Spencer Mossman

Variations of "team share" metrics for counting stats are used in prospect and in-season usage analysis for high-stakes fantasy football (NFL), and I believe it has equal utility in the football world as well.

(8/8)

"Team share" neutralizes both of these external influences to a strong degree by looking at the volume metric within the team's ecoystsem, rather than as an individual, out of context data point.

(7/8)

2) Possession: similar to team tactics, a player's counting stats are heavily influenced by how much of the ball their team sees.

(6/8)

1) Team tactics: a manager might emphasize progression through passes rather than ball carrying, in which case, a player's carry volume would be deflated (and passing accordingly inflated) relative to what they might experience on another team.

(5/8)

I used "team share" (of XYZ metric) because it accounts for a couple different factors that the traditional statistics do not:

(4/8)

#Arsenal is extremely reliant on Bukayo Saka to progress possession from the flanks. His ranks for team share of progressive actions:

7th: Carries (20.45%)
24th: Passes (6.88%)
2nd: Receptions (25.69%)

📊 Carries x Receptions

(3/8)

#AFC #PremierLeague

Saka's rank, compared to other #PremierLeague attacking midfielders, wingers, and forwards, in terms of team share of chance creation metrics:

🥇: Open Play Shot Creating Actions (18.50%)
🥉: Assists (37.04%)
🥉: xA (26.79%)
4th: xAG (25.33%)

📊 xA x Open Play SCA

(2/8)

#AFC #Arsenal

Bukayo Saka is the most relied-upon player in the #PremierLeague.

He has generated 18.5% of Arsenal's open play shot creating actions, most in the league.

There is nobody in the squad, nor available via transfer, that can replace his contributions.

Mini 🧵

#AFC #Arsenal

Conclusion pt. 3

Maybe the best thing for his future is to stay at Frankfurt, but if I was a mega club looking for my next no. 9, I wouldn't be in the market for the likes of names such as Benjamin Šeško, Evan Ferguson, or Jhon Durán.

I would be first in line for Hugo Ekitike.

#SGE

Conclusion pt. 2

The comps of his numbers since his Reims days have followed this pattern: promising ➡️ awful ➡️ pre-breakout ➡️ actual breakout.

#SGE

Conclusion pt. 1

Hugo Ekitike is having a historic year for a center forward so far this campaign.

If the collective football world had paid attention to the pattern behind his underlying metrics, the breakout would have been obvious.

#SGE

Hugo Ekitike In Possession pt. 4

The comparable datapoints:

- '21/22: Nico Williams ('23/24 & '24/25)
- '22/23: Josh Sargent ('21/22)
- '23/24: Alexander Isak ('22/23)
- '24/25: João Pedro ('23/24)

#SGE

Hugo Ekitike In Possession pt. 3

It is important to evaluate something like pass % through the lens of positional peers because the further forward a player sees the ball, the higher density of defenders to space.

#SGE

Hugo Ekitike In Possession pt. 2

To illustrate, the five center forwards with the highest npxG/90 in Europe's top 5 leagues this season, sorted by passing %:

🇫🇷 Hugo Ekitike: 75.0%
🇵🇱 Robert Lewandowski: 73.5%
🇳🇬 Victor Boniface: 72.4%
🇳🇴 Erling Haaland: 68.4%
🇳🇴 Alexander Sørloth: 64.2%

#SGE

Hugo Ekitike In Possession pt. 1

As he has matured, he has turned into more of a 'strength of pass' forward, rather than a ball carrier.

#SGE

Hugo Ekitike Goal Contributions pt. 3

The same pattern emerges as before... a solid comparable data point, a pretty bad comp, a comp to a player pre-breakout, and then a historic data point.

#SGE

Hugo Ekitike Goal Contributions pt. 2

The comparable datapoints:

- '21/22: Jonathan David ('21/22)
- '22/23: Lyle Foster ('23/24)
- '23/24: Phil Foden ('21/22)
- '24/25: nobody

#SGE

Hugo Ekitike Goal Contributions pt. 1

Same graph as before (poss. adjusted xG marks), but for U22 seasons.

Only Kylian Mbappe ('21/22) and Bradley Barcola ('24/25) join Ekitike as forwards in the 90th+ percentile for both.

Ekitike is by himself in terms of players in the 95th+ percentile.

#SGE

Hugo Ekitike: The Improvement pt. 3

As it stands, Olise's campaign last year is the best season with 900+ minutes in my model, regardless of age (slightly edging out Neymar in '22/23).

Hugo Ekitike currently rates better.

#SGE

Hugo Ekitike: The Improvement pt. 2

Some comparable datapoints to each of his four seasons:

- '21/22: Bukayo Saka & Gabriel Martinelli ('22/23)
- '22/23: Adam Hložek ('22/23)
- '23/24: Victor Osimhen ('21/22)
- '24/25: Michael Olise ('23/24)

#SGE

Hugo Ekitike: The Improvement pt. 1

The same plot as the initial (depth of touch vs. model value), but for all U22 forwards and wingers over the past four seasons.

Ekitike showed plenty of promise at Reims, hit the brakes at PSG, and has rebounded resoundingly at Frankfurt.

#SGE

Hugo Ekitike: Expect Goal Machine pt. 3

That said, he and Omar Marmoush are one of the most prolific duos in world football right now, so the environment is conducive to continuing this historic trajectory.

#SGE

Hugo Ekitike: Expect Goal Machine pt. 2

If he continues at this pace (1.03 npxG+xAG/90), he would be the only U22 player in FBRef's database to finish the season with >1 expected goal contributions per 90, so I'm confident there will be some regression.

#SGE

Hugo Ekitike: Expect Goal Machine pt. 1

There isn't a forward in Europe with Ekitike's current profile of expected goals and assists per 90.

The graph highlights this pretty well.

This has been the main catalyst behind his breakout season.

#SGE

Hugo Ekitike Data Profile pt. 5

As well, he has improved on his capabilities in the big three "possession" metrics: receiving, carrying, and especially passing.

#SGE

Hugo Ekitike Data Profile pt. 4

If xA > xAG, the player has a high volume of dangerous passes, but they don't lead to high percentage chances. If xAG > xA, the player creates high quality chances. This is Hugo Ekitike.

#SGE

Hugo Ekitike Data Profile pt. 3

For reference, xAG is a measure of the xG likelihood of passes that lead to a shot, while xA is a measure of the xG likelihood of all passes, whether or not they lead to a shot.

#SGE

Hugo Ekitike Data Profile pt. 2

For chance creation, his quality has improved (rather than quantity). To demonstrate, his xAG-xA/90 over the past three seasons:

0.00 ➡️ +0.02 ➡️ +0.06

#SGE

Hugo Ekitike Data Profile pt. 1

The biggest jump in performance this season has been his npxG+A values per 90.

In terms of goal scoring, the catalyst behind his improvement has been his positioning and shot selection. Ekitike's npxG/shot over the past three seasons:

0.11 ➡️ 0.08 ➡️ 0.23

#SGE

Thread Overview pt. 3

The graph shows all the wingers and fwds plotted by how far up the pitch they see the ball (left to right), + how well they rank in my model (bottom to top).

In this thread, I'll seek to demonstrate how much Ekitike has improved this year through the lens of my model.

#SGE