Mark Tenenholtz argues soccer analytics are highly advanced, citing graph neural networks and expected goals

@signulll They don't take as many shots from far outside the box anymore

@signulll Soccer analytics is quite advanced. xG is just one example, graph neural networks to analyze spacing are a more advanced example

@signulll Analytics are applied in soccer at every level of the game, from player profile analysis for recruiting to activity analysis for training/injury prevention to tactical analysis for game prep

@signulll should read net gains by ryan o’hanlon

@signulll also, read the story of professional gambler and billionare owner of brighton tony bloom who brought gambling analytics into the transfer market

@signulll I read a long article a dozen years ago about the new analytics in soccer: the big revelation was that the two best players were ... Messi and Ronaldo, just like a billion people had seen with their own eyes.

@signulll Funny you say that - we @nhssaa22841 are hosting Chris Dove of @Arsenal and Prof @ScottNestler of @ufsports @UFNZimmerman for a webinar on soccer analytics - link here https://forms.gle/62PYoC1Ked8A8TSg6 it’s FREE. http://Www.nhssaa.org

@signulll There is plenty of analytics nowadays in soccer. I visited the data team of Parma FC, and their technical skills were at the FAANG level. Top people.

@signulll yes analyis of sports is a militarily american method, most other countries prefer the heart of the game than the logistics of it

@signulll Functionally all pro clubs doing some version of analytics off of "event data" (on-ball actions such as passes, shots, tackles, duels)

Smartest clubs are modeling tactics & individual player decision-making by making use of spatiotemporal tracking data

definitely growing but it still feels way less developed, or at least way less visible, than in american sports

clubs track live player data, set pieces, pressing triggers, spacing, subs, load management, etc. coaches using computers/ipads during games is also way more common now

thing is fans don’t get much of that data in the broadcast, so it looks like it doesn’t exist while watching a game

@Steve_Sailer @signulll Has analytics ever disagreed significantly with general consensus on who the best players are?

@signulll Just play football man, it's not that complex. Pick players who have the ability to keep the ball, can thread a pass through a needle, can strike the ball well, can and do run till they drop. If they can do all that and have fun doing it, you have a strong team.

@signulll Exploded last 5-10 years. Scouting teams heavily use analytics to buy players cheap / sell at a huge profit to top teams and recycle back into undervalued signings (Brentford, Brighton, etc.).

Top teams also using of course too

@signulll Introduction of xG (expected goals) as a stat kind of kickstarted it.

Last year felt like a narrative shift for the first time as people now feel EPL is overly Analytical / tactical and a boring product. People hated on Arsenal winning for this reason (Spurs fan bias aside)

There are a few data providers that everyone is using. Algorithms are mostly developed in-house. One hour after the end of the game, the coaching staff had the data/insights in their hands (the data analysis pipeline starts immediately after the game ends). The preferred delivery is through visualizations.

@tonyapc42 @signulll In the late 20th Century, baseball analytics crunched vast heaps of numbers to discover that the greatest player ever was ... Babe Ruth.

@signulll Teams are starting to do it. Chelsea claims to be using data heavily for assessing transfers but their transfers have been disasters sooo

There is but mostly in player recruiting and pre and post match analysis. Some teams use it heavily for their play but it’s impossible to get right because it’s 90 minutes of open play

You can math the hell out of it but what do you do when a player decides today is the day they’re scoring from 40 yards out

@signulll As a former sports analytics minor in college (yes they have that lol), soccer was one of the furthest behind (alongside hockey) in adapting advanced analytics.

xG is basically all they got.

Soccer moneyball though based on xG would be set piece maxxing