Martinelli has failed to kick on, what does the data say is wrong?
He hasn't taken the next step we hoped he might, he is also not as useless as many would have you believe, what does the data say is going on here with Martinelli?
There is no player currently with Arsenal that has seen their stock fall further than Gabriel Martinelli. He was a player that was one of the crown jewels of a young team in 2022/23 and now there is frustration and anger among a fairly large portion of the fanbase to see his name in the first XI.
This summer has also been a big one where his place in the squad has the most uncertain and under threat since he and Emile Smith Rowe were competing for minutes. He won that competition, but it is far from certain that he will do it again.
Arsenal have spent big, bringing in three attacking players, including two in Eberechi Eze and Noni Madueke, that would be direct competition for his place, while the incumbent Leandro Trossard sticks around as well.
This is a fertile topic here at Cannon Stats, with Adam and me analyzing and opining on him numerous times over the last few years.
In December 2023, the first hints of trouble were brewing, leading to my first look into what was different for him in 2023/24 than his impressive 2022/23 season.
That next summer, Adam took up the mantel looking at the question that was starting to percolate in the fanbase on if Martinelli was still the best long-term choice and if he had slipped from being at the level required.
In October 2024, it was my turn again as Martinelli was off to a good start in 2024/25, regaining a bit of form.
This spring, it was Adam’s turn to highlight that Martinelli’s level of play after returning from injury had picked up pretty significantly. This is still probably something that is under appreciated as people have seemed to dig in on him.
The most recent article on Martinelli came again from Adam this summer. This time looking at him from the perspective of his overall productivity.
That catches up with the present.
In this piece I don’t want to retread too much on what has already been written previously but rather look more at the question of what has changed for him and if he has been able to adapt to these changes and still be able to put himself in the best positions to provide an impact for Arsenal.
The good place to start is looking at how the numbers have changed over the last four seasons for him.
This last season is interesting/worrying because it is one where none of his stat’s were at their best for this four-season span. The high-level numbers really hit this point showing that last year while not bad, was a down year for him from his previous level.
Open Play Shots, the best was 22/23, this current season is down 14% compared to the average the previous three years.
Non-Pen Goals, the best was 22/23, this current season is down 7% compared to the average the previous three years.
Assists, the best was 21/22, this current season is down 26% compared to the average the previous three years.
Open Play Key Passes, the best was 23/24, this current season is down 9% compared to the average the previous three years.
Dribbles Completed, the best was 21/22, this current season is down 42% compared to the average the previous three years.
Progressive Passes Received, the best was 23/24, this current season is down 21% compared to the average the previous three years.
It’s discouraging to see these declines (not every stat showed decline from average but still not great) and important to try and get an idea of how bad this is, what changed and why this decline/plateau is happening.
So onward we go into the details here for a deep dive into Gabriel Martinelli.
Shooting
The shots and goals are the area of Martinelli’s game that have (had?) the most potential to add superstar value.
This was the calling card for the player when he broke into the team, he played with an almost single-minded focus to find a way to get into the box and take a shot. The other big bonus for him was that this was not a player padding his stats taking most of his shots inside of the penalty area and doing so with a high average quality.
This is where seeing the volume of shots fall gets discouraging because if that does not return it is hard to see him be more than a good player.
He has maintained the good shooting locations, and the high average quality for his shots. He has even done quite well at converting those into goals (last season was only second to his 15 goal season on goal scoring rates) but the ceiling on what you can produce when you are taking 1 shot fewer every three games starts to add up.
His shooting rating reflects this; he was in the 90th percentile in his big season and dropped into the 66th percentile last year.
Looking at things a bit more granularly you can see that there has been a bit of a change in the shooting profile but nothing all that massive.
The big 2022/23 season with 15 goals was a season where he was +50% on his expected goals and even the best finishers don’t hit those levels over and over again consistently.
Overall, here the shooting locations are pretty strong, especially for a player that starts wide.
The majority of the shots are inside the width of the six-yard box and there is a big cluster inside of 12 yards. He does have a fair amount from distance and from tough tight angles on the left but he also able to tuck two of those away and put several on target so it would be hard to tell him to stop that (those are also the types of shots that pushed the goals above xG so much).
2023/24 was a tough follow up season for him. The expectations rose on the back of the 15 goals, and he just never really got it going.
He had two spells of injury that saw him lose three weeks each, one in the fall and the other in the spring. The injury absence in the spring especially seemed to hurt him as he was starting to come into some form and got piped down the stretch by a hot Leandro Trossard.
