Substitutions:

What statistics, history, and results have to say about Mike Petke’s substitution-based decision making.

(Photo: Joe Camporeale-USA TODAY Sports)


Remember how bad that Hans Backe guy was? Yeah, he was the worst. Well, maybe not the worst. I forgot Juan Carlos Osorio for a second there. Maybe my attempts to wipe that dreadful 2009 season from my memory are finally working! At least for a minute there. Anyway, back to the point.

An obsession with nordic players, inattentiveness to fans — or things important to fans such as the Open Cup — and poor cap management (although maybe that can be more attributable to Eric Soler) were among Backe’s many faults. But his most memorable deficiency was almost certainly that one tactical habit of his, that habit that would make Red Bulls fans scream at their televisions or gesture ridiculously toward the technical area at Red Bull Area: the failure, no, refusal to make substitutions.

Fittingly, Backe’s last pitfall as coach was that — in a game in which the Red Bulls’ center back had already been red carded, obviously warranting a switch— he did not make any substitutions whatsoever until the 89th minute, far too late to recover from Nick Deleon’s late, series-clinching strike, and thus ending both the Red Bulls’ season and the Backe Era.

Now, almost two years later, Red Bulls fans are left to ponder what has changed. No, the new signings aren’t nordic, but the glut of players coming from obscure clubs in Europe’s second divisions does evoke that same niche, questionable approach to scouting/team-building. Mike Petke is surely more of a fan favorite, but his comments following RBNY’s pitiful US Open Cup loss to the Cosmos mirrored Backe’s same disregard for the tournament. And, yes, our cap management, compared with the savvy of Real Salt Lake’s Garth Lagerwey or Toronto FC’s Tim Bezbatchenko, seems dismal to the casual, if restricted, onlooker. But this article is not here to address any of those worrying similarities between this and the prior regime (full disclaimer: I love Petke, I’d rather have him than Alex Ferguson, Jose Mourinho, or Pep Guardiola as the coach of my team). The worrying trend that I want to address here is the regression to Backe-esque substitution practices.

The most recent iteration of Red Bulls fans’ frustrations in this particular area came in this last Saturday’s 1–0 away loss to the Chicago Fire. Despite the absence of any meaningful scoring opportunities, Mike Petke did close to nothing — substitutions, tactical changes, expletive laden shouts from the sideline — to change the tide of the match until the 80th minute, when he inserted Bobby Convey. In all, Petke only used two of his available three substitutions, and only in the last ten minutes of the game when their impact could only be negligible.

This is consistent with his strategy all year long. In games in which the Red Bulls have been in a losing position in the second half, Petke has never used all three of his subs before the 84th minute. In a game against DC on the road, down a goal, Petke put on a sub in the 66th minute, but waited until the 84th and 90th minutes to make his next two changes. Against Toronto on the road, down a goal, Petke only made two of three available changes. Against Portland at home, down a goal, Petke elected to make one substitution before stoppage time. And against Philadelphia on the road, down two goals, Petke made all of his changes after the 80th minute.

So, is this approach valid? Petke’s explanation for his actions, courtesy of Empire of Soccer, following a listless home draw against a poor San Jose Earthquakes in which he made only one change, is fairly typical of his attitude toward using subs or even rotating his squad throughout the year: “Andre Akpan, I would love to bring on in a situation like this. I mean, we had the leading scorer up top that scored again tonight and had countless opportunities. Then I have Thierry Henry up top who can make a difference at any second … At the end of the day, it’s not that I don’t want to make a sub. I understand you and the fans might have [your] opinion, and I respect it, but at the end of the day, if it doesn’t feel like it’s going to make an impact and the way the game’s going, I’m not going to make a sub.”

It’s a fair approach. Given the choice between the now-departed Andre Akpan fresh off the bench or Thierry Henry after he had just run a marathon, I think most RBNY fans would opt for the later in a do-or-die situation. However, over the course of the season, running Thierry Henry into the ground isn’t going to do a lot of favors for the team. If Henry gets injured for the stretch run, RBNY can most likely kiss its playoff hopes goodbye.

However, just because Petke is disinclined from putting on a forward does not mean that he should make substitutions at all. According to a study (featured in the book The Numbers Game by Chris Anderson and David Sally) by sport scientist Chris Carling from the French club Lille, midfield substitutions carry a high degree of efficacy. These substitutions, according to Carling, “covered a greater overall distance and distance at high intensities and had a lower recovery time between high-intensity efforts compared to other midfield teammates who remained on the pitch.” In other words, fresh legs offer a significant, measurable benefit to teams, something most soccer fans didn’t need a study to reveal.

