Three ways Chelsea let two set pieces decide the tie against Man United

3. Trebly redundant midfield

Any two of Chelsea’s three midfielders would have been better off paired with a different third midfielder, particularly an attacking midfielder. Jorginho, Mateo Kovacic and Billy Gilmour are too similar in their preferred area of the pitch and overall shape of their game to be a functional unit. They were redundant and left gaps none of them could fill.

The first sign of things going awry was seeing Gilmour spending so much time near Chelsea’s forward line.

Of the three, he seems most suited and most likely to play deep. Gilmour makes accurate, well-struck, long-range passes; tackles aggressively, welcoming the physical challenges despite (because of?) the usual size differences; and reads the play well enough to preempt the opposition’s intent. In a somewhat stagnant game in the opponent’s half, Gilmour is not going to contribute much and would be better off protecting the back in case of counter-attack

Kovacic is proving himself most adept in transition in both directions. Like N’Golo Kante, he can tackle the ball off an opponent and within two steps be dribbling away at top speed. From there he can either power the ball through space or pass to a full-back or attacking midfielder already ahead of him. He is not as effective when dribbling at a defensive block, and does not make the passes that will open up a tight defence.

Of the three, Jorginho could possibly do the most as the most forward midfielder, if only to open up as a recycling option, particularly in Willian’s absence. His trademark short no-look passes could be beneficial in this situation as he could keep the play tightly in the final third, and his teammates were almost always in the same places from one sequence to the next. Instead, Jorginho remained the deepest of the three for the first half.

Only Jorginho, then, was in his preferred position, but none of the three were in their best position for this unit and this opposition. The number of times Jorginho observed players run past him with the ball testified to how much better Gilmour would have been both in reading the play and doing something about it.

In the second half, Frank Lampard ameliorated some of this by having Kovacic and Gilmour play more as a double pivot in transition. More than anything, though, it was simply Gilmour’s increased pace and power in passing up the middle that allowed Chelsea to move against Manchester United while they were in defensive transition. It was the first variation in Chelsea’s passing in the game, and gave Gilmour the best justification for being the deepest midfielder (his defensive acumen and aggression were close in second).

For a 4-3-3, Chelsea needed at least one midfielder making attacking runs and directly and closely supporting the offence. Any two of these three would have been a reasonable pair in a double-pivot behind a three-man attacking midfield line (Jorginho and Gilmour would be the least preferable combination). But Lampard’s need to rotate precluded the 4-2-3-1.

Frank Lampard needed to rotate his squad for this game, and should against Watford as well, despite the result on Wednesday.

He has many players seriously in need of rest, especially with Ajax on Tuesday. He has enough players to try a variety of set-ups, especially now that he knows more about what works, what doesn’t and which players need the extra instruction.