In the two months before his injury, he had 4 goals and 2 assists (over 1.1 G+A per 90) and was doing it on 0.6 xG + 0.6 xA with over 3 shots per 90 and 4.3 shot creating actions per 90. When he came back, he had just one start in the final 10 matches and that was on the final day of the season.
The big drop off here was that he went from +50% on his xG to matching it, even with solid numbers on the post shot xG and good on target numbers the big hot finishing didn’t continue (Trossard was instead the player that got that +50% season and pushed him to the periphery).
The shooting locations continued to look strong here, but he lost some of the big chances very close to goal inside of the six-yard box. He did still maintain his shot volume overall here and even cut out some of the real low-quality shots from distance and tight angles.
That brings us to the last season.
Here his goal scoring numbers picked back up again, and he was again over 0.3 non-penalty goals per 90 and did so on an xG that was just marginally worse than his 22/23 season (0.339 vs 0.345, going to three digits to show just how marginal it is). He even produced a similar level of extra boost from his shot placement adding 24% goal expectation in 24/25 compared to the 26% he did in 22/23, it just ended up where he only matched his xG and didn’t have the hot finishing season.
If the hot finishing is reversed and he gone 10 goals at 21 and then 12 goals here in an injury shortened season the story and narrative about him is almost certainly different but hey that’s the world we live in.
The granular story of the shots in this season is pretty interesting, the refinement of his shooting choices continued, and he is focused almost entirely in the “prime” scoring zones. He cut out almost entirely the speculative long shooting attempts, here is the percentage of his shots from different distances/zones in the Premier League over the last three seasons:
22/23: 46.3% in prime, 25.4% from the corners of the box, 20.9% from outside the box
23/24: 55.1% in prime, 28.5% from the corners of the box, 12.2% from outside the box
24/25: 50.0% in prime, 39.1% from the corners of the box, 8.7% from outside the box
The absolute elite wide players will be taking a bit more than a shot per 90 in those central zones in the box and Martinelli is just off of that taking 0.9 shots from open play in these spots.
The overall shot volume at just 2.14 puts him just below the 70th percentile range for wingers and the trend and overall decline from previous seasons is a bit of a concern; it is mitigated somewhat by his higher proportion of shots in more dangerous locations (85th percentile for shots in prime locations, 88th percentile for xG per shot).
It would make us all feel better if when we went to his FBref page that bar next to shots was green rather than grey and adding a few more shots would be helpful, even if they were a bit more speculative efforts is something that the team could use.
This transitions well into the next section I want to look deeper into and that is the change in where he is receiving the ball and his overall touches, and if this can explain the decline or suggest ways that he could bounce back.
Receiving
Martinelli’s receiving numbers have varied quite a bit over the last three seasons.
His 22/23 season saw him receive a slightly above average volume of passes for a winger, that increased in 23/24 before declining significantly last season. The two seasons previous saw him pushing elite levels for progressive passes received and final third entry passes received, with last season showing a notable decline.
He has maintained a high volume of touches in the final third and even in the area 25 yards from the center of the goal even with that decline from previous seasons but there is a noticeable change in the overall distribution between how wide he is now compared to what he was previously.
In his most prolific goal scoring season 22/23 13.7% of his touches in the final third came in the most central zone and adding in the half spaces that increases 46.7%. The last two seasons where his end product has been a topic of worry it has dropped to 8.2% and 35.2% in the 23/24 season and 9.6% to 39.1%. last season.
This isn’t a new diagnosis, and it seems pretty clear from watching him that he either has instruction to stay wide or has not been found by teammates or gotten himself into those more dangerous central areas at the same rates.
Getting the most out of Martinelli (or any left winger that Arsenal play) does likely go hand in hand with getting them more touches in those central locations. There are potential changes that we have seen in the squad that could make this possible, the team has signed Piero Hincapie and he has been a wide left player comfortable getting wide and overlapping (the incumbent left backs could be asked to provide this option more often as well). Viktor Gyokeres as a striker likes to peel out wide leaving space for the player on the left to fill central locations in those situations. There is the potential that Eberechi Eze could play in the left midfield spot and he is a player that would be comfortable filling the widest spots if the left winger moved inside.
Tying this back to shooting, there is a pretty obvious connection that central passes received in the final third are more likely to result in a player taking a shot and for Martinelli that is no different.
Here is where the passes that he took an open play shot were received over the last three seasons:
Only 13.5% came in the widest areas of the field with a big hot spot centrally and in the left half space at the corner of the box. This pattern holds even when you break it down by season.
A player needing to more central when they receive to generate a shot is not unusual but one of the things that separates out the good from the very good wide players is that ability to generate their own shot and this is just not one of the major strengths of Martinelli’s game.
