Photo by Marjorie Elzy
Followers of the Union know who will be starting this weekend when the team return to the field against Atlanta United.
It will be exactly the same lineup as took the field May 30th against Portland, with maybe a Burke for a Santos.
Why? Well, the Union are known for their lack of squad rotation and Jim Curtin has stuck to his data-driven mentality on the negative effects of lineup change. Yesterday is today, as will be tomorrow in Unionville.
Fans of Blue fans worry the team can’t maintain this kind of stasis without a dramatic drop in points as the season goes on. The supporters demand squad rotation, “playing the kids,” or something of the sort.
Coach Curtin’s data says that’s not a great idea, but there is ample data to the contrary. Who’s right?
After reviewing both arguments, the overarching summary is this: Squad rotation doesn’t seem to matter, or more specifically rotating or not rotating doesn’t correlate with winning or losing.
Stories from the pitch
The Union almost always start with the same group of eleven players as they did the match before.
In 2021, they’ve averaged fewer than two changes per game. Removing the forced adaptations made because of Jose Martinez’s red card against NYCFC, the number is closer to one. They’re consistent, nearly top of the league, and the only MLS side remaining in the Concacaf Champion’s League.
But it’s early days and the calendar is about to get a lot more crowded, certainly they can’t keep this kind of pace up without some help. Should the team start rotating bodies? Should they be worried about winning if they do so?
To answer this question, first there are the anecdotes.
Last season Leicester made fewer changes than any side: won league. This season Chelsea have made fewest: top of league. Rotation sucks.
— Gary Lineker (@GaryLineker) January 4, 2017
Gary Lineker sees beauty in consistency.
Sir Alex Ferguson has the opposite opinion (and more than just one league trophy to show for it).
The former Manchester United manager “went 165 games from late 2008 until March 11, 2011 without naming an unchanged starting XI. In that time, he won two Premier League titles, the Champions League, the FIFA Club World Cup, the League Cup, made another Champions League final and went on to win the Premier League and make the Champions League final in 2011.”
In the parlance of the 1980’s movie “Space Balls,” if rotation sucks, then perhaps consistency blows.
Going even further, Pep Guardiola’s 2008 Barcelona side made an average of five changes per match while winning the first treble in Spanish history and the only sextuple in soccer history.
Fans may now refer to Pep as the David Bowie of soccer because he loves “Ch-ch-ch-ch-chaaayn-gessss!”
Author’s note: You’re welcome.
In the end, none of these anecdotes is very convincing in and of itself. Why? Because anecdotes are for people who can’t handle the truth. Just because your cousin’s uncle’s 2nd niece-twice-removed’s dog got hit by what the local paper thought was a meteor does not mean the sky is falling.
Data is the only way, and the data shall set fans free.
Numbers, graphs, and charts
https://sportsanalytics.berkeley.edu/articles/soccer-squad-rotation.html
The first data set comes from the world of higher education.
This set analyzes Bundesliga matches, the number of changes made per team per match during a season, and how that number correlated with overall points.
Author’s note: There are obviously a litany of confounding variables to this equation, but that’s partially the point.
The correlation is measured by a number call r-squared (r2), which has a value in this data set of 0.004350. An r2 of 1 means two things move perfectly together, like synchronized swimmers. An r2 of -1 means two things moved perfectly opposite of one another, like looking in a funhouse mirror. An r2 of 0 means the movements of two things had nothing to do with one another, like Chris Farley and Patrick Swayze’s respective Chippendale’s auditions.
0.004350 is pretty close to zero.
Bad things can happen in matches where a team performs well though, specifically players can get hurt. The thesis goes: there’s no way the same group of players can perform at their peak during 3-a-week’s starting Saturday and ending when the pandemic is over in 2026. That sounds like it makes sense anecdotally, but there must be data!
…there is, and like the best breakfasts and cheeses of the world, it comes from France.
Author’s note: Italy has some pretty incredible cheeses too, but there wasn’t any research on soccer substitutions.
A study of irregularly intensive periods of playing and training (“six games in eighteen” days kind of intensity) found that “[t]he total incidence of injury across the prolonged congested periods did not differ significantly to that reported in the non-congested periods.” There was some suggestion that players were injured more during the training between matches in congested periods, but that shouldn’t worry Union fans as Coach Curtin has already suggested a full calendar means a lot of matches and recovery with little time for anything else.
The anecdotes of players being injured amidst congested periods of play isn’t causal, it’s just a good story.
What does it all mean?
The Union are building toward a trip to The Azteca in late summer and another shot at the Supporter’s Shield in the fall. There are only a handful of teams that can prevent them from the latter, and one monumental squad standing in the path of the former. It might take an entire village of Union players to navigate the gauntlet that stands before them.
Or it might just take eleven or so.
It would be interesting to look at how squad depth impacts the rotation. Pep and Sir Alex probably have deeper teams than Leicester or The Union.
That’s a good metric, but I didn’t see anything that used it. You would have to know how fat the bell curve is in terms of value-per-player.
The second paragraph is incorrect, the lineup in Atlanta will have at least one significant change from that against Portland. Barring something really strange happening, Martinez will miss at least the three games in June due to the Copa America. And if Venezuela doesn’t finish last in its group, he’ll probably miss the first game in July as well.
You’re right, thank you for that. I’d still say the over/under on changes is 1.5.
What does it say about me that my first reaction to the caption on the home page and ‘Unionville’ was Why is PSP writing about the Ches-Mont League?
You lost me at ‘data set’.
1) The best cheese comes Italy, Chris. Come on! I thought you’d be with me on this.
.
2) Lineker’s stat on Leicester and Chelsea is interesting, but it also probably reflects that they suffered the fewest injuries and therefore needed fewer changes. Also, Leicester wasn’t playing European football that year, so they were juggling fewer competitions and therefore needed less squad rotation.
.
3) Anytime you can get a Spaceballs reference into a column, you do it. Brilliant.
.
4) Good, thoughtful piece. Despite my comment #2, I’m still thinking about this.
1) Yes.
I think Lineker’s stat misses out on one other piece of information. Neither Leicester or Chelsea were playing European football those seasons, so the only places they would have been playing midweek would of been the “Festive Fixtures” and later in the Cup competitions. Its easier to not rotate when you are only playing once a week.
A). Good piece, I found it intriguing. It’s a good look at how every side’s narrative has some gaping holes in the actual examples.
B). To add to Dan’s #2, better teams also have the luxury of keeping the lineup the same. If your team starts losing, you might have to make some changes, whereas a team that keeps winning would have to explain why they made a change. In summary, it’s plausible that winning causes low-rotation, rather than the inverse.
C). I did my own data analysis for PSP a few years ago in MLS matches and (to save you from my lousy writing) found that keeping the same lineup tended to result in more points, but medium and wholesale change carried roughly the same penalty. (https://phillysoccerpage.net/2018/03/09/change-part-2-analysis-does-stability-improve-results/).
*). Hilariously in retrospect, that research was done because a lot of fans thought Jim Curtin rotated too frequently, and they also wanted him fired.
It’s always been a depth issue. Until Ernst, management was, let’s nicely say, ‘challenged’ and ‘cheap’ (not counting Ernie Stewart who essentially used the Union to get his current job, helped find Ernst for Jay?)
In Ernst we trust. Simple as that.
Personally I think NOT having Bethlehem forces management to make the kids 1st level bench riders. They need playing time. All of them should be playing with Union2 or whatever.
But put that team in South Jersey. Management seems to ignore SJ, focused on EPA only. Big mistake. AC suburbs or at the old Garden State Racetrack. Perfect places for a mini-Subaru Park2