Why Celtic should replace Joe Hart with Peter Vindahl Jensen

The club have been linked with a host of keepers since Hart announced his intention to retire earlier this year – Newcastle’s Martin Dubravka was one of the latest names linked while Caoimhin Kelleher and Ugurcan Cakir have also been mooted as potential options over the last few months. Interest in former Danish under-21 international Vindahl Jensen first emerged last week though, with the Daily Mail now stating the former AZ Alkmaar keeper has moved to the top of the club’s list of candidates to replace Hart.

26-year-old Vindahl Jensen, who made 47 appearances while on loan in Prague last season, only finalised a permanent move to Sparta from AZ in March and the Czech side is said to be holding out for a fee of more than £4m for a player they have only just added to their ranks permanently.

Should Celtic come to an agreement with the Fortuna Liga champions for Vindahl, what kind of keeper would they be getting? Here, using firstly Wyscout data and then StatsBomb data, we take a closer look at Vindahl, who started his professional career in his native Denmark with Nordsjaelland and had a spell in Germany with Nuremberg.

As mentioned with previous data analysis pieces on goalkeepers linked in the past few months, data never gives a full picture when analysing a player, even more so for goalkeepers who have much more limited data points compared to outfield players. Goalkeeper data is also more heavily influenced by team performance than any other position. However, there are still some key metrics that can be used to evaluate goalkeeper performances and give an insight into their profile.

Shot-stopping

Arguably most important is shot-stopping, how well is a goalkeeper performing when it comes to their fundamental role, keeping the ball out of the net?

As can be seen in the above scatterplot, Vindahl was amongst the best shot-stoppers in last season’s Czech top flight. Given Sparta’s domestic dominance, Vindahl faced the third-lowest xG per 90 of all goalkeepers in the Fortuna Liga (0.94 per 90).

His conceded goals per 90 was even lower though, at 0.84 per 90, giving Vindahl a positive prevented goals value of 0.11 per 90. This was worth just under four goals less conceded to Sparta across the season than would have been ‘expected’ and likely crucial in what was a tight race for the title with rivals Slavia.

One of Vindahl’s strongest shot-stopping domestic seasons to date, Wyscout records him conceding slightly more than expected across his whole career to date, conceding an average of 1.37 goals per 90 from 1.31 expected conceded goals per 90.

Claiming/sweeping

The next area we can use data to provide insight into Vindahl is claiming and sweeper actions. Firstly, the data shows he had the lowest exits (0.89 per 90) of any goalkeeper in the Czech top flight. Wyscout defines exits as an attempt on the part of the goalkeeper to actively play a high cross or a long aerial pass in the air, either to claim or to punch the ball.

For aerial duels, any claim that an opponent challenges, Vindahl comes in below average for the volume he is involved in (0.25 per 90). Both of these low numbers can be attributed again to Sparta’s general dominance with Vindahl simply less likely to be involved in these compared to his positional peers who were ‘busier’ during games.

That said the below graphic is a bit more revealing when it comes to Vindahl’s leaving his line tendencies with his actions per crosses faced and per set-pieces faced, the lowest of all goalkeepers in the league. So, although not involved in many, this further detail suggests that when he is faced with these balls in his area he is not overly proactive.

The next graphic gives a bit more insight into Vindahl leaving his line outside the box.

Here we can see that likely down to Sparta playing a high line, Vindahl was pretty active outside his box. His 1.92 sweeps per 90 was one of the highest amongst Fortuna Liga goalkeepers in the last year.

This graphic also highlights that Vindahl has tended to make short passes when he did find himself in these positions outside his box. He has been more effective with these shorter passes too, completing 92.9 per cent of his attempted short passes outside his box compared to his long passes (61.9 per cent). Given Celtic also play a high line, especially domestically, this is encouraging to see.

Lastly on this graphic, and leading into the next section, we can see Vindahl has received a high number of passes per 90 (17.42 per 90) compared to his positional peers in the Czech Republic’s top flight in the last year.

Distribution

Brendan Rodgers has continued to utilise the goalkeeper in build-up phases during his second spell in charge so effectiveness on the ball has likely been a key consideration when identifying the club’s next number one.

As seen towards the end of the last section, Vindahl has been pretty active in possession during his spell with Sparta – they averaged 57.1 per cent possession last season, the second-highest in the league.