Here is a new graphic that I created for this because this topic made it something I had to know about. This looks at a player’s ability to move the ball carrying inside from a wide position and generating a shot. This isn’t perfect and there are a few here that aren’t true “cutting in” type shots here but it’s a decent first attempt here and it does identify this issue in Martinelli’s skill set.
Of the 145 open play shots that he has had in the last three just 27 (and it is a generous 27) here fit the description of cutting inside from the wide spots that he so often finds himself in.
Here is what Bukayo Saka has done over the last three seasons from the opposite side to be able to receive wide and be able to generate his own shot through cutting inside.
Saka has been able to manufacture nearly a quarter of his own open play shots from his ability to receive wide and get himself into shooting locations (50 out 209 open play shots).
If Martinelli had been able to cut in and create at the same level as Saka that adds 9-10 shots. That doesn’t seem like a ton but that is again some of the minor differences between good and very good, Martinelli goes from 2.04 open play shots per 90 and that is solidly above average to 2.17 open play shots per 90 and potentially 1-3 more goals scored based on his shot quality and finishing.
I don’t know if at this age, it would be fair to expect a massive change to his ability to be able to create his own shot. This gives again a nice transition into looking at the dribbling and carrying numbers.
Dribbling and Carrying
When Martinelli first broke in at Arsenal he had electric pace and used that to terrorize the opposition full back. He had a bit of an unorthodox style of dribbling and it seemed to throw defenders off and give him the half second, he needed to be able to beat them.
This was never an elite trait for Martinelli and he excelled more in larger spaces where he could use his main weapon but it was quite effective and put him well above average compared to his peers.
Like many wingers before him it looks like he might have peaked at being able to beat his man in this early 20’s, Martinelli’s unfortunately might have come especially early in his 20s. Here is how his successful dribbles look for his career compared to other left wingers for each season.
The 23/24 season was a decline of about 10% and that’s not ideal but when the following season is a 37% decline you wish you could have gone back to that level.
You might be able to explain a bit of the decline, especially in attempts, to his hamstring injury where maybe he lacked some of that explosiveness. Overall, still it is not pretty; for the season he was not particularly effective pre or post injury at beating his man.
Before the injury he had 24 successful dribbles (1.5 per 90) out of 84 attempts (5.2 per 90) for 28.6% completion percentage.
After the injury he had 9 successful dribbles (0.97 per 90) out of 25 attempts (2.7 per 90) for 36.0% completion percentage.
The decline in success rate is a bit of a red flag here. This is typically a metric that gets better as a player gets more experience, but for Martinelli it has been the opposite thus far.
His dribble and attacking duel attempts don’t shed a lot of light on what specifically might be a problem either.
The majority come in the widest areas of the field and that is generally the areas of the field where dribble success rate is highest for a winger, when you are more central and closer to goal, it is less often that the player is one on one and that leads to more losses.
For Martinelli, even when he is wide, he is not able to get the better of his defender often enough these last two seasons.
Looking at his overall ratings for when the ball is at his feet, he rates pretty strong but with the caveat about the dribbles and total losses really sticking out.
Looking at the data presented here and from watching him play, it does often feel like the role that he is asked to fill at Arsenal is part of what leads him to much of what has led to frustration with the player.
When he gets the ball one of his go to moves is driving down the line continuing to stay wide.
There are times he can use his speed to gain enough of an advantage to be able to be able to turn the corner and angle towards the goal but often he ends up further up the field but not in an especially dangerous location where his threatening options are limited.
This flows nicely into looking at his ability to create for his teammates.
Creativity
I think Martinelli can deliver a good ball from wide but given above he too often finds himself in a poor situation and that ups the difficulty level. Crosses are already a tough way to create a chance, and they don’t need that extra degree of difficulty added. Here are the open play crosses from Martinelli over the last three seasons.
One of my areas of frustration is that he will get into the area near the corner flag and try to play the ball in and you already have the feeling the chances of something good happening is low.
Here are just the crosses from the left corner highlighted:
He is completing fewer crosses than expected here but he does actually have a bit better than average chances quality generated where he is 0.022 xA per cross compared to the average of all crosses of 0.013 per open play cross. This is a bit confounding and my guess at explaining it would be that Martinelli is not a bad crosser when the opportunity is there (better chance creation than average), he just too often will attempt a hopeless cross because he ran out of other options (poor completion percentage).
Looking at his crossing in even further detail there are some interesting things to pull out.
Most of his crosses come in the areas where he completes them at lower-than-expected areas.