Where sports science data can provide significant insight is into the timing and use of all substitutions; in other words, is Petke right not to use all of his substitutes frequently? If not, at what point in the game should he make use of them? According to an extensive study conducted by Villanova University professor Bret Myers (once again featured in The Numbers Game), a regression analysis of substitution patterns versus game results yields the finding that, in a losing situation, coaches should make their first substitution before the 58th minute, the second before the 73rd, and the third before the 79th. In leagues across the world, coaches who follow this substitution pattern cut into or erase the deficit facing them in 38–47% of cases, whereas those that do not only manage to cut or erase the deficit in 17–24% of cases. Hence, there is a statistically significant correlation between using all three subs, relatively early in the game, and achieving results from losing positions.

Skeptics of this stat may cite a greater drop off in quality between the starters and substitutes in the case of RBNY or MLS in general versus that typically seen in the global game. And yet, the rule determined by Myers holds true in MLS, even though there is arguably a greater differential in talent between starters and bench players in MLS than there is in the Premier League or Bundesliga. In Myers’ data set from the 2012 season, from losing positions, about 40% of MLS coaches opted to use their first sub before 58th minute, second before the 73rd, and third before the 79th. Of these, 38% shrunk or erased the deficit from the time of the first substitute. However, coaches who did not follow this pattern only managed to achieve this feat 17% of the time.

It is quite understandable that Petke “doesn’t feel like it’s going to make an impact in the way the game’s going” if he makes a sub, but this data suggests that he is wrong. Moreover, going back to Lille’s sport scientist’s study mentioned previously, the decline in performance of a player who has been on the pitch for 80 minutes is not noticable to Petke’s naked eye. Their overall fitness and level in play shows almost no signs of detectable drop off, but in high-intensity situations (such as 50–50 balls or runs past a defender) players’ capacity to perform is ubiquitously diminished. Thus, relying on one’s judgment of player performance may not give a coach full sense of a player’s utility on the field until perhaps an attacking opportunity is wasted or an opposing player sprints past a tired player.

Beyond the data, ample evidence exists that Petke needs to alter his approach — whether via substitutions or other means — more frequently and earlier in games in which RBNY is chasing a result. Thus far, there have been three primary examples that surface when examining games in which RBNY has been trailing and Petke has made changes both to his personnel on the field and to his tactics when in losing positions. The first of these came in the 66th minute of a home game against the Chicago Fire, at which point the team was trailing 5–2. Petke made two substitutions, and moved from a 4–4–2 to an adventurous back three. RBNY managed to get two goals back and nearly pulled off a miracle comeback. Next, against Toronto FC at home, after going down 2–1, Petke inserted Bobby Convey very late in the game and switched to a 3–4–3 of sorts, in which Convey and Alexander sat deep lobbing balls in, Sam and Steele put in several crosses from the touchlines, and Henry, Cahill, and Wright-Phillips waited in the box for service. This brazenly route 1 approach yielded the tying goal in stoppage time, when a long ball from Alexander found Cahill’s head, which in turn found the feet of Wright-Phillips at the death. Finally, with his hand forced by injury to Tim Cahill and a red-card to Matt Miazga, Petke adopted a diamond midfield with one up top, with Alexander coming in for Cahill as a shuttler on the left side of the diamond, and his team came back from a goal down to win 2–1 despite playing with one less player.

In all three of these circumstances, Petke used his bench and a tactical adjustment in order to change the game. In all three of these circumstances, the team totally turned the game around, scoring a total of five goals and conceding none in approximately 75 minutes combined (the cumulative time after Petke made the substitutions and tactical adjustments until the end of each game). Now, this is not a blind endorsement of substitutions or tactical adjustments for their own sake, but evidence from around the world, Major League Soccer, and the present Red Bulls season points to the conclusion that more substitutions earlier in the game are necessary in order to improve results, especially from losing positions.

While thin on depth, the Red Bulls still possess ample options to come off the bench. Depending on the starting 11, Mike Petke can choose from players like Ambroise Oyongo — who offers speed and one vs. one ability — Peguy Luyindula — who can open up the game against a compact team defending a lead — or new signing Saer Sene— who, although he doesn’t offer enough from a counter-attacking or defensive standpoint to be considered an option to start on the wing in a 4–4–2, offers an attacking option in the last 15 minutes on either side or up top. In losing situations, Petke tends to stick to his guns; if he needs a result, he tends to view the same eleven players he chose at the start of the game as those most likely to do the job of changing the tide of a match. However, data, experience, and the attributes of Red Bulls’ bench players conclusively indicate that a more aggressive substitution strategy is necessary. If such an approach is not undertaken, fans should not be confident of coming from behind in the games that count as the Red Bulls look ahead to the group stage CCL, the stretch run of the regular season, and hopefully the MLS Cup playoffs.