The first distribution scatter here shows how Vindahl performed when it came to short/medium passes. Again, heavily influenced by team style, we can see he attempted an average of 18.17 of these types of passes per 90 with his accuracy rate of 97.55 per cent coming in just under the average mark for Fortuna Liga goalkeepers last season.

He attempted one of the lower numbers of average long passes (7.27 per 90), highlighting Sparta’s tendency to favour shorter passes over longer ones when building up from the back.

Vindahl’s accuracy rate was not as impressive at these longer distances (62.84 per cent). However, given Celtic tends to go long from the back even less, Hart averaged just 2.89 long passes per 90 in the league last season, this would not be too much of an issue, should Vindahl make the move to Scotland this summer. The graphic below shows Vindahl’s distribution in greater detail, mapping all of his passes from goal kicks and open play across all distances in the last calendar year.

His goal kicks were mostly short, 110 of his 222 goal kicks were within his own third, and he was pretty efficient here, only failing to find a teammate with two of the 110. His goal kicks to the middle and final third were less frequent, and in line with the scatters, his accuracy rate is lower than most other goalkeepers in the Czech top flight at these longer distances.

Open play passes paint a similar picture, again much more passes shorter, to decent effect, and fewer at longer distances, which he was less effective in. The open play passes graphic highlights a high number of quick breaks too, the second highest per 90 of any goalkeeper in the league in the last 12 months, indicating Vindahl is good at quickly getting attacks started when he has the ball in hand.

Overall, his xGBuildup, a metric that assigns a value to those involved in deeper-lying playmaking or buildup play that leads to shot, is by far the highest of any goalkeeper in the Fortuna Liga in the last year, underlining Vindahl’s effectiveness when it comes to contributing to build-up play from the back.

StatsBomb Europa League data

Vindahl played 14 matches in European competition with Sparta last season, most of which came in the Europa League, with StatsBomb data providing some insight into how his domestic performances scaled a higher level in European competition.

Firstly, as we can see in the above scatter, Vindahl struggled to replicate his shot-stopping on the European stage. Sparta of course had a difficult campaign, which ended in a heavy loss over two legs to Liverpool in the last 16, and Vindahl faced the second-highest post-shot xG of all goalkeepers in the competition (minimum 300 minutes played). 

His goals saved above average though, StatsBomb equivalent of prevented goals, was -4.8, the fourth lowest of the 36 goalkeepers that played more than 300 minutes. So, although expected to concede a high number of goals, StatsBomb records Vindahl as conceding almost five goals more than he ‘should’ have as well.

Vindahl’s radar further highlights his struggles on the shot-stopping front in the Europa League. Some of the other metrics on his radar here are consistent with what was seen in his domestic data though such as holding a high position to sweep – he came in above average for Statsbomb’s ‘goalkeeper aggressive distance’, an average value of how far from the goal a goalkeeper is coming forward to perform defensive actions.

It is a bit more positive on the distribution front too with the data provider attributing Vindahl with a high number in the ‘positive outcome’ metric, a possession that connected through the player and resulted in either a shot or free-kick in the attacking half or a corner. Again, this is consistent with his high xGBuildup in the league.

No doubt heavily impacted by his shot-stopping, StatsBomb doesn’t score Vindahl’s overall contribution in last season’s Europa League too favourably at all, highlighted by his low goalkeeper OBV (-0.37 per 90). StatsBomb’s goalkeeper OBV assigns a value to a goalkeeper’s goal-preventing actions – such as shot-stopping, cross-claiming, and sweeping actions. Ultimately an indicator of whether a goalkeeper is making a net positive contribution to their team or a net negative one.

Conclusion

Vindahl looks as if he would be a solid number one for Celtic. At 26 years old, he is approaching peak age for a goalkeeper which offers the potential to provide stability between the posts for several years to come. 6’5” - the same height as Hart - he would provide a big presence and should cope fine with the physical aspects of Scottish football. Domestically, he would be more than steady when it comes to shot-stopping and would be capable of contributing effectively to build-up from the back.

Although not an out-and-out sweeper-keeper he has recent experience of sweeping, to good effect, in a possession-dominant side that holds a high line too, making him a good fit in that regard. He would likely not be overly proactive in leaving his line to deal with crosses though.

The biggest question mark though, after a difficult Europa League campaign, is how he would perform at the Champions League level. That would remain to be seen but, in a domestic setting certainly, it seems Vindahl would be more than capable of taking on the role of Celtic’s number one.