It hasn’t necessarily made teams better though. PSG have continuously flopped in Europe and look at Chelsea finishing in mid table below teams with a fraction of their budget like Crystal Palace & Brentford in the EPL despite spending €600m+. Good coaching and smart investment can often get the better of crazy spending. Bayern are reportedly now strongly considering spending €110m on Declan Rice, which is an insane figure. He’s a good player but nothing special. Imo Alvarez for €30-40m is much better value. As was Schlotterbeck for €25m and Sule on a free when the likes of De Ligt, Hernandez, Upamecano & Pavard all cost way more. And then Adeyemi for a smaller fee and considerably lower wages then Mane. That’s the way for BVB to get the better of Bayern imo, buy smarter.
IMO there is 2 problems with "buy smarter" 1. There is no reason to think you can be smarter than the other teams all the time. So you might have some good fortune but in the end it is not sustainable. 2. BVB mostly have to sell their best players whereas Bayern do not. Imagine if you still had Haaland. BTW - i am not a Bayern fan. I put it as my team years ago as a pisstake and I can't change it
While that is true, the point about buy smarter was that it's not by choice but the only chance we got. We have to do a better job in the transfer window than the other clubs, not only Bayern. Other clubs in the league are getting more out of their money. That is coupled with getting young talents that need to be given time to develop who we could never buy from other clubs if they already produced like they do for us. We cannot buy the 80 million player. We can only compete the few times when the 15 million youngster turns into the 100 million player after two years with us. The only element missing is actually using the narrow window to win something as long as the Haalands, Sanchos and Bellinghams are still here. We only have ourselves to blame that we blew this chance twice in the last 5 years.
We used to have a lot of these discussions on the Arsenal board, but the problem is what would our alpha in the transfer market be based on? Data?
I'm pretty sure teams use data science and machine learning to predict which players will do better. Liverpool is the one that's well advertised. I'm speculating, but we too must have some kind of secret sauce that we started using during the Mislintat days (if not earlier.) I think we did do a better job than other teams picking out the good players.
That really summarize the situation well imo... Imagine Keeping all the players we let go... Just Lewandowski and Gotze back then. Having that trio intact on our side, maybe it would have been beneficial for Gotze too... Can you imagine Bayern without Lewandowski the last 10 years?
The real miracle this year that no one talks about is Union Berlin!! With a total salary mass that is lower then 7 Bayern players at 16.161 M Euro made the champions league and out perform 14 other BL teams with the 4th lowest annual salary ...
And if it wouldn't have been Union, it would have been Freiburg with an even lower payroll. And applied to us, Mainz outran us by 5.5km with nothing to play for when 3 of our players on the pitch together earned more than their complete team.
Can you throw some insights on how machine learning and data science can predict which players will do better? I am curious
Yes you can find those edges - the problem is they tend to get closed up because other teams adopt them. Look at Liverpool - they really fell away this season because there is no secret formula to always outperform the market
Yes! I'm not an expert in ML nor in ML for football, but I understand the concepts behind ML. In this case you have a lot of historic data on young players. You have some "normal" players and you have the type of players who then became the Messis, CR7s, Phenomenos, Zidanes of the world. (Or even the good players that are below these.) So the question is: Can you identify anything in the data that can tell these two groups apart from a young age? So you take the 2 groups and label them as "good player group" and "normal player group." You take only their data when they were young. Then, for a new young player, you want to know whether his data look more like the good players or the normal players. In technical terms this can be treated either as a classification problem or as an anomaly detection problem. But! Of course the devil is in the details. If you do that blindly, using just normal stats you will probably find that there's a lot of overlap between the two group, such that if you try to classify a young players you wouldn't be able to. So the trick I think is to use your intuition and start figuring out what particular stats or what combination of stats or what other non-football related data you can use to make this prediction. (Technically, what are the features that you want to build your model on.) That's the first challenge I see in figuring this out and I'm sure there are a million more I don't know about.