I don’t like Dušan Vlahović, and I know how that may sound in this particular moment. The other day, Vlahović was involved in all three of Juventus’ goals against Frosinone in a 3–2 win, scoring their first two and assisting the winner. That improved his numbers to 15 goals and four assists this season. The fact that I’m writing an article about why I don’t like Dušan Vlahović right now may come across as me being stubborn about my opinion on him, and it may convey pettiness, but I think this might just be the perfect time for me to express my opinion on Vlahović. Last December I wrote an article where I ranked the top 100 players in Serie A, in which I ranked Vlahović 100th (read it here), explaining some of my issues with him. At that point, it was somewhat controversial to claim that there were countless players better than Dušan Vlahović in Serie A, but it was easier to justify with his form at the time. Since then, however, Vlahović has been on fire, but my opinion on him has not changed. Of course, I could just be wrong, but I truly do believe that Dušan Vlahović isn’t much more than an average striker on a continental scale, and in this article I would like to explain why.
Contextualizing the numbers
Dušan Vlahović’s story, as it pertains to the subject of this article, really begins with his 2020–21 season with Fiorentina. He was a promising talent at Partizan, sure, and he wasn’t completely anonymous in his first two seasons at Fiorentina, but it was in 2020–21 when it all really started. Dušan would start off the season quite slow, but would put up exceptional numbers in the second half of the season, ending it off on 21 goals and three assists. There are reasons to already start adding context to these numbers, but I think it’s more appropriate to also look at Vlahović’s 2021–22 season, in which he first scored 17 goals and assisted two with Fiorentina, before adding seven goals and an assist with Juventus in the second half of the season. I think this final run with Fiorentina — particularly referencing Vlahović’s run in the calendar year of 2021 — is the core to what I believe to be a misunderstanding of Vlahović by many. To a great extent, it doesn’t just come down to a misunderstanding- or negligence of underlying numbers, but rather a misunderstanding of more basic end product numbers.
In his final full season with Fiorentina, Vlahović scored 21 goals, and in his final half season with Fiorentina, he scored 17 goals, meaning he would almost definitely have reached the 20-goal threshold with Fiorentina that season if he’d just stayed — including his time at Juventus that season he ended up scoring 24 (all of this exclusively pertaining to Serie A). The magic number in regards to goals is 20, and in football discourse the 20-goal threshold holds a lot of weight, as exemplified to the reactions to Vlahović’s performances at Fiorentina. Now, the narrative that many have created since Vlahović’s departure from Fiorentina is that he has struggled due to factors outside of his control, with the supposed incompetence of Juventus manager Massimiliano Allegri being the most popular factor. However, the misunderstanding here is that Vlahović has largely played at the same level at Juventus as at Fiorentina. However, it is that magic 20-goal threshold which blinds us. It never ever seems to fail to manipulate us — once we see a 20-goal season, we automatically associate it with greatness, but this is an extremely narrow interpretation of stats in football. The 20-goal threshold is not necessarily synonymous with robust final third impact, and is not an effective measurement of effectiveness in that department. One of the best examples of how I think we misinterpret these kinds of stats, is Gabriel Jesus. Jesus is notorious as an unreliable striker who can’t finish and isn’t good enough in the final third, but in reality he has proven to be a phenomenal final third player throughout his entire career, despite never even eclipsing 14 league goals, let alone 20. Here are the reasons why:
- Gabriel Jesus’ final third impact does not merely come in the form of pure goals, but also assists.
- Gabriel Jesus has rarely taken penalties throughout his career.
- Gabriel Jesus has often struggled with playing time, be it as a result of high competition in attack or because of injuries.
