Home » Tactics » A data-driven approach to squad usage

A data-driven approach to squad usage

Sports aren’t static. Big changes happen, sometimes suddenly and other times gradually – but either way, they happen. Baseball’s starting pitchers used to go nine innings routinely. Basketball players didn’t dunk. Ice hockey teams had never heard of line changes. And soccer has a long history of tactical innovations, to which this post might just contribute.

First, let me say that I’m not a coach. I’ve played thousands of hours of soccer, but I’ve never coached a single match. I also have a background that’s very different from most coaches in the world’s top leagues. So what I recommend here may be naive for one of several reasons: clubs may already be doing it, clubs may have tried it and rejected it, or clubs may find it impossible to implement.

On the other hand, sometimes a fresh pair of eyes helps. Unlike most coaches, I’ve spent thousands of hours examining detailed data from soccer matches and validating what I find using video, and vice versa. That kind of research teaches you a few things about what works and what doesn’t. Here are a few things I think I’ve learned:

  • Teams with stable lineups do better than those with heavy rotation, adjusting for the intensity of schedules
  • Teams that play one system well do better than those that switch systems all the time
  • The best way to kill a game off is by keeping possession and using up the clock
  • Fresh attackers have an advantage in scoring in the second halves of matches

Viewed with an open mind, these simple truths – not always true, of course, but with a tendency to be true – suggest a variety of potential innovations in coaching. If I took control of a team right now, I’d have little idea of how to train them and manage them man to man. But here are a few things I would do, subject to the constraints of available resources:

1. Use data and video from previous seasons and pre-season matches to assess the strengths and weaknesses of all players, including who combines well with whom.

2. Choose an ideal XI to play an ideal formation, then make these players do everything together: train, eat, sleep, play Fifa, you name it. Each player should understand every other player’s subtlest visual, verbal, and physical cues. They should know each other almost as well as they know their families. They should sit down for meals at a table shaped like their formation on the field, in all of their respective positions. Their interactions and interconnections should become second nature. The XI would have its own assistant coach as well.

3. Create four special teams. The first pair of teams would be for an attacking blitz: three forwards (strikers, attacking midfielders, or wingers) each. These teams would share an assistant coach. The second pair would be for killing a game off. One would have midfielders adept at dribbling and keeping the ball, for games where the team was able to hold possession. The other would have defenders and defensive midfielders, for parking the bus or elevating the press in games where the team had a hard time keeping the ball. Each would have its own assistant coach. These players would do everything together, too. They would sit on the bench together during matches with their assistant coaches, discussing how to play should they come on.

4. The attacking blitz special teams would be used relatively early in matches when the team was losing, entering on the hour mark at the latest and possibly as early as the start of the second half. (Think of Ezequiel Lavezzi in the 2014 World Cup final – 45 minutes of total commitment, compressing 90 minutes of energy into one half, leaving everything on the field.) The game-killing teams would enter when the team was winning, perhaps a little later. (Think of what Argentina should have done, 0-1 up against Germany in 2006.) During draws, the strategy would depend on the coach’s judgment.

5. When the whole team got together to watch tape, afterwards it would break into the same subgroups, each with their assistant coach and an analyst as a guide, to discuss what they saw.

6. Training matches would pit the subs against the first XI. Then, just as in real games, the special teams would sub into the first XI: attacking blitzes to simulate when the first XI was losing, and game-killers to simulate when the first XI was winning. Each special team substitution would result in taking off the front three, who generally tire the fastest. The front three would be replaced ahead of the other eight players by the incoming trio, except for the game-killers whose job was to park the bus. In their case, the new players would slot in between the defenders and the midfielders as an extra line, or behind the defenders in order to maintain contact between the seven starters.