Trying to infer this, but it would seem that his right footed crosses (or areas he is more likely to be able to use his right foot) are stronger than the areas where he would use his left foot. I would love to see how this looks on weak foot vs strong foot but that is unfortunately not in the data that I have here.
He might need to reconsider what the best options are when he gets deep near the touchline/edge of the box area because trying a cross is not working as an effective way to generate chances (outside of winning corners which given Arsenal maybe washes out)
One area that I think he could get better at but would still be hard given that he is an inverted player is turning this type of possession into cutbacks.
He has only been attempting 0.4 cutbacks per 90 and this is an area where he can up the volume to generate more good looks for teammates.
Pulling out a bit deeper now looking at his overall creative numbers, this is an area where again he rates out above average to quite good.
He lags behind a bit on his teammates finishing the chances he creates but overall, he has good volume of completing passes into dangerous areas, key passes and the expected value from these.
Here is the overall view of his key passes over the last three seasons.
The big thing that I notice here is that his numbers are hurt significantly that the passes he has completed to teammates have not been converted at the rate that the quality of chances would have suggested. He has just 9 passes assisted by passes (he has more that have come off of rebounds or were deflected by a defender) from an expected 18.2.
His overall creative numbers have been consistently quite good.
The 2023/24 season showed a decline from the season’s prior, but it is to a ‘good’ level rather than a ‘very good’ level.
Final Thoughts:
I had kind of hoped to come to the end of this post and have a solid crystalized take to be able to sum up and I just don’t here.
On the one hand, it is frustrating and concerning that Martinelli has seen his overall level decline but has also had patchy form. On the other hand, the view of what his level is in the wider fanbase is way out of whack with what the data shows and he is still quite a good player.
If you are wanting to look at things generously looking at the last three seasons shows that this is a damn good player.
He is still in his prime and the sport has seen plenty of examples of players that has some rough patches figure things out a bit more at this age and go on to have long productive careers. I think that it would be unwise to project him to “World Class” levels but it wouldn’t be a crazy outcome if he was a very good wide attacker and the type of player that can be a secondary attacker who isn’t the star but plays a lot and is a major contributor on a Premier League or Champions League winning team.
If you want to be less generous you can focus more on the trend line of the production. In 2022/23 he was at that star level production, that fell to solid Champions League level starter and then fell further to a player that is above average Premier League level starter. The next step here is a player that isn’t even that and like many young players before him shone bright early but burnt out too quickly.
I have been playing around with ways to try and guestimate and project future performance (I have lots of inspiration on this from Dom Luszczyszyn and the graphics he makes for Hockey), using the players past performances and how players in that position have improved/declined historically.
The error bars on the 2025/26 season are pretty large, he has big production in the recent past, his trend is not pretty. He’s also in his prime and that would a bit of anomaly for the “decline” phase to start. That gives you a spot where trying to guess what comes next is tough.
If he ends up bouncing back to something close to star level, I would not be shocked by that. If he ends up looking like a player that can’t start at the highest level of the Premier League, I would be disappointed but also not shocked.
I am still a believer in the talent here and I hope that he has what it takes to show the doubters that they were wrong about him, but it is getting to the point that he will have back that up on the field now.
Clearly Arteta values him. Perhaps for his defensive work, perhaps because he stretches the field by staying wide. I don't know. I do think that this season is a critical season for him. Will he learn to produce an end product with all those very productive players around him? WIll he learn to drive into the area and take on his defender that way, the way Noni does? Will he find Big Vik in the box more now that we have a real box poacher? All questions. As for me, I like him, and I really hope he comes good esp this season. We need him to be effective no matter if he starts games of finishes them.
I do not know if he is going to be a role player or a regular starter going forward. But he is outstanding defensively, which Arteta values, and he provides a speedy attacking wide player, which contrasts with Eze/Trossard. Ironically for all the discussion of Martinelli, I am still convinced that Trossard is the player who should have left. I suspect that Eze is going to take all the minutes that could have been Trossard, as a skilled player, moving inside, and strengthening build up play. Martinelli is going to (probably) continue to be the defensive emphasis wide hug the line attacking player. Trossard did get some minutes against Forest, but looking forward, to me it seems like Arsenal will have Eze and Martinelli on the left as two contrasting players and Madueke and Saka on the right. All 4 players can swap sides (as saka started to do more last season). As long as Eze is healthy, I think Trossard is going to get very few minutes outside the League Cup. And Martinelli will continue to get minutes when Arteta (at a minimum)wants to strengthen defensively to control games in the last 30 minutes.