So, what is it really that a lot of people miss about Gabriel Jesus? Well, in all competitions in his Manchester City career, he averaged 0.82 non-penalty contributions per 90, over a sample size of 159.8 90s. For reference, Harry Kane averaged 0.76 non-penalty contributions per 90 in all competitions in his Tottenham career, although that was over a sample size of 378.9 90s (via FBref). If you want to include Gabriel Jesus’ time at Arsenal as well, his non-penalty contribution rate is still 0.80 per 90, over a sample size of 201.8 90s (via FBref). So, going back to Dušan Vlahović, whereas the narratives that his performances were performing and he was being held back at Juventus had already gained prominence by the summer of 2022, the actual productivity-numbers are quite comical: he averaged 0.68 non-penalty goal contributions per 90 with Fiorentina the prior season, and 0.67 non-penalty goal contributions per 90 with Juventus (via FBref). So, what are the lessons here? Productivity is usually a better indicator of ability than production; penalties can be misleading; strikers can be creators too. In the case of Vlahović, he was never as productive as his basic end products implied: he only averaged 0.55 non-penalty goal contributions in his breakout season with Fiorentina in 2020–21, and even his non-penalty goal contribution rate of 0.68 in his final half-season at Fiorentina was not as impressive as the raw end product numbers suggested. This is to say, he may have hit the 20-goal threshold in back-to-back Serie A seasons, but this vastly misrepresented how productive and impactful he really was in the final third in these seasons. Again, the essential misunderstanding of Vlahović is illustrated by the fact that he was perceived to have dropped off in form at Juventus, but in reality his productivity remained almost exactly the same.
Sustainability of performance — Finishing
Now, let’s really get into the issues with Dušan Vlahović, which is where this article may become somewhat controversial. We’ve established that Vlahović’s productivity in the final third has widely been misunderstood throughout his career, but his underlying numbers are what indicate his actual limits as a player. We can explore this by going through his seasons one by one, starting with his breakout 2020–21 season. In this season he averaged 0.55 non-penalty goal contributions per 90, as mentioned, and not only is this not particularly good in and of itself, but his underlying numbers also suggested a minor overperformance: he averaged 0.49 npxG+xAG per 90. As has more or less been established, his productivity would rise significnatly in his final stint with Fiorentina, but whereas the overperformance of his underlying numbers the prior season was realistically sustainable, his overperformance with Fiorentina in 2021–22 did not seem that way. In fact, his npxG+xAG per 90 more or less remained the same, actually decreasing by 0.01, while his non-penalty goal contribution rate improved by 0.13 per 90. In 2021–22 with Fiorentina, the disparity between the npxG+xAG per 90 and that of the non-penalty contribution rate was 0.20, which is difficult to sustain in theory, and also turned out that way in reality. As a matter of fact, Vlahović’s npxG+xAG per 90 improved to 0.53 in Serie A with Juventus in 2021–22, while his non-penalty goal contribution rate essentially remained the same, as mentioned… yet, that was still a clear overperformance of the underlying numbers, so what does that tell us?
In terms of evaluating the sustainability of end product, it is perhaps more important to look specifically at the finishing, as the potential deviation between xAG and assists largely comes down to randomness, whereas an overperformance of npxG can imply an inherent ability to finish well, and therefore sustain an overperformance. In 2020–21 Vlahović overperformed his npxG by 0.04 per 90; with Fiorentina in 2021–22 by 0.18 per 90; in the 2021–22 Serie A season as a whole by approximately 0.17 per 90 (via FBref). Unfortunately, FBref does not allow us to compare average npxG-overperformance with an expansive sample of players, over an expansive sample size (every season countless players are able to overperform their xG, but only the great finishers are able to do it consistently over extended periods of time). Such a comparison would allow us to put these aforementioned numbers into perspective in a more effective way, but I think Erling Haaland — as one specific point of reference — is quite effective in this scenario all the same. He is probably the best finisher in the world, and in the top five leagues (for Dortmund and Manchester City) he has averaged an npxG-overperformance of 0.12 per 90. With this in mind, it would’ve been fair to say that Vlahović’s npxG-overperformance in 2021–22 was unsustainable, and that projection would seemingly be proven to be accurate about a year later, following the conclusion of the 2022–23 season (via FBref).
2022–23 is a complicated season to assess for Dušan Vlahović. It remains by far the worst season of his career post 2019–20, and while he did indeed fail to overperform his npxG the way he did the season prior, this was not the main reason for his end product plummeting. In Serie A his non-penalty goal contributions per 90 had regressed from 0.68 to 0.47; his non-penalty goals per 90 from 0.58 to 0.37; his npxG-overperformance from 0.17 to 0.03. It is hard to explain exactly what happened to Dušan this season, because — for instance — Juve’s chance creation numbers would not change drastically, in fact they would improve: their xG per game rose from about 1.46 following Dušan’s arrival in 2021–22, to 1.57 per game in 2022–23. The theory here, of course, is that it’s harder for an attacker to be productive the less dangerous its team is in the final third, and easier the more dangerous the team is. However, it’s hard to gauge exactly how an attacker’s productivity and/or threat is affected by their surroundings from a statistical point of view, because they themselves have a responsibility in the team creating chances — both through movement and through their own chance creation. This is to say: if a team struggles to create chances, that may not merely be caused by a lack of service for the striker, but also a lack of contribution from the striker themself. Nonetheless, the slightly increased chance creation by Juve in 2022–23 compared to the season prior, implies that the simultaneous regression of both productivity and threat from Dušan Vlahović was not a matter of the team regressing to any extent. In any case, it is notable how this season wasn’t just plagued by the theory that Vlahović’s performances were not sustainable (like the season prior), but by clear, concrete struggles in terms of his actual output; this season we actually saw Vlahović fail to contribute, as opposed to the prospect that it could happen in the future. This was partly through an overall regression, which remains difficult to explain, but also through the halt of his enormous npxG-overperformance, which implied that he wasn’t the best finisher on the planet after all, in fact he may not even have been one of the best ones… except now he’s back at it again.
