In Defence of Turnovers – or – Seizing the Means of Concession

Let’s do a quick thought exercise:

Most football matches are played in league competitions. League competitions are won by collecting the most points possible. Most leagues now award 3 points for a win, 1 point for a draw, and 0 points for a loss. This point distribution system creates an interesting phenomenon: more points are allotted, in total, for matches where a side wins, than matches where the two sides draw. This results in a curious dynamic: a side that draws every match in a 38 game season will end the year with 38 points, while a side that wins half of its matches and loses half of its matches will end the year with 57 points, despite the two efforts having been seemingly the same (half success, half failure). 

Now, if we are attempting to win a league title, let us assume that we have greater quality players than the majority of the league. Sides with lower tier talent generally do not target league titles, they target relegation survival, a mid table finish, or occasionally a top half finish. That is not our hypothetical objective. We want title glory. So we have players of quality to match that ambition.

From here, let’s make a jump. Let’s assume that, as long as we don’t shoot ourselves in the foot, we will outperform sides with less quality than us. To operationalize this, let’s say the quality of our team is 2.0, as a holistic value. Let’s say we are playing a hypothetical opponent whose quality can be represented as a value of 1.0. Based on this, in the average match, you would expect our team to create ⅔ of the chances, and score ⅔ of the goals, to thus on average collect ⅔ of the points. I realize this doesn’t read intuitively (what do you mean you score ⅔ of the goals and win ⅔ of the points? A 2-1 win means you win 100% of the points!), but bear with me. We are discussing the average result here. Imagine we are playing this match 1000 times. This is where you’d expect us to wind up.

Let’s now say that every match in this 1000 match series has 2 goals. That means, on average, we collect roughly 1.3 goals, and our opponent collects roughly 0.7 goals, for an average goal difference of 0.6. Obviously, because this is an average result, we will deviate from this. There will be 1-1 matches, and 2-0 matches, and occasionally we will lose 0-2. But the frequency of these results will be defined by the probability of their deviance from the average 1.3-0.7 result. 0-2 results will happen less than 1-1 results, which will likely happen more than 2-0 wins. 

Now let’s up the goal count. Every match in this 1000 match series now has 6 goals (I know it’s a lot, but you’ve made it this far, so just trust that I have a reason). We now collect 4 goals per match on average, to our hypothetical opponent’s 2. Once again, deviation will occur. 4-2 will probably be the most frequent result, but there will also be plenty of 5-1s, 6-0s, 3-3s, and 2-4s, as well as the occasional 1-5 or 0-6. But consider that the frequency of these results will again be characterized by their deviation from the average results. We will be losing and drawing less than we were in the three goal a match series, on average. As a result, we are likely to collect far more points in this high goal environment than we were in the low goal environment. What is the key takeaway here? 

Good teams can collect more points – without increasing their quality relative to their opponents – by playing a style of football that increases the frequency of goals on either end of the pitch. 

So, now that we’ve got that out of the way, let’s move onto the next issue. How do we increase goals? It’s one thing to say, ‘oh, just make more goals happen in a match without changing how those goals are distributed.’ It’s something else entirely to do it. 

The answer lay in how the ball is used of course. A rather formative notion, and one I touched upon in my piece on Georginio Wijnaldum, is the idea that winning football matches isn’t so much tied to having possession as it is to not conceding it. Now, these may seem like the same concept, but they’re actually quite distinct. Not conceding possession isn’t about how long you have possession for – it’s about giving your opponent fewer opportunities in transition than they give you. Why is this? Because counterattacks produce goalscoring opportunities more frequently and of greater quality than any other means of creation in open play. This has been born out statistically more than once in more than one place. It’s also the reason why some of the world’s best sides (read: Liverpool FC) have predicated their entire systems around the counter-press – an organized means to create chances in transition. The reasons this is true are many: defences are disorganized in the moments after a team concedes possession, there are often insufficient players behind the ball, and there is frequently greater space in behind the back line of the team conceding possession. To be honest, though, we don’t really care about the why. We care about the what. And the what is this: we want to maximize goalscoring opportunities in matches against smaller sides, and the best way to do this is to maximize the number of times we win the ball in transition. 

