Skip to Main Content
Audience Engagement, Fan Engagement, Media & Tech

How To Replace A Legend: By Wojciech Szczesny


From an inconsistent performer at Arsenal to the heir apparent to Gianluigi Buffon at Juventus, Oliver Hopkins uses Stats Perform’s advanced goalkeeping metrics to analyse the rise of Wojciech Szczesny.

By: Oliver Hopkins

It is not easy to replace a legend. Just ask David Moyes or Unai Emery.

How then do you replace Gianluigi Buffon, one of the greatest goalkeepers of the modern era? A Juventus legend, a man who’s made 516 appearances for the Old Lady in a career spanning two decades, picking up nine Serie A titles in the process. An Italian legend, the most capped player in the history of the Azzurri national team (176) and a World Cup winner. How can you possibly follow that?

Easy, just ask Wojciech Szczesny.

The Journey to Turin

Szczesny’s ascension to Juventus’ first choice goalkeeper was rubber-stamped earlier this year with the Pole signing a new four-year deal that will see him remain in Turin until 2024. There can be no greater endorsement of how effectively he has replaced Gigi Buffon between the sticks.

It has been the career trajectory that many Arsenal fans cannot fathom.

After rising through the ranks of the youth setup in north London, and after a successful six month loan spell at then-League One club, Brentford, Szczesny became first choice for the Gunners during the 2010/11 Premier League season, displaying great athleticism as a shot-stopper capable of pulling off spectacular saves. He had the makings of a fine ‘keeper, but, as has become notorious with Arsenal’s defending in general, allowed individual errors to creep into his game far too often.

Indeed, from the start of the 2011/12 Premier league season – Szczesny’s first full season as Arsenal number one – until his departure in 2015, only four goalkeepers: Asmir Begovic (14), Joe Hart (13), Tim Krul (13) and Tim Howard (12), made more errors leading to goals in the competition than the Pole (11).

The mistakes weren’t just confined to league play, with his mix-up with centre-half Laurent Koscielny – a mistake which gifted the League Cup to Birmingham City in 2011 –  among the most high profile. For many, it summed up Arsenal’s defensive fragility, coming just weeks after they had surrendered a four-goal lead away at Newcastle.

The 2013/14 Premier League season would be Szczesny’s best at Arsenal. His 16 clean sheets were enough to tie him with Petr Cech for the Golden Glove Award. However, disciplinary incidents tested the patience of Arsene Wenger, notably after Szczesny was allegedly caught smoking in the changing rooms after a game, and the Pole was subsequently loaned to AS Roma in 2015. It is easily forgotten that at Roma, Szczesny kept future Liverpool goalkeeper Alisson out of the team.

Still, it seemed unthinkable that five years after his Arsenal exit he would be the heir apparent to Buffon.

Evaluating Goalkeeper Performance

There is no doubt that Szczesny is enjoying a fantastic season with Juventus, recording the joint-highest number of clean sheets (10 – tied with AC Milan’s Gianluigi Donnarumma and Udinese’s Juan Musso) and the best save percentage in Serie A (80%) among those ‘keepers to have played five or more matches.

However, evaluating goalkeeper performances using traditional metrics has its limitations, as these numbers can be heavily biased by team and defensive strengths. Szczesny has been part of a dominant Juventus side and you’d rightly expect him to perform highly in these metrics regardless of his ability between the posts.

We can isolate an individual goalkeeper’s performance using our expected goals on target (xGOT) model. Using this model, we can predict how many goals a goalkeeper would be expected to concede, given the quality of the shots location (xG) and the end goalmouth location of the shot. As a result, we can directly evaluate the contribution of the goalkeeper and see how many goals they prevented for their team.

Click to enlarge

So far this season, Szczesny was expected to concede just over 20 goals according to our xGOT model. Given that he’s actually only conceded 13 (excluding penalties and own goals), the Pole has prevented over seven goals with his saves this season. This is by far the highest figure in Serie A, with SPAL’s Etrit Berisha clocking in at second place with just shy of four goals prevented. Looking around Europe’s top five leagues, Szczesny’s tally of 7.1 goals saved is enough to rank him in sixth position.

But hang on, don’t those goalkeepers who face more shots have an opportunity to prevent more goals? To allow for a fair comparison between goalkeepers, we can standardise for the number of shots each ‘keeper faced by looking at their goals prevented rate (xGOT conceded divided by goals conceded).