Vlahović is currently enjoying the best season of his career, and that’s actually coming in many different forms so far. His phenomenal non-penalty goal contribution rate of 0.97 per 90 is by far the best of his career; his npxG+xAG per 90 of 0.60 is the best of his career; his npxG-overperformance per 90 of 0.23 is the biggest of his career (via FBref). So, okay, you can say that the npxG-overperformance is unsustainable, and that his numbers will regress over the remainder of the season — for one, but perhaps more importantly in pertains to the long-term. However, the problem with this is that he IS doing “it” right now; he’s been incredibly productive over a respectable sample size (17.6 90s), and even the underlying numbers are good — on the pitch he’s delivering, and he’s giving Juventus a lot in the process. This is what admittedly makes Vlahović such a complicated, unpredictable player; how can you just write off what he’s doing right now? The only fair answer to this is that you have to account for- and respect hot streaks such as the one Vlahović is enjoying at the moment. However, as for Vlahović’s volatility I think there is an explanation for it, which makes him a very limited, and ultimately average striker on a European level, for me.
Sustainability of performances — Shot selection
Having established that Vlahović is one of the most volatile strikers in the world, the main question becomes what the reason for this is, and to me it’s quite clear: it’s his shot selection. Vlahović has never been exceptional at amassing npxG from a pure output standpoint, and beyond that his npxG has largely been accumulated through low-quality shots. Our selected timeframe begins with the 2020–21 season, and within this timeframe this was Vlahović’s most efficient season in terms of shot selection. He didn’t take too many shots, with his attempted shots per 90 ranking in the 57th percentile in Serie A that season, but his npxG per 90 ranked in the 76th percentile, and his npxG per shot ranked in the 77th percentile (via FBref). However, these are not exceptional numbers anyway, and it was all down from there.
Moving on to the 2021–22 Serie A season (with both Fiorentina & Juventus), Vlahović’s npxG per 90 would still rank in just the 76th percentile, decreasing from 0.42 to 0.41 despite a 1.01 increase in attempted shots per 90, compared to the season prior. Vlahović ranked in the 52nd percentile for npxG per shot, and in the 38th percentile for average shot distance (the higher the distance, the lower the percentile). This implied that Vlahović was taking a lot of low-quality shots, but clearly he was still finishing at a high level, not only backed by his npxG-overperformance, but also the fact that he ranked in the 67th percentile for percentage of shots on target, and in the 87th percentile for shots on target per 90 (via FBref). This was not something he would be able to sustain, however.
Unsurprisingly, Vlahović’s poor 2022–23 season was plagued by further poor shot selection, and inevitably also poor finishing to some extent. His npxG per shot would practically remain the same, regressing from 0.12 to 0.11, thus ranking in the 42nd percentile. His average shot distance was also in the 42nd percentile, reminiscent of the season prior, however what changed was the outcome of Vlahović’s shots. Again, we partly saw this through Vlahović’s npxG-overperformance regressing from 0.17 per 90 to just 0.03 per 90, but also through the general accuracy of his shots, as he plummeted to the 30th percentile for percentage of shots on target (via FBref). It was a season that was plagued by individual failure on almost all fronts, but of course this season has been quite the opposite in some ways, so how has Vlahović’s shot selection and general shooting looked as his end product has exploded? The truth is there are still quite a few red flags. What has been impressive is Vlahović’s npxG per 90 of 0.51, which ranks in the 90th percentile, however it is perhaps here that the volume of his npxG is at its most misleading. His npxG per shot has essentially remained the same, which — in fairness — is impressive given his attempted shots per 90 has improved from 3.04 to 4.38 compared to last season. However, it’s still indicative of how many low-quality shots he attempts, as further illustrated by him ranking in the 28th percentile for average shot distance. Furthermore, his general accuracy has managed to remain low, as he ranks in the 42nd percentile for percentage of shots on target.