This still doesn’t feel like a revelation, though. So what, everyone should counter-press? Well, I’m not not saying that, but it’s also not the point. Instead, I have one more logical jump for us to make together: let’s go back to winning the ball. We know we want to do this more than our opponent. We want to do it as much as possible, in fact, to maximize the number of chances we create. Because we are playing against a club of lesser quality, let us assume that when we concede possession, it is less likely to result in a goal than when our opponent concedes possession. Essentially, let us assume we have a superior defensive structure, whether that be through counter-pressing or not. This is vital: basically, irrespective of how we do it, our turnovers cost us less, because we can more frequently turn our turnovers into a subsequent turnover by our opponent. If this is true, then there becomes a logical next step: if we have a replicable system for winning the ball back after conceding possession, and we want to win the ball back as much as possible, we actually want to concede possession, under the right circumstances.We want to seize the means of concession, so to speak.

In this chart, Southampton obviously stands out as the side with the most turnover intensive style of play. This isn’t particularly surprising, as they, like Liverpool, press high and systematically. However, I’m going to skip over the Saints in favor of discussing a handful of others.

Let’s consider one of our principles again: we want to win the ball more than our opponent, while giving it away less. The larger the margin, the better. With a quick glance at this graphic, you might say ‘well, Liverpool looked to maximize on both ends, but it looks like Manchester City actually looked to minimize their turnovers.’ You would be correct. There are a lot of explanations for this. For starters, Manchester City have some of the best possession based players in the league and Guardiola’s positional play reliably creates chances against set defences more than any other attacking system, perhaps of all time. Actively conceding possession to create transitions isn’t as much of a priority when you can cut through a deep block like a wet tissue. Not all of us are so privileged, though, which is what brings us back to Liverpool. 

The league winners have perfected this high turnover football. Not only have they created a replicable system to recover the ball, but they’ve also created a systematic and high upside means for conceding it in the right spots: crossing.

As you can see here, Liverpool, and specifically their fullbacks, are England’s leaders in hocking the ball into the box. This might sound derogatory, but it’s not meant that way. The deep crossing that Liverpool repeatedly harken back to against set defences in the final third has a plethora of functions. First off, it puts the ball between the penalty spot and the goal. That’s a great place to score from. You knew that part, though. Arguably more importantly, it’s a great place to win the ball back from. Cleared crosses are an excellent opportunity to engage the counter-press, win the ball back, and score in transition. Liverpool cross just as much to recycle possession as they do to create chances directly. 

Once you know this, it puts some things in perspective. For instance, Trent Alexander-Arnold’s turnover count suddenly doesn’t look so hideous.

To be clear, am I saying giving away the ball is good? No. But, when winning the ball is valuable, if you can control how you lose it in such a manner that it facilitates winning it back again, that can have positive effects. Liverpool have made a well oiled machine of this strategy, but it applies in a couple other places in a less pure, deliberate form. Take Bruno Fernandes.

Manchester United were very pedestrian in attack for the first half of the 2019/20 campaign. Paired with what was actually one of the league’s best defences, the club played some inconsistent, non incisive football against smaller sides. Chances, ideas, and goals were few and far between for much of the Autumn and into the Winter. A big part of the inability to create chances was down to stagnancy in possession: the side lacked line breaking passers (see also: Pogba, Paul – Injury), and would often pass it around the back until playing Daniel James down the line into a corner. Most of their chances, and in turn goals, came on the counterattack. This state of affairs meant quite a few dropped points, and the side was clearly screaming for a creative midfielder. Enter the aforementioned Portuguese magnifico.