Szczesny’s goals prevented rate of 1.55 this season means he was expected to concede 1.55 goals for every goal that he actually concedes. The Pole ranks first in this metric in Serie A amongst first choice goalkeepers.

Serie A - xGOT rankings 2019/20 (500+ Mins)

RankTeamPlayerMinutes PlayedGoals ConcededxGOT ConcededGoals PreventedGoals Prevented Rate
1JuventusWojciech Szczesny1,9801320.17.11.55
2SPALEtrit Berisha2,3403639.83.81.10
3VeronaMarco Silvestri2,6102831.63.61.13
4LazioThomas Strakosha2,61024262.01.08
5CagliariRobin Olsen1,52518202.01.11

While xGOT is a great measure of an individual’s performance in comparison to the ‘average’ goalkeeper, each stopper in the league will face different types of shots over the season. These shots will be influenced by the defensive style of their team and the opponents they face and hence, may or may not be favourable to an individual goalkeeper’s strengths and weaknesses.

Using Stats Perform’s Goalkeeper Index, we can now personalise these estimates to each goalkeeper and simulate how they would perform over the exact same sample of shots.

How Does The Stats Perform Goalkeeper Index Work?

The Stats Perform Goalkeeper Index uses AI modelling techniques with several descriptors to capture the influence of a particular goalkeeper on the shot outcome.

On top of the start and end location of the shots, the Goalkeeper Index personalises the predictions by considering a goalkeeper’s historic ability to save shots based on factors such as shot destination and angle of the shot.

These features capture goalkeeper strengths and weaknesses on a high level and act as proxies for factors such as handedness, anticipation, and reaction time for shots of varying angles and destinations.

How Does Szczesny Rank?

Using the Stats Perform Goalkeeper Index, we can simulate how each goalkeeper would perform if they faced every single one of the shots faced by goalkeepers in Serie A so far this season. This gives us the number of goals that each goalkeeper would be expected to concede.

Serie A Gk Index Rankings

Again, Szczesny is head and shoulders above the league, and is expected to prevent 0.41 more goals per game than the league average.

What Would It Look Like If Buffon Had Been In Goal This Season?

We can take the model further and directly compare two goalkeepers over the course of the season. This means we can answer fun hypothetical questions like, “what if Buffon was still Juventus’ number one?” The graphic below shows a simulation for how the two shot-stoppers would fare if they faced every single one of the 114 shots on target conceded (excl. set pieces) by Juventus this season. Szczesny, the model suggests, would concede 3 goals fewer than his deputy, with the Pole notably stronger in the centre and to his upper left.

His natural ability has never been in question, but where Szczesny has really improved is the elimination of some of the costly mistakes that became common at Arsenal, having made just one error leading to a goal in almost three full seasons at Juve. Five years on after leaving London, there is a strong case to be made that Szczesny has now become best ‘keeper in Serie A in 2020.

The Rankings Across The Rest Of Europe:


Goalkeeper Rankings Premier League Week 32

RankPlayerGame PlayedSimulated Goals Conceded (Per Game)Simulated Goals Prevented (Per Game Compared To League Average)
2David de Gea320.890.22
4Vicente Guaita290.930.18
5Bernd Leno300.950.16
6Dean Henderson301.010.10
7Kasper Schmeichel321.030.08
8Paulo Gazzaniga171.040.07
9Lukasz Fabianski191.050.06
10Hugo Lloris151.080.03


Goalkeeper Rankings LaLiga Week 33

RankPlayerGame PlayedSimulated Goals Conceded (Per Game)Simulated Goals Prevented (Per Game Compared To League Average)
1Sergio Asenjo280.620.26
2Jan Oblak310.630.25
3Thibaut Courtois280.670.21
4Alex Remiro240.750.14
5Jasper Cillessen200.750.13
6Unai Simon280.760.13
7Marc-Andre ter Stegen300.780.11
8Aitor Fernandez310.780.10
9Ruben Blanco290.810.07
10Manuel Reina300.840.04


Goalkeeper Rankings Bundesliga Week 34

RankPlayerGame PlayedSimulated Goals Conceded (Per Game)Simulated Goals Prevented (Per Game Compared To League Average)
1Yann Sommer340.830.38
2Manuel Neuer330.900.31
3Peter Gulacsi320.970.23
4Lukas Hradecky341.000.20
5Kevin Trapp221.030.18
6Alexander Nubel261.080.13
7Roman Burki311.090.12
8Rune Jarstein291.120.08
9Rafal Gikiewicz331.160.04
10Alexander Schwolow241.180.02