The Athletic has analysed the style and performances of those linked to the Tottenham job to see how they might do as Mourinho's replacement
theathletic.com
From Fonseca to Rangnick: using data to scout Tottenham’s next manager
By
Tom Worville
Recruiting a manager is arguably one of the toughest tasks in all of football.
The initial filtering process is fairly straightforward as there’s only a limited pool of options to choose from. The repercussions of having the wrong man at the helm can be far more damaging though in both the short and long term compared to a player signing.
That’s due to several reasons, chiefly the pay-off required when sacking them, the churn required to get in new players for the next person in the job and the opportunity cost of the time spent with the wrong manager in the first place.
Of course bringing in, say, a sub-par central midfielder carries the same sort of potential downsides. The impact is lessened though as not all players need to (or can) play all the time. Most players retain some degree of their value too, which the club can cash in on at some point.
Some clubs are able to insulate themselves against these associated costs by making the manager a head coach instead, reducing the scope of the role just to making the team better on the field.
That’s not the case with Tottenham though, who at this moment in time do not have a sporting director, although
Daniel Levy is believed to be considering bringing one in. Franco Baldini was the last director of football at Tottenham back in 2015, with technical performance director Steve Hitchen carrying out similar duties with less of an influence.
Compared to a player signing, the questions asked of potential candidates are slightly different too. Will the manager have a short-term impact? Are their performances consistent? Do they blood younger players? Do they improve the team? What’s their playing style like? How has it changed over time? Does it fit the squad? The list seems never-ending, yet each answer helps reduce the uncertainty around who the right candidate is.
These are the sorts of questions that Hitchen and Levy are likely to ask each potential candidate, so we shall too, with the help of data…
Data scouting for players is very widespread these days, but the same analysis of coaches is less commonly seen, at least in the public domain. There’s plenty of publicly-available data on teams, but with managerial tenures stopping and starting throughout any given season, it’s not always the case that a team’s stats for a season are reflective of a single coach.
It’s also true that the stats we see for teams aren’t
exactly the numbers we want to look at for a coach. We care more about progression and evolution of a team rather than just single-season performance, and there’s the non-performance-related data which is interesting too. What’s the age profile of the team? Does the manager always rely on a core group of players? Do they need to make lots of transfers?
With the help of
Piotr Wawrzynow, a football data scientist from Krakow, Poland, we can appraise some of the names linked to Tottenham. Wawrzynow’s work (a detailed piece on his methodology can be found
here) led to him creating a dashboard that splits out a coach’s performance and playing style.
The basis for how Wawrzynow measures performance lies in a coach’s ability to regularly progress the ball into threatening positions and stopping the opponent from doing the same. Think of this as an extension of the expected goals model (a measure of chance quality based on the shots a team takes), as there may be times that a team gets the ball into a good location but for one reason or another doesn’t shoot.
Additionally, separating results and performances is important, as some coaches are able to grind results out over a period of time even if the team isn’t playing very well. Seeing how performances change over time is important, too: have things steadily improved, declined or stayed about the same?
This performance rating then powers some of the
Tenure Profile ratings. These are:
Average Rating, which looks at how well the team played across the course of the manager’s tenure;
Performance Impact, which helps us understand if performances got better, stayed the same or worsened over time, and;
Short Term Impact, which measures whether the side improved under the coach from when they first arrived to their first pre-season.
In terms of the play style of the coach, Wawrzynow devised some stylistic measures to understand
how a team plays. There are eight measures which cater to both in and out-of-possession traits, with the dashboards showing how these tendencies may change over time and who the manager is most similar to.
Diving into the profiles for seven coaches who have been linked with Tottenham in recent weeks, we can get an objective view of what the data says about them, and whether they’re likely to be a good fit for Tottenham.
Graham Potter
Starting with Graham Potter, his overall coaching performance rating of 97 out of 100 makes him one of the best coaches in the Premier League today.
Brighton’s performances (blue line on the right-hand chart) indicate a side which is great at getting the ball into good positions, regularly doing so better than their opponents. It’s quite a distance above their results in parts (the red line) which is based on an Elo model which adjusts for each match depending on how likely or unlikely the result is.
That trend echoes much of what we’ve seen from Brighton and Potter this season: that the process is good, but the results haven’t quite caught up yet.
Looking at the “most similar coaches” section of Potter’s dashboard, we see that there was a clear difference in style between him and Chris Hughton. Potter favours a possession-heavy approach with plenty of build-up and a focus on dominating the territory in the final third, rarely relying on long balls and counters.
