Sounders reliance on veterans could be costly

"With age comes wisdom, but sometimes age comes alone." - typically attributed to Oscar Wilde -----

One year ago, I put together an analysis of recent MLS Superdraft history and first-team playing time. The article demonstrated (among other things) that draft picks tend to find more time on the field for some teams than for others, even when the impact of draft position is removed (i.e. some teams pick well - or at least pick players who see more time with the first team - even if they pick late):

"Some teams pick well" is a reasonable interpretation of the data, and that's the primary conclusion dealt with in the the original article - though I made some allusion to the alternatives. Opportunity matters, and observers tend to default to that narrative where convenient ("of course he found time for a bad team"). How a team manages its roster influences the perception of player-development success. Indeed, since developing players learn from encounters with higher levels of competition, successfully training a draftee, Homegrown Player, or any other youngster to contribute at the MLS level depends on such measures of playing time. Brooke Tunstall made a similar argument - albeit with a focus on luring players out of their college years - back in July on AmericanSoccerNow:

The thinking [motivating the formation of Project 40 and its successor, Generation Adidas]... was that with its short season and restrictions on training hours that college soccer wasn’t a good enough model to allow MLS and the U.S. national team to develop world class talent. Getting top players into a pro environment at an earlier age would enhance their skills at a key developmental stage.

Setting aside differences in target age, I agree that the construction of a playing environment is an important issue for improving the standard of play in MLS and U.S. Soccer in general.

How do we measure playing environment in MLS?

So, how do different MLS clubs perform with respect to finding playing time for different age groups? Previous articles on roster age have clearly established differences exist within the league, and have varied over MLS history due to changes in roster rules - noted by Elijah Miller in successive articles here

Over the years, looking at the MLS average age can explain quite a bit of the league’s history: consistent from 1996 to 2001, dropping in 2002 after contraction and rising upon expansion and the introduction of the Designated Player rule.

Now for the second straight season, the league’s average age is poised to drop as Home Grown players grow in influence and younger foreign stars are more easily signed.

The average age for the 2011 MLS season so far is 27.15 years, down from 27.36 in 2010 and 27.57 in 2009.

and here:

First things first: The big news is that with about one-seventh of the season [2012] complete, the league-wide average age is 27.94 years old. Last year’s ended up at 27.06, which means that it’s on pace for a 3.3 percent increase. That may not seem like a lot, but the average has never moved more than 1.3 percent in either direction in a single season in league history.

Indeed, after two years of getting younger, this season is on pace to surpass 1998’s 27.78 average as the oldest on record. Although, it’s likely that it will decrease by the year’s end, so it’s not a sure thing. The 2011 season’s final average ended up being about a 10th of a year lower than its first month. Some of this may be explained by the Olympic qualifying call-ups, but more than half of the MLS players that took part weren’t starting anyway.

How do things work out in 2014? Before I get into the numbers, I need to make a couple important points concerning methodology.

• Goalkeepers develop on a very different track than outfield players. They represent 9% of minutes played, but age the league overall by about 0.2 years on average. The difference with and without keepers is slight, but I prefer to exclude keepers altogether. All numbers presented below are outfield only.
• Roster age analyses typically use Whole numbers (age as of a specific date). For the sake of simplicity, I'm going to do this as well... but again, for simplicity, my date is a little bit odd. Age for the 2014 season is calculated as of 1/24/2015. If you wanted to compare my numbers to Elijah's above, subtracting about half a year might be appropriate.
• Not all players on a roster play. You will see me use the term "Roster Age" below, which is the mean age of all players on a roster who accumulated at least a minute of play during the season. Similarly, not all players on a roster play the same amount. Accordingly, I'll be adopting the phrase "Playing Age" for weighting players' contribution to the average age by their share of total minutes played. If you embodied every outfield player on a roster or in MLS over the course of a season in a single person, this number would be his age.

Finally, the data...

Team	Roster Age	Playing Age	Dif	PPG
New York Red Bulls	27.92	29.86	1.94	1.47
Seattle Sounders	27.86	29.02	1.16	1.88
San Jose Earthquakes	28.08	28.93	0.85	0.88
LA Galaxy	27.20	28.89	1.69	1.79
Real Salt Lake	26.35	28.40	2.05	1.65
DC United	26.56	27.95	1.39	1.74
Chicago Fire	27.29	27.90	0.61	1.06
Montreal Impact	26.39	27.75	1.36	0.82
Portland Timbers	27.77	27.65	-0.12	1.44
Houston Dynamo	26.81	27.56	0.75	1.15
Chivas USA	27.38	27.44	0.06	0.97
Vancouver Whitecaps	25.75	26.98	1.23	1.47
Philadelphia Union	26.32	26.83	0.51	1.24
Toronto FC	26.26	26.56	0.30	1.21
FC Dallas	25.79	26.53	0.74	1.59
Sporting KC	25.65	26.36	0.71	1.44
Columbus Crew	26.08	26.34	0.26	1.53
Colorado Rapids	25.67	26.08	0.41	0.94
New England Revolution	25.92	25.90	-0.02	1.62
MLS	26.68	27.52	0.84

