How important are substitutions?

[The Fish 10964]

Are Substitutions important? If so, how much?

 

http://www.telegraph.co.uk/sport/football/competitions/premier-league/11115075/How-important-are-substitutions-in-football.html

 

 

Sometimes you change something, and it goes wrong. In the 2004 Champions League semi-final, with the score Monaco 1-1 Chelsea, and Monaco a man down, Claudio Ranieri decided to go for the jugular. On went Jimmy-Floyd Hasselbaink for Mario Melchiot, then Scott Parker for Robert Huth, and in their total disarray Chelsea conceded two late goals. Back in the dressing room, one player angrily asked Ranieri: "What the f*** was all that about?"

 

Sometimes you change something, and it goes right. In the Rome derby in 2010, Ranieri's Roma were 1-0 down to Lazio. At half-time Ranieri, who had been in the job for a few months, did the unthinkable: he took off not just any two players, but Francesco Totti and Daniele de Rossi - two gods of Rome. Not a problem: Roma scored twice in the second half and won 2-1.

 

Sometimes you do nothing, and it goes right. Bayern Munich v Manchester United, in the 2010 Champions League. Franck Ribery came into the game carrying a slight knock, and as the second half began he was struggling. But with his side 1-0 down, Louis van Gaal decided he couldn't afford to bring him off. Ribery scored the equaliser as Bayern won 2-1 in the 92nd minute.

 

Sometimes you do nothing, and it goes wrong. May 2012, and Tottenham are fighting for a Champions League place. They need to beat Aston Villa away. As the minutes leak away, Harry Redknapp refuses to make a change, even though he has Jermain Defoe and Louis Saha on the bench. Finally, on 88 minutes, he twists. Off comes Rafael van der Vaart, and on comes… Scott Parker. The game finishes 1-1, Chelsea win the Champions League, and while the rest isn't history, Redknapp is.

 

The point is that when it comes to substitutions, there are no hard and fast rules. I was interviewing a manager in his office once, and noticed a piece of paper on the wall on which were written the optimum times to make substitutions: 59 minutes, 68 minutes and 77 minutes. Curiously enough, this sacred wisdom wasn't enough to save his job.

And yet, different managers have different reputations when it comes to making their subs. Sir Alex Ferguson and Louis van Gaal are seen as gamblers. Arsène Wenger and Brendan Rodgers are seen as fairly conservative, despite the intrepid style of their sides. Against Chelsea last season, Wenger declined to make a single change. "We had a good balance," he shrugged afterwards.

 

So which managers make the most substitutions and which the fewest? Which managers are more likely to make a triple-substitution at half-time, and which more likely to wait until the dying minutes? And what does this tell us about the styles of different managers?

 

Let's look at some numbers...

 

Number of substitutions

 

We studied every current Premier League manager's substitution record over the last season-and-a-bit, with the qualification that they had to have been in their current job for at least 20 games. That's a total of 16 managers taking charge of 826 games, making a total of 2250 substitutions. To increase the sample size, we considered all competitive games.

Unsurprisingly, most managers used a high proportion of their available substitutes. I mean, why not? But it's interesting to see which managers routinely use all three of their changes, and which don't.

 

 

Jose Mourinho pretty much always does. In fact, only twice last season did he fail to use all three subs, against Aston Villa and Paris Saint-Germain, and Chelsea lost both those games. Against Norwich, he changed the game by bringing on Willian and Eden Hazard just as he sensed the tide beginning to turn.

 

"During the game I am not nervous," he said. "I have feelings. I smell things, and I had a smell that they would score a goal. That is why I had Eden warming up at 1-0, because I smelled that."

 

Garry Monk is another manager who hardly ever wastes a substitution, in contrast to his predecessor Michael Laudrup, who left 14 per cent of his substitutions unused. At the other end of the scale, Brendan Rodgers and Paul Lambert are among the managers who use fewest substitutions.

 

Neither, though, can hold a candle to the ultimate champion of leaving things as they are: Sean Dyche. If Dyche puts you on the bench for Burnley, it's a good bet you won't be getting your boots dirty. The Ginger Mourinho left a staggering 30 per cent of his substitutions unused.

 

Now. Let's find out why any of this matters.

