As a child of the 1980’s, back when chronic boredom still existed in childhood, one of the VHS tapes that I wore out watching was Top Gun. One of my dad’s few indulgences was AV equipment, so the entire house would rattle from the surround sound during the opening scene, when the fighter jets flew across the screen.
New Toy ALERT!
Wyscout recently added some new metrics for the top 5 EU leagues relating to physical movement. If you have noticed that male football players have been wearing ‘bras’ for a number of years, that is due to various technologies designed to track things like player movement, fitness, etc. - NOT a bro or manzier?
The history is limited and I will only have temporary access, as Wyscout does not currently cover the Scottish Premiership with these metrics, so whatever incremental cost they will demand is unlikely to be worth it. But in the mean time, I get to play around with the data!
A Celtic player with related data available for his last season at Augsburg is Arne Engels. Give me access to new data and the outcome is predictable - benchmarking time!
A quick summary of the terms:
High Speed Running - 20-25 km/hour
Sprinting - 25+ km/hour
Medium Acceleration - 1.5 meters/second to 3 meters/second
High Acceleration - 3+ meters/second
As I often do with ‘off the shelf’ metrics, I have added various ratios via X per Y % in order to try and tease out additional potential context and signals.
This exercise offers some potential lessons on the distinction between data and analytics. One of the ratios I have been generating as part of the transfer benchmarking exercise, available in a dedicated section of this Substack, is accelerations per progressive run.
My theory on calculating that ratio was/has been that it may offer a signal for pace - basically, I was trying to work with the data I had available knowing that it was VERY limited relative to what I was trying to gauge. You can review the broader transfer benchmarking exercise here, where I happened to use Arne Engels as the case study.
Here were the relevant metrics from the exercise:
The number of progressive runs can be materially impacted by systemic factors - the team for which a player plays, relative competitive level, and the style of play implemented by the club/manager. The ratio metric offers some potential signal that Engels’ level of speed was better than what the raw metrics may be suggesting.
As Alan and often say, we are interested in analytical accuracy rather than data precision. The interplay between the information presented in the two graphics above highlight the importance of this distinction. The introduction of the tracking data-related metrics is an obvious and material upgrade to the simple ratio I have been using. Hooray! But it is the underlying analytical process which is preeminent vs data quality.
Now back to the actual analysis - important caveats here are that this is a limited data sample size and I am new to the metrics and data set. Familiarity and experience tend to improve intuition and the sort of pattern recognition that AI/ML algorithms attempt to replicate.
When going through Engels’ data within the broader peer group, my eye happened to be drawn to two Atalanta central midfielders: Jose Ederson and Ibrahim Sulemana.
I decided to add them into a comparative analysis for the obvious reason of the proximity of next week’s Champions League fixture, but also because of Atalanta’s reputation as an analytics-focused club when it comes to recruitment.
Ederson, now age 25, was signed prior to the 2022-2023 season for 22.9 million euros, a season after Salernitana signed him for 6.5 million euros from Corinthians in Brazil.
Sulemana, now age 21, was just signed this summer for 7.5 million euros from Cagliara, a season after they paid 4 million euros to Hellas Verona.
Ederson is a focal point at the base in Atalanta’s midfield within their typical 3-4-3 or 3-4-1-2 setup. Sulemana appears to be a development player within the usual conveyor belt of Atalanta squad planning, and has not yet played material minutes. His data shown above is from last season at Cagliara.
My read of the data and Engel’s benchmarking is that he is an extremely vigilant player that operates at a high effort level. The consistency of his percentiles related to running at speed vs overall running is the basis for this conclusion. At the same time, his raw speed appears to be lacking.
Of course, physical attributes like speed and quickness are only part of the picture when it comes to players, but natural raw speed and quickness cannot be taught. In addition, my view remains that Celtic need to be much smarter about allocating resources - like Atalanta.
Matt O’Riley offered a lot of attributes that helped offset what my simple benchmarking exercise suggested was, at best, mediocre pace at a lower English league level. So far, Engels’ profile does not indicate that sort of decision making or on-ball skill set. We now have potential evidence that his speed is, at best, mediocre at the aspired for European level.
This all leaves me asking a question - is being a hard worker with a good mentality worth a record transfer fee? I could see the answer being “yes” if the player displays higher end speed, but…
With reports of Paul Tisdale consulting Celtic over the summer, I hope part of that experience and his appointment results in a remit to materially raise the level of the analytical framework being deployed with player recruitment.
If I recall correctly, Joe Ledley wasn't blessed with speed. The experts said that if he was faster he'd have played for a top 6 EPL club. Maybe Engels is the same.
I’m really liking the benchmarking articles James. Find it much easier to follow and consume on this platform.
On the article itself, it all sadly resonates with my ‘at the game eye test’