His average rating for a low block and high pressing indicates a difference in style to both what Spurs played under Jose Mourinho and how Brighton played under Hughton, but also his effectiveness at instilling a completely different approach in a short period of time.
Looking at Brighton’s player-age matrix below also shows a faith in younger players, with Potter’s side having the fourth-lowest weighted average age (calculated by minutes played x age for all players) of just 26 years old.
Brighton’s finishing woes are well-documented, but under the hood they’re a top-ten level Premier League side with the
15th-largest total income in the league in 2020. If Brighton had Harry Kane on the end of their chances they’d likely be sat in mid-table. If Tottenham had Graham Potter at the helm, they’d have a coach who’d play an eye-catching brand of football while also setting the club’s course for potential title challenges in the future.
Brendan Rodgers
Brendan Rodgers’ time so far at Leicester has brought about some excellent league performances. Leicester just missed out on Champions League football last season but have already beaten their points total with four games to spare in 2020-21, and have an FA Cup final against Chelsea to look forward to.
Rodgers has largely overseen excellent performances per Wawrzynow’s model, although recent results are above where you’d expect performance levels to be, the repeatability of which remains to be seen.
He does rate favourably in terms of his short-term impact, which is important given the presence of the European Championship means that pre-season will be a short one this summer.
Stylistically Rodgers looks similar to Potter, where he places emphasis on build-up at the back, also utilising a back-three similar to the Brighton coach. His team employs a slightly higher press, but both coaches share a focus on creating chances from open play — something that definitely suits the likes of Kane and
Son Heung-min. He’s also shown he’s capable of changing his style too, as Leicester in 2020-21 have looked to build and cross from wide areas less than previous Rodgers sides.
The current style of play differs hugely compared to predecessor Claude Puel, evidence that Rodgers has been effective at coaching his players and not requiring a huge amount of squad churn. Since his arrival at the club, the only two notable departures have been Harry Maguire at the end of 2018-19 and Ben Chilwell after 2019-20, with most of the other departures being rotational options anyway.
Rodgers, like Potter, has shown trust in youth too, with Wesley Fofana, Youri Tielemans, Harvey Barnes and James Justin all contributing significant minutes this season.
Rodgers has shown at Celtic that he’s not just a good coach but a winning one too, and with Leicester he’s demonstrated that he’s capable of developing a side over time, even if the models think there’s more than meets the eye.
Erik ten Hag
Erik ten Hag is now an outside shout for the Tottenham job given he’s signed a one-year contract extension keeping him at Ajax until 2023, but he’s still a name worth assessing to understand how good he actually is.
The models suggest that he’s been excellent at Ajax since becoming head coach in 2017-18. The majority of his tenure hasn’t seen much improvement or decline in performances, with Ten Hag rated as a very good manager per Wawrzynow’s numbers overall, and excellent for the Eredivisie.
Stylistically, he’s got a very clear style that resembles the likes of Pep Guardiola when at Bayern Munich and Unai Emery when at Paris Saint-Germain. That’s because of how clearly defined Ten Hag’s style of play is: looking to build up from the back, dominate territory in the final third and press with a high intensity.
The main concerns with Ten Hag are that he has never managed at a higher level than the Eredivisie and European competition aside hasn’t regularly faced opponents with the quality of which he’d come up against in the Premier League.
There’s also no evidence that he’s able to instil this sort of playing style on a group that haven’t played that way before. Ajax are one of only a handful of clubs in world football who have such a distinct style, so there’s an argument to suggest that Ten Hag didn’t have much of a challenge on his hands to tweak the way the squad played when taking over in 2017 to suit his own game model. That’s informed by the fact that the difference between his predecessor’s style and his own is minimal.
Of course, there’s no candidate in this whole process who is a perfect match, but Ten Hag represents one of the more risky choices who may take a bit of time to get the players he needs and the team playing the way he wants.
Marcelino
Athletic Bilbao’s Marcelino has only been at the club for a matter of months after being appointed in January 2021. For that reason, the data from his time at Valencia is more suitable to analyse the sort of manager he is.
His tenure was extremely successful, with back-to-back fourth-place finishes in his two seasons in charge, a Europa League semi-final appearance in 2018-19 and a Copa del Rey win in the same season.
From his dashboard below, Marcelino grew and grew into his tenure, with his side making a big jump in the quality of their performances in his second season, with Wawrzynow’s model seeing no regression in performances across his 79-game stint in charge.
