Tuesday, January 24, 2017

Can City's Twente Contingent Break Through Next Season?

When Pep Guardiola was brought in to Manchester City, one of the areas he might have been expected to improve on his predecessor Manuel Pellegrini was in bringing through young talent to the first team. From the La Masia core of his Barcelona teams to his promotion of Coman and Kimmich at Bayern, he has consistently been able to incorporate youth into his sides. However, the fact remains that only Aleix Garcia has broken through from the academy this season in the Premier League and Kelechi Iheanacho, who solidified a first team spot in the prior season, is on pace for under 1000 minutes again. Though this may change as the relatively old squad heads down the stretch, it does seem that there will be less of a youth movement this year than previously thought. That said, some of the brightest youth talents at City are currently abroad on loan, and thus maybe next year is the time for their assault on the Premier League. In particular, City have three players in Holland's Eredivisie with FC Twente who might add something to the squad next year: Bersant Celina, Enes Unal, and Yaw Yeboah. In order to see how they would potentially fit at City next season, I took a look at their Usage Rates so far this season at FC Twente, as that can be a helpful guide to assessing the role they might play.

Below are the Usage Rates for Twente for the first half of their 2016-17 campaign (limited to players with 500+ minutes):

Player Successful Passes Total Passes Unsuccessful Passes KP A G S Unsuccessful Take Ons Minutes Possessions Used Usage Rate Positive Negative % Positive
Unal 179 283 104 8 2 10 56 24 1204 194 14.43% 66 128 34.02%
Klich 632 766 134 19 0 3 16 4 1170 173 13.25% 35 138 20.23%
Celina 349 422 73 19 1 4 36 23 1066 152 12.77% 56 96 36.84%
Andersen 234 304 70 1 0 0 4 0 576 75 11.66% 5 70 6.67%
Yeboah 316 403 87 17 0 2 13 34 1188 151 11.39% 30 121 19.87%
ter Avest 357 489 132 10 0 1 13 8 1310 163 11.15% 23 140 14.11%
Mokotjo 941 1065 124 14 1 1 5 2 1374 146 9.52% 20 126 13.70%
Trajkovski 180 258 78 4 3 0 6 6 982 97 8.85% 13 84 13.40%
Ede 261 321 60 7 4 0 8 3 836 82 8.79% 19 63 23.17%
Thesker 577 676 99 5 2 0 0 0 1205 106 7.88% 7 99 6.60%
Bijen 723 831 108 6 1 0 9 2 1530 126 7.38% 16 110 12.70%
van der Lely 258 310 52 8 0 0 0 0 833 60 6.45% 8 52 13.33%
Marsman 223 301 78 0 0 0 0 0 1530 78 4.57% 0 78 0.00%

As you can see, from an attacking perspective, Twente are basically Man City junior. Unal, Celina, and Yeboah are all in the Top 5 in Usage Rate and have played a large percentage of available minutes. Twente typically play a 4-3-3, so often the three City youngsters are in the same front line. That said, it's pretty clear that Unal and Celina are the better prospects. Both have a higher Usage Rate and are able to use their possessions more efficiently than Yeboah. Unal's stats look pretty close to Kelechi's last season (13.2% Usage Rate, 33% of possessions used with a positive result) and Celina's similar to Nolito this season (13% Usage Rate, 34% of possessions used with a positive result). Obviously, Celina and Unal could be valuable members of the squad if they could match the production of those two.

However, while the stats for Unal and Celina this year at Twente are good, it's important to not just take these at face value. There's a strong consensus that the Eredivisie is a lower level of competition than the Premier League, backed up by relative results of the two leagues in European competition. It makes sense therefore to adjust the numbers to account for the relative difficulty of the two leagues. In order to attempt to do this, I looked at the Usage Rates for all players transferred from the Eredivisie to the Premier League in the past three years that played at least 500 minutes in the seasons before and after their transfer (unfortunately, I don't have more historical data available to me). That left me with a sample of just ten players, admittedly way too small to draw any satisfying conclusions. Still, there were some things that jumped out from the data below:


Season of Transfer Player Old Team New Team Eredivisie
Usage Rate
Eredivisie %
Positive Possessions
PL Usage Rate PL % Positive Possessions PL Usage Difference PL % Positive Difference
15-16 Depay PSV Man United 19.12% 41.73% 12.54% 30.86% -6.58% -10.87%
14-15 Pelle Feyenoord Southampton 17.23% 34.30% 14.74% 27.78% -2.49% -6.52%
14-15 Tadic Twente Southampton 17.13% 37.21% 13.23% 36.24% -3.90% -0.97%
16-17 Janssen AZ Tottenham 13.57% 43.86% 8.18% 49.23% -5.39% 5.37%
15-16 Clasie Feyenoord Southampton 11.69% 18.25% 9.37% 16.47% -2.32% -1.78%
14-15 Blind Ajax Man United 10.79% 19.76% 10.13% 18.45% -0.66% -1.31%
14-15 Janmaat Feyenoord Newcastle 10.75% 18.42% 12.18% 14.15% 1.44% -4.27%
15-16 Wijnaldum PSV Newcastle 9.93% 39.55% 8.75% 31.45% -1.18% -8.10%
16-17 van der Hoorn Ajax Swansea 9.88% 9.15% 10.77% 5.19% 0.88% -3.96%
15-16 Martina Twente Southampton 9.34% 6.90% 9.84% 11.90% 0.50% 5.01%

A drop in Usage Rate makes intuitive sense, as playing with better teammates is likely to mean a player's share of the ball is going to drop, and we do see that to some extent in the data. However, due to the fact this is something we see in basketball when players from lower leagues move to the NBA, I thought the effect would be stronger than what we're seeing here. In fact, all three defenders actually see a slight increase to their Usage Rates, which was somewhat surprising. This may be due to the fact that certain positions in football are inherently more likely to have higher Usage Rates given their proximity to the goal, something which isn't true in basketball. As a defender, since most of their Usage Rate is driven by turnovers rather than shots or key passes, it would actually make sense that being in a tougher league would increase your turnovers, and thus your Usage Rate. Also, as this excellent Statsbomb article states, team composition likely has an effect on the Usage Rate as well, and that's something I've seen in the data on Premier League transfers. Unfortunately, without having a larger sample of players there isn't enough to tease out the league effect (if there is one) versus the team effect versus the positional effect, but it is interesting to consider. The drop in the percentage of possessions used with a positive result is smaller by comparison, but the trend is much more clear as this chart shows:



This is probably due to the fact that the Eredivisie consistently features more shots per game and fewer passes per game than the Premier League, so the percentage of possessions that result in shots, key passes, or assists are inflated slightly.

So what does this tell us about how Celina and Unal would look at Manchester City right now? If we used a linear regression of the albeit limited data, we'd expect Celina's Usage Rate would drop to about 11% and the percentage of possessions used with a positive result would drop to 33%; Unal's Usage Rate would be about 12% and the percentage of possessions used with a positive result would be 31%. Again, given the sample size, these are extremely provisional numbers, but they can serve to give us a ballpark idea of what we might expect from the player in the Premier League, and at least to attempt to account for the difference in league. If they do in fact hold, Celina's Usage Rate and % of possessions used with a positive result would be very close to that of Jesus Navas (11% and 28% respectively). Given Navas is probably out the door in the summer, Celina would seem to be a perfect replacement that would a) cost nothing and b) almost certainly provide more end product given his xG/xA numbers. Unal is tougher to find a role for, given City currently play with just one striker and his translated numbers aren't as good as Aguero's (11% Usage Rate and 47% of possessions used with positive result) or Kelechi's (13% and 43%). For him, I would think a loan spell somewhere in the PL or Championship might be the best bet next season.

Of course, a key thing to keep in mind is that Celina, Unal and Yeboah are still under the age of 21, and it's certainly possible their numbers will improve. Playing at a good level in a decent league is an accomplishment in and of itself, and it is obviously hard to break into a team that is still one of the best in the world, recent results notwithstanding, at such a young age. Unal and Yeboah may get there yet, but in Celina I think we have a player who will be playing at Manchester City sooner rather than later.