7. In case of injuries during the season, subs would be promoted into the special teams and first XI. In case of underperformance, data, video, and interviews with the players would be used to identify and solve the problem. If the problem was a player, a sub would be promoted. If a change in the team’s tactical system was needed, it would take place during the longest possible layoff between matches. Subs could also challenge players in the first XI for their places through a structured process involving tryouts with the rest of the starting team.

To me, teams too often expect one or at best two players to change a game. Very few players can do this, and even the best ones struggle to do so regularly. By contrast, three new players executing a practiced strategy against tired legs might make an impact. Moreover, making a big change early forces the opposition to react. For instance, the most sensible response to an attacking blitz is to freshen the defensive line. But then the opposition has no chance to replace its own attackers later on.

A foundation of this approach comes from evolutionary biology, where researchers may want to find out what species can successfully invade the territory of another. Similarly, game theorists often ask what sort of strategy can win against a group of opponents playing a different strategy. Once everyone is playing the new strategy, a stalemate arises. But the innovator always has the first-mover advantage – the first hawk among the doves, until only hawks are left.

My approach assumes that it will be one of the first hawks. Not everyone will use so many subs so early, and therein lies the advantage. Yet apart from this strategic edge, my approach would have other more general benefits:

  • All players know their roles
  • All players know why they’re in those roles
  • All players know with whom they’ll be playing
  • All players know their partners’ slightest tendencies
  • All players know in what kind of situations they’ll play
  • All players know what to look for in the game before they come on

There are tradeoffs, naturally. Some of my approach’s costs are obvious, too:

  • It’s harder to surprise other teams if you play the same system
  • Other teams will find it easier to prepare for specific match situations
  • Subbing in lines of three players raises the risk of having to play with 10 men

This last risk is quantifiable. Injury-related substitutions happen in the second half of about 20-25% of games in the English Premier League. Of course, making early tactical substitutions raises the share of players on the field with fresh legs, likely reducing this risk somewhat. And it’s easier to play with 10 men when three are fresh, too. In any case, it’s a cost that I believe will be outweighed by the benefits of my approach.

The personnel requirements of this approach are somewhat flexible, given that different kinds of players can make up each of the subgroups. As an example, though, consider a 4-4-2 with a diamond in midfield. The attacking blitz special teams would be made up of two strikers and an central attacking midfielder for like-for-like switches. The game-killers with possession would be more attacking midfielders and/or wingers. The game-killers without possession would be a central defensive midfielder and two fullbacks, or a center back and two full backs, or a central defensive midfielder and two center backs, etc. – all depending on where they would be used. In a 4-2-3-1, the attacking blitz special teams would be made up of a striker and two wingers; the rest could remain more or less the same. And so on.

Either way, the subgroups would take up a total of 23 roster spots. That may sound like a lot, especially when 12 of them are for role players. Yet the point here is to use each spot not for maximum minutes but rather for maximum contribution to results.

By this point, some readers will have noted similarities with mainstream strategies in other sports. That’s no accident. I don’t think cross-pollination between sports has exhausted all the useful ideas – in fact, I think it has barely started. After all, how many soccer coaches in the top leagues are familiar with the intricacies of American football, ice hockey, baseball, and basketball? I’m no expert in rugby, but there would probably be some useful ideas there, too.

The overall theme of this approach is specialization and commitment – specialization of skills and situations, commitment of players to each other and to an overarching strategy for improving results. I believe that for most teams, the benefits of specialization outweigh the luxury of flexibility in how players are used. I also believe that total commitment on the part of all personnel is necessary to test a new approach fully. It’s risky, but the kind of change that can help a small club to upend a big one always is.

I’ve deliberately offered my ideas in a radical way, too, because I think that’s the best way to spur debate and change. I’m sure I’ll get my fair share of ridicule. And of course, my approach might not work. Even if it does, it probably won’t survive in exactly the same form I’m proposing here. That’s fine; I know I’m a decent data analyst, but I have no idea how good I am as a coach. I’m just convinced that the status quo is vulnerable to new ideas, and I hope teams will dare to try them out.