Stylistically, Marcelino is different to the possession-dominant managers we’ve seen up to this point, often defending in a low block and looking to attack teams either on the counter or from wide. His second season saw more of a focus on the wings, with less time spent building at the back.
Marcelino also gave young players like Carlos Soler and Jose Luis Gaya a chance, with his 2018-19 Valencia side being the fifth-youngest in La Liga overall.
The main question with Marcelino is whether his style suits a Tottenham squad that is of a higher quality than what he had at his disposal at Valencia, or if he himself needs to adapt in order to succeed in a league dominated by teams who have plenty of the ball and look to press high and effectively.
Ralf Rangnick
Ralf Rangnick’s 2018-19 season at RB Leipzig saw them finish third in the league, playing a very intense style that thrived on pressing high up, countering but also dominating the ball in the final third at times too.
Rangnick’s Leipzig hit the ground running and never really relented. Looking at their Average Rating per Wawrzynow’s model they were excellent throughout the season, consistently dominating opponents. That tallies up with their expected goal difference per game which, according to Statsbomb data on fbref.com, was +0.76, the second-best in the league after Bayern Munich.
There’s also something Bielsa-esque about Rangnick’s approach to managing games, as the “key traits” section shows. His Leipzig looked to dominate games irrespective of the scoreline, never taking their foot off the gas. There was also plenty of positional freedom for Leipzig’s players, something the Spurs squad hasn’t been exposed to under Mourinho.
Rangnick perfectly suits Tottenham’s need for a manager who is able to both play and improve young players. Looking at his squad-age matrix below, he put plenty of trust in those aged 24 and under, all of which he had overseen the signing of in his role as Leipzig’s director of football since 2012. Leipzig’s average weighted age was just 23.5 years old, making them the youngest side in Europe in that season.
Hiring Rangnick would likely require Levy to cede some control and let the German build a similar project at Tottenham over the course of a couple of seasons to the one he had overseen at the Red Bull clubs. With only limited recent managerial experience — and none in England — he’s perhaps a better fit for a sporting director role.
Paulo Fonseca
With Roma’s Paulo Fonseca being replaced at the end of the season by Mourinho, Tottenham could feasibly stop part of the European manager roundabout before it gets going by bringing in the former Shakhtar head coach.
Fonseca’s tenure at Roma started well, with his side finishing fifth in Serie A, an improvement on predecessor Claudio Ranieri’s sixth-placed finish. Recently though performances have really tailed off alongside results, ultimately leading to Roma seeking an alternative manager for 2021-22.
Fonseca’s playing style — according to the “most similar coaches” section of the dashboard below — wasn’t overly similar to the Portuguese. Fonseca focuses heavily on building from the back, making him most similar to the likes of Bruno Genesio at Lyon and Unai Emery at Woolwich.
It’s not just been possession for possession’s sake this season either, ranking as the 12th-best attacking side in Europe when looking at non-penalty expected goals per game. For context, Tottenham have been the 30th-best side in Europe, and just the eighth-best in the Premier League.
Fonseca has been particularly unlucky with injuries of late, which explains some of the downturn in form. Centre-backs Chris Smalling and Marash Kumbulla have missed ten and seven games respectively either due to injury or suspension since early February, with only Gianluca Mancini playing more than 80 per cent of minutes this season.
A lack of Premier League experience may go against Fonseca, but he’s shown he’s able to create a formidable attacking side in the last couple of seasons. It’s hard to shy away from his defensive shortcomings though, most recently on show in Roma’s 6-2 hammering away to Manchester United in the Europa League.
Nuno Espirito Santo
Nuno received plenty of praise for taking Wolves from the Championship and making them an established Premier League side in a short period of time, but they’ve regressed this season.
Per Wawrzynow’s models, Nuno is a manager who takes a while to get going and has seen limited improvements in performance since Wolves’ promotion in 2018-19 when looking at the “performance vs results over time” section of the dashboard below.
Stylistically, Nuno is one of the least possession-oriented managers in the conversation. He has a large focus on using the wings to stage attacks, crossing plenty and rarely looking to dominate possession in the final third. Despite Wolves being a side who defend in a low block, they don’t counter as much as you might expect either.
If Tottenham were looking to continue Mourinho’s defensive approach, he would be an ideal fit. Given they’re looking to get back to a
Pochettino-esque style of play though, Nuno’s shades of his compatriot are likely to be something of a turn-off for Levy and Hitchen.