x icon

Our knowledge is at your service

FiveThirtyEight – the probability forecasting site – predicted ahead of the season that Manchester City would finish top of the table, Liverpool were expected to finish 2nd, Chelsea 3rd and Manchester United 4th. The order of the top three was 100% correct, but as for the final Champions League position, it was a surprise to see Antonio Conte’s Spurs finish 2 points ahead of city rivals Arsenal and 13 points ahead the Red Devils, who have undoubtedly had better days. In the first chapter of our two-part analysis, we will attempt to answer the question of how each team performed in terms of their attacking play using both classic and advanced statistical indicators. Part 2 will assess the quality of the defence.

Goals scored and the distribution within the team

Excluding own goals from the calculations, Tottenham scored the most goals of the three teams (63), while the fact that Leicester City (62) scored more than both Arsenal (60) and United (56) says a lot about the attacking potential of the three teams. In terms of chances created, however, the Red Devils were the only team to have overshot their xG (+1.5), as both Spurs (-1.67) and The Gunners (-8.88) fans could rightly feel that there were still a few more goals left in their players’ legs. However, it looks pretty bad that the 20-time English champions were also beaten by Leeds United (55) and West Ham United (58.06) in expected goals scored (54.5), which is presumably not something that has happened very often in PL history. Now let’s take a closer look at which players proved irreplaceable.

Manchester United’s signing of Cristiano Ronaldo last summer was presumably done in the hope that the Portuguese superstar would make the Premier League trophy a real possibility. CR7, with 18 goals from a xG of 15.69, has produced what he was signed for. It’s another matter that there are certain aspects of his game – defensive work, number and intensity of pressures – that may not be acceptable if his current team is fighting for serious goals, although Manchester United were probably aware of that. So 32% of United’s 56 league goals came from the five-time Golden Ball winner, with Bruno Fernandes coming second with 10 goals. . His finishing was also good as he scored those 10 goals from an xG of 8.76. Together, the two Lusitanians accounted for no less than 50% of the Red Devils’ PL goals, a staggeringly high percentage. Also, an indication of the level of exposure of the team’s attacking play, or at least the ability of these two players to convert chances into goals. When they weren’t on their game, the burden of scoring was on the shoulders of Mason Greenwood (5), Marcus Rashford (4), Fred (4) and Jadon Sancho (3), but as their numbers suggest, they didn’t really live up to expectations. Having said that, the 4 players have managed to score 16 goals from 11.54 xG situations in aggregate. So, we’re not sure if fingers should be pointed in their direction, especially in case of the Brazilian midfielder, whose primary role in an well-functioning team would certainly not be scoring goals.

While MU had 5 players who scored at least 4 goals in the 2021-22 season, Arsenal had 8 players who surpassed this level, so the distribution of goals within the team was much more even. If we look at chance creation, we can see an even more drastic difference, as the Red Devils had only 3 players with a minimum xG of 4 at the end of the season, while the the Gunners had 9. Most goals were scored by Bukayo Saka (11), Emile Smith Rowe (10), Martin Ødegaard (7) and Gabriel Martinelli (6). Gabriel Magalhães (5) did really well too , as chipped in with goals from free kicks and corners. There are four players who deserve a special mention, one of which is ESR, still only 21, who has outscored his xG by about 4.5 goals, which, although a stunning effort, is unlikely to be maintained in the long term. Also to be applauded is Eddie Nketiah, who was getting more playing time at the end of the season (and has since signed a five-year contract with the club), and who needed only 948 minutes to reach 5 goals. The sample size is small, but if he can continue to average a goal every other game next season, he could pose a serious dilemma for Mikel Arteta when planning his attacking line-up. At the other end of the scale, Pierre-Emerick Aubameyang, who signed for Barcelona in the winter, and Alexandre Lacazette, who moved back to Lyon recently , have both scored 3.95 and 4.79 goals fewer respectively, than an average quality top-flight striker would have done from the same chances.

If the problem for United was the insane proportion of goals scored by the two Portuguese attackers, the statement is even more true for Tottenham: 63.5% (!) of the goals scored in the league last season were scored by either Son Hung Min (23) or Harry Kane (17). Without their goals, Spurs would have averaged 0.61 per game, a terribly low figure. The South Korean is without exaggeration going through his most prolific period of his career, because he never scored so many occasions in a single campaign. The previous record was 17 goals in 2020-21, topped by 10 assists. But what is quite unique about his numbers in recent months is his relentless rate of taking advantage of his opportunities, as he has surpassed his xG of 14.38 by 8.62. Also playing a key role in making the attacking play more fluid has been Dejan Kulusevski (joining from Juventus midway through the season), who has netted 5 goals and gave 8 assists in the spring: a highly impressive debut in the world’s most powerful league. If the Lilywhites are to dream even bigger next season, they will need a much bigger contribution from the rest of the squad in front of goal, as unexpected injuries can happen at any time, and it is far from certain that Son will always be so effective in converting those chances.

