivanovic, Cesar Azpilicueta, John Obi Mikel), 2 Old players (Petr Cech, Ashley Cole). Is that team too young?
Well, no: there's actually an example of a team that's composed in exactly the same ratios, and no bonus points for guessing that it's the 95-96 United team that couldn't win with the kids. So there's hope. Our squad is slightly younger than the 'average' great team by a percentage of players in the younger categories, but we are within the bounds of reason.
What does the 'average' great team look like?
So, it looks like the 'optimal' team is composed of the majority being in their prime ages and having more good youngsters than good old guys. We're set in that regard.
So, I think there's this belief that young guys aren't really that capable of winning due to their mentality, or whatever. Well, I don't think I agree with that and the data doesn't really support that; you can win with kids, as long as the kids are good enough. The United team that broke through in 95-96 had Giggs, Scholes, Beckham, Gary and Phil Neville, and Nicky Butt all fighting for first team spots when they were under-23; all of those players were pretty good. We have KDB, Oscar, Eden, Van Ginkel, and Lukaku; not bad at all, I'd say.
I think the stereotype associated with winning while young has more to do with perception than fact; however, there are certainly a few factors which explain why teams are generally between 23 and 30. In my opinion:
1) Young players are harder to prospect than old ones. There is a high level of difficulty associated with finding the best players before those players are 23: small sample sizes and bad competition levels abound. Essentially, at the youngest age, players do not gravitate towards big clubs; they're more likely to be playing for a local team than a large team, so the youth talent is distributed much better throughout Europe than at the high levels. Thus, it's harder for big clubs to corner the market on hot prospects.
As you move up in skill level, the bigger clubs start noticing those talents, so the market tends to squeeze towards the top. This process is generally not completed until a player is 25 or so; by that time, great players are usually with great clubs.
2) Simple numerical possibility. The under-23 cutoff in age is actually a small categorization, because it only consists of about three or four 'high performance' years, while the 23-30 age range consists of seven or eight.
- Mourinho was kind of a tinker man in his first title-winning season; the standard deviation was almost as large as the average in that one. This is possibly because the squad was so big, I'm not sure.
- Benitez was ridiculously a tinker man. I only ran the appearance numbers or whatever, but essentially he played his defenders every game (Carragher started every single match in 04-05) and then just rotated everyone else with a hatchet... This sounds familiar. Poor, poor Dave (and David Luiz).
- Squads were smaller in the '90s. The Milan team that I could pull up the data was quite a small squad and all of them were old. Crazy.
- The Blackburn team was older than the rest of the teams that won the PL. Maybe team cohesion and experience was the fallback for the less talented side?
- Arsenal's 97-98 team was weird. They had an extremely young squad (only our team this year projects to have more Olympic players doing important things) but only had one of those players in the starting XI: Patrick Viera. Anelka was the 12th man that year.