So, what is this all to say? Well, the main reason for Vlahović’s volatility is his reliance on low-quality shots. Is he a great finisher? Well, with the shots he attempts it doesn’t necessarily matter how good of a finisher he is, because with the difficulty of the shots he attempts, not even the best finisher in the world can sustain a significant overperformance. We have seen this on multiple occasions: last season he only overperformed his npxG by 0.03 per 90, and in 2020–21 he similarly overperformed it by just 0.04. However, as Vlahović is a very talented finisher and ball-striker, there will be periods in which he is able to convert some of these low-quality shots, and my theory is that this perpetuates an overall improvement of his level of performances through a boost of confidence. Because it should still be noted that it isn’t true that Vlahović’s underlying numbers rarely fluctuate, while his end product does. Furthermore, this shot selection-discussion also paints a hyperbolic picture of the shots Vlahović attempts, and it’s obviously not uncommon for him to get into really dangerous positions and attempt high-quality shots. As for why this occurs to a greater degree in certain periods, I would again theorize that this happens when Vlahović’s ability to finish is rewarded, which boosts his confidence and leads to him performing at a higher level overall. Right now we are seeing the most extreme example of this, but I think it’s inevitable that Vlahović is going to cool down significantly eventually, and then he’ll return to being an unreliable striker. You simply cannot rely on a striker with a shot profile comparable to that of Dušan Vlahović. On average, over an expansive sample size, such a striker will deliver to some extent, but it is also likely they will go through extreme dry patches, contrasted by highly impressive hot streaks.
Conclusion
At the end of the day, I think Vlahović is currently enjoying a purple patch, and I know that may sound harsh, but if we look at the bigger picture I think it’s a perfectly reasonable theory. First of all, something we haven’t delved into in this article is how hot and cold Vlahović has a tendency to be within a season, which is emblematic of his unreliability, and a problem which may not be illustrated through his output in the form of averages. This season is an excellent example of this, as Juve’s game against Frosinone in December was a huge turning point for Vlahović. Prior to this game, Vlahović had amassed four non-penalty goal contributions, 3.4 npxG, 0.5 xAG over approximately 9.6 90s; 0.42 non-penalty goals per 90, 0.36 npxG per 90, and 0.41 npxG+xAG per 90. Since then, he has amassed nine non-penalty goals, 5.6 npxG, and 1.1 xAG over a sample size of 8.0 90s; 1.1 non-penalty goals per 90, 0.70 npxG per 90, and 0.96 npxG+xAG per 90. The former sample size is actually slightly greater than the latter, and consists of really poor numbers. The latter sample size, on the other hand, admittedly consists of absolutely phenomenal numbers, and I think this almost epitomizes Vlahović, because it really is hard to overlook the latter sample size.
I, myself, have gone back and forth on my opinion on Vlahović countless times, but I really do think it’s just a matter of time before his numbers take a significant dip. Again, it wouldn’t be the first time. In 2020–21, it took until January for him to score his third non-penalty goal (he had previously amassed just two over a sample size of 11.3 90s). Similarly, he started off his 2021–22 season really poorly, scoring just two non-penalty goals over his first 9.9 90s, before scoring a hat-trick against Spezia and turning the season on its head. So, he does have his impressive periods, but overall I don’t think he’s anything too special on a European scale. This article has only touched on his final third contribution, and if you look beyond that there isn’t much that helps his case. For his size, I think his mobility and technique are relatively impressive and underrated, but I simultaneously think his technical consistency and decision making are both quite poor, which makes him fairly in that department. In the defensive phase I also believe he’s always been underwhelming. So, in Dušan Vlahović, you have a striker who is poor in the final third at worst; exceptional at best; solid on average, while being fairly average in buildup play, and a non-factor in the defensive phase. Ultimately, the final third is the most important area for a striker, so his relative prowess in this department should not be overlooked, but on a European scale I think this makes for an essentially average striker. When he’s firing in goals game after game, it may be difficult to visualize, but when he’s failing impact a game to save his life, I think his flaws are quite evident. He’s a strange player, and it’s nothing personal, but I don’t like him.
