Average age of MLS players

[mpruitt]

Originally posted by ChrisE
I realize that I didn't really provide as much information as I should have. Because those numbers are focused specifically on the US nats pool (and spec. the US players who got 500+ minutes), it's not exactly possible to interpret trends about MLS in general using them. So, here's a little bit about how the Nats pool compares to the total MLS population:


Year lg min nat min >500 min US % >500 US%
1996 315981 215559 205413 0.682 0.650
1997 316037 218495 211293 0.691 0.669
1998 378028 251416 240246 0.665 0.636
1999 378670 259794 248453 0.686 0.656
2000 387215 273969 265458 0.708 0.686
2001 318993 221996 214674 0.696 0.673
2002 282540 199340 193148 0.706 0.684
2003 306758 222986 215486 0.727 0.702


Looks like we're trending toward being awfully jingoistic. Good for us.

[ChrisE]

Originally posted by maxim-1
Looks like we're trending toward being awfully jingoistic. Good for us.

Yeah, I'd been meaning to start a thread about that.

[kenntomasch]

With serious thanks to ChrisE, I now have birthdates for 472 of the 476 players who played in the A-League in 2003.

The average age of those 472 players as of July 1, 2003 was 25.39.

Weighing the age for playing time, the average was 26.01.

Makes sense. I would imagine the second level would be slightly younger than the one above it. And even if the PDL wasn't by mandate predominantly a U23 league, that average should be even lower. If anyone wants to try their hand at that study, be my guess. The PDL average is probably between 20 and 21. Just a guess.

Number of players of each age:


Age...#
18....2
19...13
20...10
21...19
22...63
23...59
24...54
25...53
26...56
27...31
28...16
29...24
30...19
31...14
32...18
33....5
34....9
35....4
36....3


The big clump was from 22-26 and then there was a significant drop. It makes sense that there would be a lot of 22-year-olds - that's usually the age at which you graduate from college.

This is some valuable information that I don't think anyone has ever put together before - thanks to Chris and Phil for their work on this and the MLS age study. I just finished the numbers for the 2004 A-League season.

With a reduction of the A-League from 19 teams to 16, there were (not surprisingly) fewer players in the A-League in 2004. 476 players played at least one minute in 2003 (25.1 players per team), and 386 played at least one minute in 2004 (24.1 players per team).

(The interesting thing is that, as near as I can tell, only about half---200 of the 386---played in the A-League in 2003 as well as 2004. I have to explore that a little more deeply, but it's interesting.)

Anyway, the average age of the 386 players on July 1, 2004 was 26.3. Weighted for playing time, it was 26.7. Both figures are slightly higher than last year.

You see the list above of the number of players of each age in the A-League in 2003. Here is the 2004 list, with percentage of the total:

Age...#....Pct.
18....4...1.0%
19....6...1.6%
20....7...1.8%
21...12...3.1%
22...21...5.4%
23...55..14.2%
24...41..10.6%
25...42..10.9%
26...45..11.7%
27...31...8.0%
28...24...6.2%
29...17...4.4%
30...19...4.9%
31...13...3.4%
32...13...3.4%
33...12...3.1%
34...11...2.8%
35....7...1.8%
36....3...0.8%
37....1...0.3%
38....2...0.5%
Tot.386


Another way of looking at this is at the number of players in certain age ranges from 2003 to 2004:

..........2004..........2003....
Age......#....Pct......#....Pct.
<22.....29....7.5%....44....9.3%
22-26..204...52.8%...285...60.4%
27-31..104...26.9%...104...22.0%
32+.....49...12.7%....39....8.3%
Total..386..100.0%...472..100.0%


The percentage of players age 22-26 was up, and the percentage of players age 32 and over was down.

I have to explore this further and see what's happening as far as what percentage of each age played in the league both years (and simply aged a year) and who was pushed up and out.

Comments?

[kenntomasch]

Actually, I've just discovered there are some rounding errors there, folks, so disregard that for the moment while I re-do the numbers using Absolute Values.

[ChrisE]

I actually was doing this last night, too, kenn (though I've still got a 3 or 4 unknown birthdates).