The 2014 result does not indicate any kind of significant change in league composition in comparison to Elijah's 2011 and 2012 analyses (bearing in mind the methodological differences). If anything, the potential age increase observed partway into 2012 was not sustained.

The difference (dif) between Roster Age and Playing Age is an important measurement, as it assesses - to some minimal degree - a team's success in utilizing a full roster, young players inclusive. Notably, almost all MLS teams favor playing the older members of their roster (and this is really no surprise). Exceptions to the overall rule (say, with New England) can represent both young rosters and also a significant effort to integrate early-career players into the lineup. However, dif is also an easily skewed metric (e.g., RSL's heavy reliance upon the 35-year-old Javier Morales).

Several "win-now" ("retire-later") philosophy teams clearly rank among the league's oldest, with New York leading the field by a substantial amount (in fact, the 0.84 margin between NY and Seattle in 2nd is the largest disparity between adjacent ranks). Nevertheless, the relationship between Playing Age and success (measured by points per game) is weak.

Represented visually...

The positive trend is negligible and the fit to the data poor, plotting PPG vs. Playing Age. Though it may come as no surprise (particularly to New England fans)... aging a roster is not a reliable method of improving it. On the other hand, designing a roster after the fashion of 2014's Seattle Sounders or LA Galaxy, measured by single-year success, cannot be considered a definitive flaw in itself.

How do the Seattle Sounders handle roster management compared to their peers?

Averages are misleading (an admission you all are happy to hear, 1,100 words later!). It's reasonable to expect that a team's Playing Age (as well as the league's) should fall in the 26-29 years old range that roughly defines the developmental peak (the primary motivation is short term success). The age distribution should skew slightly young, because there is a competitive advantage to identifying similar performance at lesser cost (setting aside the broader motives I named earlier, as well as the fact that cost-for-value is not necessarily consistent by position or player origin). The distribution of the Sounders' playing minutes compared to the league in general is a good example of "Playing Age" telling an incomplete story.

Because all outliers deserve a name, let's call that diminutive high-minute green youngster on the left "DeAndre Yedlin." There were 6 under-25 Sounders on the roster to get minutes with the first team. 3 received less than 90 (Sean Okoli; Aaron Kovar; Tristan Bowen), 2 less than 700 (Andy Rose; Dylan Remick), and only 1 received time above the league trendline. Seattle did not simply have an old "average" player, but also found less time for young players than MLS teams in general.

In any specific year with any set of players, the question of usage patterns isn't separable from the people in question. It may be that choosing not to overexpose Okoli or Kovar to MLS play in their rookie seasons, for example, was the right decision (or that many of you reading this feel that way). Maintaining that bias towards on-peak players over multiple seasons is another matter. We have 6 years of data of Sounders' roster management, along with evidence of Sigi Schmid's behavior before he came to Seattle. Is 2014 an aberration, driven by circumstantial player choices?

Year	#players	Roster Age	Playing Age
2014	22	27.86	29.02
2013	26	27.38	28.30
2012	25	26.96	27.93
2011	25	26.40	27.14
2010	26	26.96	27.00
2009	19	26.16	26.39
2008	24	26.46	26.91
2007	26	25.54	27.33

Yes! I mean... no... I mean... kinda?

Using those wonderfully whole numbers, Sigi's team Playing Age of 29 is abnormal (though it should really come as no surprise that the Sounders' generic outfield player is, essentially, Brad Evans), but not particularly far beyond a solid 28 in the 2 years before. His generic player was 27 with Seattle in 2010/2011 and the same goes for his last two years in Columbus. The expansion 2009 Sounders, DeLoreaned 5 years into the future, would have been the 5th youngest team in the league. Sounders' management is clearly willing to commit minutes to young, unproven players just as they did with Yedlin in 2013. Seattle's roster management over the past 6 years, taken as a whole, falls in between LA Galaxy and Real Salt Lake over the same time period:

However, putting together an expansion team roster is a very different task (a task Seattle performed exceptionally well) than maintaining an MLS squad. The limitations of player acquisition, the resources available to put a team together, and the expectations placed on the club in 2009 were all unique for the managerial group including Schmid, Adrian Hanauer, and Chris Henderson. The Sounders of 2014 are the result of a (mostly successful) 6-year effort to modify the team to meet some lofty goals. The analysis taking all the years together is weighted heavily towards a young core towards a relatively young core group including Steve Zakuani and Fredy Montero. In the early stages of the Sounders' MLS existence, Evans, Osvaldo Alonso, and Jhon Kennedy Hurtado also received key first team roles at pre-peak ages. Applying some brute force to my spreadsheets, we can remove this bias in the data, and try to see how Seattle has modified the 2009 team in the subsequent 5 seasons.

Remove the performances of each player receiving playing time in 2009 from all subsequent years (keeping later years where the player was dealt away and then re-acquired, as with Alan Gordon). Treat the LA and RSL rosters in the same fashion. The result:

Team	#players	Roster Age	Playing Age
Seattle	76	26.38	27.74
Los Angeles	82	25.85	26.77
Salt Lake	67	24.37	24.88

The apparent similarity between the teams is heavily dependent upon the young core of '09. Seattle, LA, and RSL have exhibited different approaches to player acquisition in the subsequent 5 years. Taking this restricted data set, we can additionally remove all player performances above the age of 25, and sum up the appearances, starts, and minutes:

Team	#players	Apps	Starts	Minutes
Seattle	35	377	204	19577
Los Angeles	48	593	400	36339
Salt Lake	44	534	300	27703

Players 25 years old and younger acquired after 2009 have accounted for just over 12% of the 1660 possible starts for the Seattle Sounders in the period 2010-2014. 3.3% is DeAndre Yedlin.

So what?

The Seattle Sounders, since 2009, have been measurably bad at integrating new developing players into the 11 players taking the field. The team has been less effective in this particular aspect of player development than similarly successful (and, in some cases, more successful) teams. Low position in the Superdraft does not explain the trend. There are additional, anecdotal reasons to believe that Seattle's ability to identify MLS-capable youth players is strong - Servando Carrasco, Andy Rose, Alex Caskey, and Mike Fucito all represent strong value for the draft slot in which they were acquired. Managerial preference is the best remaining explanation for the data.

The pattern of aging the roster will likely need to change in the near future. Seattle was an old team in 2014, will be a year older, and its key lineup change of the offseason to date is the replacement of DeAndre Yedlin with either Tyrone Mears (32) or Brad Evans (30). Replacing the core roster with veterans in the manner of previous years would likely be expensive, between salary costs and potential transfer costs. The team may still be looking to its present set of developing players, particularly in the aftermath of an impressively active 2015 draft and the signing of 2 new homegrown players. Introducing Sounders 2 will impact player development tracks... but the team will have to avoid such pitfalls as those that impacted the reserve league (e.g. low levels of competition, positional inconsistency for developing players, lack of on-field interaction with first team players). Seattle has several players on the roster for whom time in USL was clearly a key component of professional growth (Alonso, Neagle, Zach Scott, and Michael Azira)... but actually taking the field in MLS is a vital step (which the first three examples clearly demonstrate). At some point Seattle needs to commit players to the first team. MLS regular season play provides ample opportunity to field developing players without significant negative impact on competitiveness.

Recruitment is a final, vital reason why increasing the usage rate of under 25s matters to Seattle. If the team hopes to be identified as a strong path from high school or college to MLS, it has to demonstrate those playing opportunities exist. If the team hopes to develop players to move on to a higher level of competition than MLS, it must demonstrate opportunities to play and excel within the league. What examples do prospective players like Jordan Morris see in Seattle? The 5 top success stories are Zakuani, Yedlin (nothing to complain about with these two, apart from Brian Mullan), Lamar Neagle (a player acquired on 3 separate occasions), Andy Rose (25 starts, 2596 minutes in 3 years to date), and Servando Carrasco (traded away for 386 less-effective minutes from Adam Moffat). Breaking into Seattle's first team is hard. This is not to suggest the Sounders mistreat prospects. If the team can't envision first team minutes for a player in Seattle, finding other teams where they may get that chance at a similar competitive level is the sign of a good organization. We can say certainly say this about Sean Okoli, Eriq Zavaleta, David Estrada, and Alex Caskey, and wish them luck. Estrada and Caskey arguably received adequate opportunities in the northwest, each logging single 1000+ minute seasons. Nevertheless, Seattle still stands to benefit from proving it can give equal chances to its MLS peers.