Much has been made of Bruno Fernandes’ impact at Manchester United. The media has almost certainly overblown his direct effect on proceedings, partially because of his inflated goal contributions due to taking penalties and set pieces, and partially because of his undoubtedly valuable yet unquantifiable leadership ability. In turn, rival fans have – in a classic case of footballing allergic reaction – undersold his impact. What’s been touched on in all of this qualitatively, but not quantitatively, is the effect he’s had on United’s pace of play. Many obtuse references have been made to how ‘United look like United again’ and ‘They’re playing fast, attacking football and taking chances.’ These statements, weirdly enough, are true. 

United did not concede possession more, or win the ball more, after Fernandes arrived. In fact, they did both less. However, the nature of their ball concession changed, specifically in the attacking third:

United gave the ball away in the final third at the same rate, but they were doing it in a different way: rather than losing the ball to tackles, United were conceding possession via failed pass attempts.

This is because, rather than playing safely, Fernandes was forcing the issue by taking risks in possession with his passing. Take note of how United’s most frequent front 4 pre-Fernandes (Daniel James, Andreas Pereira, Marcus Rashford, and Anthony Martial) each conceded possession with the ball at their feet more than three times per match. In a lot of cases that was down to safe passing and holding onto the ball too long. Bruno eschewed that in a big way. 

None of these instances are turnovers, but that’s not the point: they’re high risk high reward actions that increase the chances of a goalscoring opportunity in the face of an increased probability of conceding possession. United sorely needed that. And, thanks to what was statistically a strong and reliable defence, the side could benefitted from this higher risk play in the final third. In fact, United’s defence actually improved after Fernandes’ arrival. This is a bit of an aside, but one explanation for this is that Fernandes’ progressive play actually resulted in opponents winning the ball deeper into their own halves – thus decreasing the threat of their possessions to begin with. 

Is this to say that Fernandes’ being wasteful is definitely only a good thing? Certainly not. As United improve as a side, Fernandes’ risk taking will presumably take on less value as the side around him develops a greater proficiency for creating chances in settled possession. However, there’s no doubt that his rather pedestrian underlying metric do not tell the whole story of his impact on the side. 

Where Fernandes’ high risk play likely improved United, there are also instances of low turnover play perhaps not suiting other presumably high quality sides. Take, for instance, Arsenal’s Granit Xhaka. 

This year in the league, Arsenal have been the most turnover-stagnant team in the Premier League. No side has had fewer total changes of possession between themselves and their opponents in the average league match. Their turnover differential was, similarly, also extremely poor, with a value in the same realm as Newcastle United, Aston Villa, and Burnley. This likely had a large role to play in the stagnant buildup play that the side found themselves engaging in for much of the year. Short-lived club captain Granit Xhaka played into this, rarely winning or conceding possession.

Before we can properly evaluate what Xhaka’s impact in this vein is, we have to again go back to the issue of ‘is conceding possession bad.’ Objectively, yes. But, as we explored earlier, this is not always true if you have the quality to win it back, and if you have the quality to more reliably create chances when your opponent does the same. Regardless of how poor Arsenal were this year, they still have greater quality relative to most of the Premier League. As a result, this low turnover style of play undoubtedly hindered their ability to collect points, and the low variance nature of their matches certainly fed into their league leading number of matches drawn. As such, Xhaka’s conservatism on the ball and inability to contribute to regaining it when out of possession was likely a net negative.

With Granit, though, there is hope. 

Under Arsène Wenger, in the 2017/18 campaign, Xhaka was used far more aggressively, indicating a capacity and suitability to higher variance play. For most, this probably falls more in line with their perception of Xhaka – his defining characteristic is his distance passing, a trait not known for its ‘safety’ or ‘conservatism.’ If Mikel Arteta is so inclined – and, perhaps, knows what’s good for him – Xhaka may see a return to a role that takes advantage of his talents and puts Arsenal in a better position to win matches – not just not lose them.

Before I leave you to stew with this, I’d like to stress this once more – I am not saying turnovers are good. What I am saying is that they don’t have to always be as damaging as they appear in the moment, or on the stat sheet even. In some cases they can even be a tool. This reality is worth considering the next time you go to evaluate a player, or a club, on their characteristics in possession.