These are, by my calculations, the ages (using your 7/1 cutoff) for the last three A-league seasons:


2002 26.27
2003 26.51
2004 26.69


We've seen a pretty steady escalation which, in my opinion, is a product of the increasing stability of the league. The average age for the 13 teams that played both in 2003 and 2004 actually fell - 27.0 in 2003, 26.8 in 2004. The real change came in the loss of six teams - Calgary Storm, Charlotte, Cincinnatti, El Paso, Indiana, Pittsburgh - only one of whose average ages were above 27.0 (Charlotte, 27.1), and who in total had an average age of 25.4. These teams were replaced by three expansion teams, PR, Edmonton, and Calgary Mustangs, whose average age was 26.1, and obviously the more established teams bore more of the burden.

[kenntomasch]

I have all the birthdates, Chris. Let me know who you need, and I'll supply it since you've done the research more thoroughly than me.

[numerista]

I have all the birthdates, Chris. Let me know who you need, and I'll supply it since you've done the research more thoroughly than me.

Hey Kenn,

Since you're volunteering, here are a few that would help fill out my records (all I keep are birthyears).
Paul Marcoullier
Randy Merkel
Michael Richardson

Thx.

[ChrisE]

Hey Kenn,

Since you're volunteering, here are a few that would help fill out my records (all I keep are birthyears).
Paul Marcoullier
Randy Merkel
Michael Richardson

Thx.

I have Merkel as 3/20/76, and Michael Richardson as 10/8/68 numerista. I guess the website has foundered, but I think it would be a good place to locate a comprehensive list of US-based soccer birthdates.

[ChrisE]

I have all the birthdates, Chris. Let me know who you need, and I'll supply it since you've done the research more thoroughly than me.


I certainly wouldn't say that, kenn. If you like, I'll obviously send you what I've got (I don't think anyone has ever actually taken me up on that). I'm also trying to get nationality information, which was pretty simple for 2004, but gets a lot harder further back. Anyway, players I'm missing:

Domingo Gomes
Carmalo Rago

I guess that's it for ages, although I'm missing several nationalities. If you're willing, I'd like to see whatever database you're using.

[numerista]

Thx Chris,

While we're on this subject, I find it interesting that starting in the birthyear 1976, there seems to have been considerable fluctuation in the number of American MLSers per year. 1976 and 1977 were high, presumably spurred by expansion, and then 1978 and 1979 were low, which may suggest that opportunities were dwindling even before 2002 contraction ... was there a reduction in roster size in 2001? Since then, things seem to be on the rise, indicating that the Development Player program is doing its job. (Sorry I don't have data handy.)

Viewed in conjunction with the Transitional International program, it appears that the league's front office has developed some good strategies for growing its talent pool.

[kenntomasch]

I discovered that "Carmalo Rago" is actually "Carmelo Rago", and he has a bio on canadasoccer.com. He's like 19 or something. Gomes I'm sure I have because I have everybody, but it's on another computer.

It's just Excel, my best buddy. Nationality exists in better form for 2004, you're right (but not for everybody), but going back, good luck.

My sense is that there are a heck of a lot more foreign players in the A-League than in MLS. Even percentage wise, given that the A-League is almost twice as large as MLS. I can't remember how many we could have - five visa players? Something like that. But I know there are a bunch of Eastern Europeans in the A-League.

[ChrisE]

I discovered that "Carmalo Rago" is actually "Carmelo Rago", and he has a bio on canadasoccer.com. He's like 19 or something. Gomes I'm sure I have because I have everybody, but it's on another computer.

It's just Excel, my best buddy. Nationality exists in better form for 2004, you're right (but not for everybody), but going back, good luck.

I don't care what it's in kenn, I thought it would be useful to compare your numbers to mine and to see if there was anything that one of us was doing that the other isn't. Everything I've done is in Excel, I didn't learn there were other database programs until a couple of months ago, and I wouldn't have an idea how to use them.


My sense is that there are a heck of a lot more foreign players in the A-League than in MLS. Even percentage wise, given that the A-League is almost twice as large as MLS. I can't remember how many we could have - five visa players? Something like that. But I know there are a bunch of Eastern Europeans in the A-League.

This is a difficult thing to analyze, considering the A-league's presence in Canada and Puerto Rico - how should we count these players? Are Americans foreigners in Canada, etc. - like the age situation, there are a lot more extenuating circumstances that make MLS look damned orderly by comparison.