The best creators

In addition to the number of goals scored and the conversion rate, it is equally important to look at which players excelled in serving their teammates and distributing the final pass. Staying with Spurs, Son (8.56) and Kane (5.08) created by far the most xA, but the high-flying wingers also played their part, with Sergio Reguilón (3.31) and Emerson Royal (2.95) also among the top five in the expected assists chart. Kulusevski’s 8 assists have been made from a xA of 2.89, but it is hard to expect this ratio to remain the same on a larger sample. Lucas Moura has assisted 6 times for Tottenham, the fourth most in the team. After the English striker (24) and the Brazilian attacker (18), Pierre Højbjerg (17) has the most successful through passes, but Harry Winks (13) and Eric Dier (10) have also contributed. By far the most successful key passes came from Kane (24), but the aforementioned wingers and forwards both had over 10 for the season as a whole, as did Ryan Sessegnon (10), who got only 1105 minutes to prove his worth.

For city rivals Arsenal, creativity was mainly provided by Saka, Ødegaard and Martinelli, who stood out not only in terms of assists (7, 4, 6) but also in terms of xA (6.89; 5.79; 4.48) and key passes (20, 19, 16). It’s safe to say that the Norwegian genious was the Gunners’ most remarkable attacking midfielder in the 2021-22 season: he had the most through passes with 18, while his 19 key passes were the 2nd most in the team. In the former category, Granit Xhaka (14) is up there, and he can also boast 8 key passes, along with Lacazette.

For the Red Devils, only three players have created at least 3 xA. Sancho finished the season with 3.2 xA, 3 assists and 16 key passes, which are not elite figures compared to players with same numbers in rival teams. However, the 22-year-old English winger, who arrived for 85 million euros, was still one of United’s most creative attackers. Only Fernandes (7.67) created more chances for his teammates than Luke Shaw’s 3.66 expected assists, and it is astonishing how different his level of football was from the others. The Portuguese virtuoso’s 33 through passes were three times (!) more than the next closest man, Diogo Dalot (11), but the 31 key passes he dished out were also unmatched. So if the portuguese had an off day or was taken out by the opposition, there was often a gaping hole in United’s attack.

Other important metrics

Last but not least, we will look some other relevant statistical indicators that give us a deeper insight into which players have provided an extra edge in the composition of attacks. In addition to key passes and through passes, Fernandes was the absolute king of progressive passes (297) and successful passes to the final 3rd (185) for the Red Devils, but Harry Maguire (232) also performed well in the former and Fred (161) in the latter. Scott McTominay (103) was also brilliant in ball recoveries in the opposition half, alongside the Portuguese (97) and Brazilian (104), while Ronaldo (150) was the best in touches inside the opposition’s penalty area and Sancho (124) was the best in progressive runs.

Gabriel (248), White (242) and Xhaka (222) were the main progressive passers for the Gunners, but they also made the progressive passes into the attacking third, along with Thomas Partey (160). In the opponent half, the Swiss midfielder has made 114 recoveries this season, the Norwegian midfielder 110, while Saka and Martinelli have typically excelled in touches inside the penalty area (194 vs 101) and making progressive runs (134 vs 69).

Among Spurs players, the trio of Højbjerg, Dier and Ben Davies also stood out in terms of progressive passes (306, 279, 247) and passes distributed to the final 3rd (290, 140, 147). The Danish midfielder has had by far the most ball recoveries in the opposition half (147), while the highest number of touches were made by Kane (149) and Son (132) inside the box. When it comes to the volume of progressive runs, the South Korean striker (81) and the Welsh wing-back (67) deserve the highest praise.

At the end of the analysis, it’s hard to leave out the assessment of the Expected Threat model, which quantifies the extent to which individual players increase their team’s chances of scoring goals through their various moves – or, with other words by getting the ball from point A to point B. Taking all types of actions into account, the greatest threats from the Gunners’ whole squad for the entire 2021-22 season were created by Saka (8.7), ESR (8.3) and Ødegaard (7.3). For Manchester United, only Fernandes (9.2) produced higher than 5 xT, followed by Fred (4.3) and Sancho (3.6). As for Tottenham, Son (12.3) created the most threat, followed by Højbjerg (4.4), Kulusevski (4.0) and Kane (4.0). Taking only passes into account, the Lilywhites have Dier (3.6) and Davies (2.7) alongside the Dane (3.8), MU have Fernandes (4.0), Alex Telles (2.6) and Dalot (2.2), and Arsenal have Xhaka (3.4), Cédric Soares (2.7) and the White-Gabriel duo (2.4) leading the cumulative xT index.


Analysing the teams’ attacking play gave us a comprehensive picture of which players in each role provided the extra spark that fundamentally determined the effectiveness of the team. At Manchester United, Ronaldo scored every third league goal, at Tottenham the Son-Kane duo (23+17) accounted for almost 2/3 of their PL goals, while at Arsenal only Saka (11) and ESR (10) reached double figures. In terms of creating chances and getting the ball into dangerous positions, Højbjerg and Kulusevski stood out for Spurs, in addition to the former two stars, while for The Gunners, Ødegaard and Martinelli were the standout players alongside the young English winger. The Red Devils’ number one conductor was once again Fernandes, who, without exaggeration, was practically responsible for almost all the progression and creating good quality chances.

(Cover photo: )