 

The effect of substitutes

 

A few months ago, Opta did some research into the effectiveness of the half-time substitution - click here. Looking at every World Cup game since 1998, they found that if you are losing by a single goal at half-time, making a substitution increases your chances of getting a win or draw from around 24 per cent to 40 per cent.

 

Then there is the research of Bret Myers, a professor from Pennsylvania, who devised the ideal time to use your three substitutions: 58 minutes, 73 minutes and 79 minutes. Myers claimed that if you follow his plan when you are losing a game, your odds of getting a draw or win are more than doubled.

 

Then there is the work of Daniel Altman at BSports - click here - and Colin Trainor at StatsBomb - click here - who found that Premier League substitutes score at a significantly higher rate than players who play the full 90 minutes, even when you allow for the fact that substitutes are more likely to be attacking players. Think of players like Javier Hernandez and Edin Dzeko, whose goals-per-minute rate is much higher when coming off the bench.

 

Altman found that fresh players coming on had a clear advantage when pitted against tired defences. Trainor found that not only did substitutes score at a higher rate than players who had played the full 90 minutes, but so did the players they replaced. The effects of fatigue on performance in the later stages of a game have been vastly understated.

 

Statistical and physiological evidence is increasingly converging upon a firm conclusion: there is really no excuse for a manager not to use all three of his substitutions. Not only do substitutions minimise the risk of injury to tired players, it also provides managers with the best chance of influencing the game in their favour.

 

If, that is, they're used early enough.

 

Time of substitutions

 

We've all screamed at a manager to make a substitution when things aren't going well. We've all rolled our eyes when the change is eventually made in the 91st minute. For a while, Emmanuel Adebayor's entire job at Manchester City was pretty much to come on as an injury-time substitute and jog around for about 100 seconds.

So, taking our 16 managers, we worked out the average time of their first substitution. We excluded all changes made in the first half-hour, which would have been almost exclusively forced by injury or an early red card.

If the evidence suggests that substitutes are a benefit, then clearly it makes sense to give them enough time to have an effect. And there's one Premier League manager who understands that perhaps better than anyone else.

 

Allardyce loves his stats. Perhaps he's even had some hi-grade version of this analysis in a ringbinder on his desk for years. Whatever it is, Allardyce is the king of the early change. His numbers last season were warped by a high number of first-half injuries, but even allowing for that, Big Sam makes a habit of throwing a curveball with his substitutions.

This next graph shows which managers lose their patience the quickest, and which wait longest to make their first change. If a manager didn't use any of his substitutions inside 90 minutes, we used a nominal value of 91. (And if you don't know what median means, just ignore it.)

 

Allardyce reigns supreme. There's the time he took off a bemused Carlton Cole after just half an hour against Everton. Then there's the double first-half substitution against Chelsea last season. In his disgust, Joe Cole disappeared straight down the tunnel and only turned up eight months later at Aston Villa. "Nobody wants to be substituted before half-time," Allardyce explained. "Purely from a tactical point of view, I had to change something, the way the game was at that stage. That's my job."

 

Once more, Dyche proves himself among the Premier League's most conservative managers when it comes to making substitutions. But you'll notice that Mourinho and Monk, the two managers who use the most substitutions, are at different ends of this graph. Monk uses all his subs, and uses them early. Mourinho uses all his subs, and uses them late. The distinction is explained by a factor we haven't yet considered: what the score is when a substitution is made.

 

Game state

 

The next graph shows what different managers do when their team is in the lead, and when their team is level or behind. In theory, a manager that's drawing or losing will be making substitutions earlier, in an attempt to change the direction of the game. And so it proved, across the board.

 

Along with Allardyce and Redknapp, Manuel Pellegrini is one of the most proactive managers when it comes to making substitutes, whatever the state of the game. At the other end of the scale, Roberto Martinez and Wenger are among the more conservative, leaving their first change an hour on average, even when their side is losing or drawing.

 

The timing of Mourinho's substitutions, on the other hand, is overwhelmingly dependent on the state of the match. If Chelsea are chasing the game, he is one of the quickest to change things around. If Chelsea are winning, he is one of the last to act. And this tallies with the view of Mourinho the arch-pragmatist. "Substitutions are a consequence of how you read the match," he has said. "You can't predict them before the game, because it is unpredictable."

 

Some other conclusions that the data threw up, but are too boring to put on a graph:

1) On average, Gus Poyet makes his third and final change earlier than anyone else - on 79 minutes. No prizes for guessing that Sean Dyche leaves it latest: 85 minutes.