Here's the minutes played distributions from 2004 - I divided it according to FIFA regions, but excluded Mexico and Argentina/Brazil because I think they're unique situations, and Canada and Puerto Rico because they have teams in the league (and grouped them with the US in the last group):


USA 47.7
CONCACAF 6.0
CONMEBOL 2.2
UEFA 8.4
Africa 3.6
Asia 0.1
Oceania 0.7
Arg. & Brazil 5.5
Mexico 0.9
Canada 22.5
Puerto Rico 2.4

A Lg. Nations 72.6



The US numbers are obviously way lower than MLS (though above the 10/16*72.6= 45.4 that would be expected by team proportions), but those numbers are actually surprisingly close to the 72.7% of minutes played by Americans in MLS last year.


These are the ratios and raw numbers of actual players by nationality (who garnered at least 1 minute):


USA 174 44.8
CONCACAF 24 6.2
CONMEBOL 9 2.3
UEFA 34 8.8
Africa 18 4.6
Asia 2 0.5
Oceania 3 0.8
Arg. & Brazil 21 5.4
Mexico 2 0.5
Canada 90 23.2
Puerto Rico 11 2.8


The Americans garnered more minutes than would have been expected by their number of players (extending a trend that has been growing in MLS as well), Byron Alvarez and Hugo Alcaraz made Mexico hugely overrepresented, Argentina and Brazil were slightly better represented - all other regions got fewer minutes than should have been expected, with Puerto Rico and Africa being the most significant underachievers.

[kenntomasch]

I certainly wouldn't say that, kenn. If you like, I'll obviously send you what I've got (I don't think anyone has ever actually taken me up on that). I'm also trying to get nationality information, which was pretty simple for 2004, but gets a lot harder further back. Anyway, players I'm missing:

Domingo Gomes


I show Gomes as 1/1/80, but I show four players as 1/1/80 and I wonder if that's actually true or just a placeholder that means "I don't have his birthdate right now, but he was born in 1980."

Anyway, Chris, I emailed you because I'm having trouble getting Excel to calculate ages correctly. It's rounding up, causing a guy born in August of 1979 to be 25 on July 1, 2004, and that's not right. And I can't figure it out short of just going through and manually changing ones that fall between certain dates. Which defeats the purpose of Excel.

[kenntomasch]

Okay, I sorted the age thing by rounding down.

I thought I had everybody's DOBs for 2003, but I don't have these guys':

Carrasco, Javier (El Paso)
Finklea, Timothy (Cincinnati)
Narita, Junro (El Paso)
Thomas, Keegan (Calgary)

Without those guys, I show the median age in 2003 as 23. Is that right?

I show 2003's non-minutes-weighted average as 25.40, and the weighted average as 26.02.

I show the median age in 2004 as 25.

I now show 2004's non-minutes-weighted average as 25.76 and the weighted average as 26.17.

Chris, I've emailed you my spreadsheet to compare with yours.

[ChrisE]

Without those guys, I show the median age in 2003 as 23. Is that right?

I'm not sure what you're doing here, kenn - I'm using your speadsheet, and get a median of 25 years (more specifically, 25.3 years) - I'm not sure this is indicative of much, however, since the median is heavily skewed by the fact that Cincinatti fielded a dozen or more 22-year old recent graduates when their team mutinied against them.




I show 2003's non-minutes-weighted average as 25.40, and the weighted average as 26.02.

I disagree with rounding ages in this case, kenn - by rounding ages down, you cut about .5 years off the average age (because a player who is 25.9, e.g., is counted as a 25 year old) - I don't see any reason to round to the nearest year, except when counting how many players there are of age x.




I show the median age in 2004 as 25.

I now show 2004's non-minutes-weighted average as 25.76 and the weighted average as 26.17.

For these, I got the same thing plus .5 years - median 25.65, non-weighted average 26.26.

[ChrisE]

I show Gomes as 1/1/80, but I show four players as 1/1/80 and I wonder if that's actually true or just a placeholder that means "I don't have his birthdate right now, but he was born in 1980."


I thought the same thing - I think odds are pretty slim that four players, all relatively obscure, had such a conspicuous birthdate - and, considering that there weren't any 1/1/82's or 1/1/79's, I'm worried that that is just the default birthdate, and that the 80 is no more meaningful than the 1 or the 1.

[kenntomasch]

I'm not sure what you're doing here, kenn - I'm using your speadsheet, and get a median of 25 years (more specifically, 25.3 years) - I'm not sure this is indicative of much, however, since the median is heavily skewed by the fact that Cincinatti fielded a dozen or more 22-year old recent graduates when their team mutinied against them.

You're right. 25. I divided 472 by 2 and got 286 instead of 236. The median for 2003 is 25.


