Revisiting playing time: how important is a stable lineup in Major League Soccer?

Two weeks ago, I published an article that analysed the distribution of playing time for Columbus Crew teams over the past 14 years. One of the apparent conclusions of that investigation was that, for a team to be successful, it is necessary to have a relatively stable core of players. The underlying charts identified that the more successful Crew teams had a group of 6-7 players who played at least 80% of the season, and a relatively small number of players who played more than bit roles, usually 15 players who appeared for at least 25% of the time.

This article was originally published at Examiner.com

The issue of how to distribute playing time, important in any season, will be particularly significant in 2010, with the Crew facing the prospect of competing in four competitions in a season that will also see some of the team’s best players potentially missing for six weeks around the World Cup.

In order to better understand this issue, then, I extended the playing time analysis in several ways.

The first step was to average the playing time curves between Crew teams that made the playoffs and teams that did not. While it is true that the playoff teams did indicate a larger core of players than those that missed out, the difference was not particularly significant. However, the playoff threshold has changed over the last 14 years as the league has expanded, meaning that the point total to qualify for the postseason in one year may not be enough in the future. The patterns become clearer when a 40-point threshhold is applied, rather than simply whether a team made the playoffs. Teams that earned more than 40 points in a season showed a more significant core of players, where teams that fell short of that mark had playing time that appears much more linear.

Playing time distribution of MLS teams, grouped by regular season points. The more points a team earned, the larger their core of players.
Playing time distribution of MLS teams, grouped by
regular season points. The more points a team earned,
the larger their core of players.

But what about other teams? Is the Crew’s experience with playing time an anomaly? To find out, the next step was to expand the analysis to include all MLS teams. At this point, I started looking only at seasons since 2004, as that was the year that the league did away with overtime, allowing tied games to stand after 90 minutes.

Over these past six seasons, then, it becomes clear that the Crew’s experience is not unique. 11 different MLS teams have earned at least 50 points in a season, and those teams have established a larger core of players than those which earned fewer than 50 points. Moving farther down the standings, teams that earned 40-49 points had a larger core of players than those that earned 30-39 points, which in turn had a larger core of players than those that earned under 30 points.

A few caveats

For as clear as this pattern is, however, several things remain uncertain. Primary among them, is the issue of causality. Are teams with a consistent core of players likely to earn more points, or are teams that earn more points more likely to turn to the same lineup, game after game? At the bottom of the table, it seems obvious that teams that aren’t winning are more likely to shuffle their lineup in search of success.

One possible lesson that might be drawn, is the importance of keeping faith with underperforming players. Robert Warzycha took considerable heat at the beginning of the 2009 season when the Crew opened without a win in their first seven games. However, he resisted the temptation to dramatically overhaul his lineup, and the team soon responded by climbing to the top of the standings.

While overall patterns are clear, individual years can vary widely. For example, the league-worst 2004 Fire followed a playing time curve very similar to the league-best 2009 Crew.
While overall patterns are clear, individual years can vary widely. For example, the league-worst 2004 Fire followed a playing time curve very similar to the league-best 2009 Crew.

Additionally, while the average playing time curves break down very nicely to support the argument that a larger stable core correlates with greater league success, there is still considerable variety between individual years. In fact, putting the playing time curves for the league-best and league-worst teams onto a single chart shows that these patterns sometimes overlap. Playing time on the 2004 Chicago Fire, who finished dead last with 33 points, actually shows a larger stable core of players than the 2009 Crew, who won the Supporters Shield with  49 points.

Another aspect the playing time question is how teams handle the demands of multiple competitions. All the investigation thus far has focused solely on league play, but since 2004 the league has expanded its international obligations with expansion of the Champions Cup into the Champions League, and the creation of the SuperLiga.

In 2008, the New England Revolution seemed to be tripped up by their participation in the SuperLiga, allowing the Crew them – including an emphatic 4-1 victory over a tired Revolution side that set both teams spinning in opposite directions. In 2009, by contrast, it appeared that the Crew juggernaut was derailed by the demands of the CONCACAF Champions League, seeing their record home winning streak end first against Cruz Azul before closing out the season with losses in their last two home league games, and one home playoff loss.

Disclaimer: Soccer, as a sport, has traditionally avoided the type of numeric analysis which is so prevalent in sports such as baseball or chess, where every action can be quantified, analyzed, and dissected. This does not mean, however, that analysis is fruitless. Some types of analysis, particularly those that look for larger trends, can be quite instructive. If you have any comments on whether you find this type of examination helpful, please leave a comment below.

Leave a Reply

Your email address will not be published. Required fields are marked *