2) Poyet also leads the way when it comes to double substitutions. A third of his games feature the eye-catching "double substitution", compared with just five per cent of Monk's.

3) The manager most likely to make a change at half-time? Harry Redknapp. He does it 35 per cent of the time. Next comes Alan Pardew on 30 per cent. At the other end of the scale, Dyche makes a half-time substitution five per cent of the time, followed by Wenger on 11 per cent.

 

What does any of this tell us?

 

First of all, this isn't about picking out "good managers" and "bad managers". Also, there are a number of factors that determine whether and when a manager decides to make changes. The quality of substitutes available, for instance. In the dying minutes of Chelsea's win at Anfield last season, as Liverpool threw everything in search of a equaliser, many wondered why Brendan Rodgers didn't use of all of his substitutes.

 

Then again, his options - Luis Alberto, Iago Aspas, Aly Cissokho - didn't exactly scream "cutting edge". Mourinho, meanwhile, was able to bring on Gary Cahill, Fernando Torres and Willian, who scored to wrap up a 2-0 win.

 

But maybe there's something a bit more personal to it, as well. There's so much more to management than tactics or philosophy. Risk aversion is the idea that changing something and failing is worse than failing when you haven't changed anything. Making an early change, or using all three substitutes, might be seen by some managers as an admission of failure, that they got their original starting line-up wrong. Better, surely, to give the lad five more minutes and see if he comes good.

 

And yet the evidence is there. Managers who don't use all their substitutions, or who use them too late, appear to be giving up a significant advantage as a result. Maybe a manager who sticks with his original XI for just that little bit longer is possessed of a steadfast belief, not just in his players, but in himself.

 

 

 

 

[ClubSpinDoctor 0]

Important, especially when you've got as many options as we apparently do, according to SOME on here

[toonotl 3110]

Very interesting article.

[luckyluke 2]

It is but one of the most influential factors is going to be who you've got on the pitch, who the subs are and who you're playing. I'd have thought they make a bigger impact on what subs you make.

[Park Life 71]

Personally think subs need cutting back to one allowed (outfield) and a goalkeeper. Would cut back on the advantage of the rich clubs sitting there with two fekin first teams...

[Noelie 103]

Substitutes are very important, but it's a question of knowing when, and who, to bring on.

During Pardews tenure he hasn't shown much of a clue of the right time and the right player.

His consistency in playing players out of position adds to his cluelessness.

[The Fish 10964]

That's something I'd like to see, who does Pardew habitually take off and who does he bring on and when? Is it different to the league standard and which subs seems to have the biggest impact on games (strikers, midfielders, etc.)?

[ewerk 31201]

It's all about the left backs.

[Meenzer 15716]

It's all about the left backs.

 

It's all about the dum dum da da dum dum(mett).

[Park Life 71]

The earlier you do them the more panicky it looks.

[The Fish 10964]

The earlier you do them the more panicky it looks.

Rodgers' first half change against West Ham was essentially saying to himself

[David Kelly 1260]

It's no surprise that the lowest three managers in terms of the percentage of available substitutions are Dyche, Lambert and Rodgers. Burnley and Villa have small squads with quite a drop in quality from those players who play regularly and those who dont. Liverpool have a better squad but when you're trying to influence an attacking outcome (which what I believe most tactical substitutions are) they don't have players on the bench anywhere near the quality of Suarez, Sturrige, Sterling and Gerrard. Since those players stayed fit most of the time last season you wouldn't change them, you'd just hope they could pull something out of the bag.

 

With regards to Pardew he takes off Goufrran or a leftback in the majority of his substitutions (and Riviere now) but he rarely takes the chance of bringing on someone who is plays a more attacking role. Saturday showed this clearly. We were getting beat 2-0 and he takes off our only striker and brings on a new one (he did exactly the same against Citeh although later iirc). He got lucky partially because Bruce made a cock up with his substitution taking off Jelavic who was holding the ball up well and pushing us back. That allowed us to move the ball forward easier. He also got lucky that Cisse actually started hitting the target and McGreggor dropped a bollock for the first goal.

[trooper 940]

Pardew showed a masterclass in substitutions last week by taking off Sissoko. Then replacing him with Sammy Ameobi in the 89th minute when we were 0-4 down against Southampton pure genius.