I disagree with rounding ages in this case, kenn - by rounding ages down, you cut about .5 years off the average age (because a player who is 25.9, e.g., is counted as a 25 year old) - I don't see any reason to round to the nearest year, except when counting how many players there are of age x.

I can see that, I guess. But I was dealing with ages as an absolute. If we're going to start getting into half-ages, it's just as abritrary as picking July 1 for the reference date. A player who's 25 and three days and a player who's 25 and 360 days are both 25. Too many decimals might or might not be too high a level of implied accuracy for the broad generalizations we're trying to make here.


For these, I got the same thing plus .5 years - median 25.65, non-weighted average 26.26.

I'm not so much concerned with what the numbers are as what the numbers mean. Is the A-League more or less a league of young kids than MLS? Is the league getting older or younger? What are the effects of contraction on the A-League - do older or younger players get forced out?

Now that A-League stats are a little more thorough (there've been years where they couldn't tell that something was a typo, and they created a whole new player based on the typo, when it was obvious which player the stats belonged to) we can do a bit more in terms of having a complete record of an A-League season. We can also see which players move in and out of the league, on their way up or their way down the ladder. We can see the effects of MLS expansion and where those guys come from. We can see what happens to MLS draft picks who don't make it.

This is good information to have.

[ChrisE]

You're right. 25. I divided 472 by 2 and got 286 instead of 236. The median for 2003 is 25.

(You know, Excel has a median function )



I can see that, I guess. But I was dealing with ages as an absolute. If we're going to start getting into half-ages, it's just as abritrary as picking July 1 for the reference date. A player who's 25 and three days and a player who's 25 and 360 days are both 25. Too many decimals might or might not be too high a level of implied accuracy for the broad generalizations we're trying to make here.

Here's my problem with your example. If you're measuring from July 1, and have a player who is 25 and three days and a player who is 25 and 360 days, you would measure them as the 25-25=0, the same age. However, if someone else chose to measure (arbitrarily) from August 1, he'd measure them as 26-25=1 year apart. Whereas, if you don't round, they're 357 days apart no matter when you measure from. Although, yeah, we don't need to be measuring ages to the thousandth, I think it's better to calculate using the raw numbers, and only then round, rather than use rounded numbers from the start. Your average age of 25.4 implies, to me, that the average player would be 25 and 3 and a third months - if you're rounding before you do these calculations, this simply isn't true.




I'm not so much concerned with what the numbers [i]are as what the numbers mean. Is the A-League more or less a league of young kids than MLS? Is the league getting older or younger? What are the effects of contraction on the A-League - do older or younger players get forced out?


I agree completely - what's important isn't that the A-league's average age is 24.853, it's that it fell (or rose) .3 years from last year, etc. However, I think it's necessary for all of us to use the same method when measuring these things, and I think (of course) that it's better to use as much information as we have until we actually get down to interpreting the statistics.



Now that A-League stats are a little more thorough (there've been years where they couldn't tell that something was a typo, and they created a whole new player based on the typo, when it was obvious which player the stats belonged to) we can do a bit more in terms of having a complete record of an A-League season. We can also see which players move in and out of the league, on their way up or their way down the ladder. We can see the effects of MLS expansion and where those guys come from. We can see what happens to MLS draft picks who don't make it.

This is good information to have.

Well I'm glad I wasn't around for the bad old days. Agree completely with the rest (although I'd be really interested to see anyone draw conclusions from the transfers in/out of the A-league).

[ChrisE]

Another year, another drop in average age for MLS. This makes four straight years that the league has seen its average (minutes-played) age drop, after four years of relative stability. Furthermore, while ages became more and more spread out each of the first 5 seasons, this trend seems to have tailed off in recent years, perhaps because of competition with ages becoming more compressed around the 22-26 range.


Ages st. dev
1996 27.20 3.43
1997 27.44 3.78
1998 27.52 4.10
1999 27.44 4.11
2000 27.48 4.36
2001 27.35 4.57
2002 26.98 4.34
2003 26.72 4.14
2004 26.43 4.28

[kenntomasch]

And I think we know intuitively that the league is getting younger. The first wave of guys in the league may have included a bunch of guys who had been holding on, waiting for a Division I league to get started, along with the inevitable aging foreign players in for a holiday in the States (that's just a guess, I don't have the breakdowns).

It seems like we're seeing more and more young Americans now...really young Americans (besides Mr. Adu), which is dropping the average.

Good stuff, Chris.