Stats in Football

canadiancherry

canadiancherry

Star Player
Very fascinating article in the New York Times on Liverpool's use of stats. They have a bunch of guys with PhDs in physics down the hall from Klopp's office! I wonder what statistical models we are using?
https://www.nytimes.com/2019/05/22/magazine/soccer-data-liverpool.html

Very interesting is that the use of advanced analytics isn't just for winning games; it's for finding transfer targets that can later earn you money:
Another acquisition may have been even more important. Soon after arriving at Liverpool, Graham was asked to research a left winger at Inter Milan, Philippe Coutinho. His data strongly endorsed Coutinho. Liverpool bought Coutinho’s rights for about $16 million. Over the next five years, Coutinho’s play contributed to Liverpool’s revival. But his most important contribution was to accrue value. Last year, Barcelona paid Liverpool about $170 million for Coutinho. Soon after, Liverpool spent more than $200 million on three new players: Alisson Becker, the goalkeeper; the midfielder Fabinho; and the fullback Virgil van Dijk. All became crucial contributors this season. These were known commodities, and none came at a bargain price. But without the profit made by selling Coutinho, Henry assured me, those players would not have been acquired.
Click to expand...
 
Very detailed article thanks for sharing, although it did slightly grind on me in its theme that Liverpool’s use is revolutionary - I think most top level clubs (and others in lower leagues, Brentford being a prime example) have been using data in such ways for a number of years as well - the difference is probably Liverpool willing to lift the cover on their operation more than others with the release of this article!

To my knowledge AFCB haven’t really acknowledged use of data in transfer policy (use of resources probably the main reason with EH preferring to use money elsewhere -although I do love the thought of Fletch and Richard Hughes spending hours locked in a bunker reviewing their statistical analysis!) however, there have been some articles in the past with regards to use of data and accompanying video analysis in player development and match day prep such as this example (ironically on goalkeeping given criticism of that department!)

http://www.espn.com/soccer/english-...the-way-in-goalkeeper-training-and-technology

Although part of me despairs when I see my beloved game broken down into data like this, it does give me comfort to know that “success” of that research ultimately comes down to how the manager passes that through to the players and whether they in turn are successful in carrying out those instructions on the pitch - there are so many variables once your cross the white line into the pitch that there is still uncertainty....hopefully the result is not too many Pulis type anti football machines whilst enigmas and genius’s like Messi, Le Tissier, will always find a way to produce provided they pass the fitness tests!
 
If Spurs win the Euro Final with an own goal or fluke....all that data gets knocked into a cocked hat! Thats just one example!
However I wont totally dismiss this Liverpool revelation...this modern day version of the Don Revie ' dossiers' !
Not while Afcb has bits of Liverpool Data Material milling around DC with little subsequent 'stats' value ....and the proceeds of said Data sitting in their bank!
 
If one were cynical, one could conclude that ate the real takeaway is, never buy a player from Liverpool (or sell to them either!)
 
LewSwimmin said:
Interesting article.
However, you don't need a computer device to see that Asmir parries too many low shots instead of catching and smothering them like Artur does.
Click to expand...
But what stats might show is Asmir’s positioning when he makes these parries and what angles the shots are coming from when he parries. Perhaps there’s a trend, maybe there isn’t but that’s the sort of thing an analytics team can show you rather than a sweeping generation that a keeper is useless as he always parries.

In my opinion there’s also a difference between stats and analytics. Stats is just the really basic stuff, possession, pass accuracy, assists etc. As the article mentions when discussing Keita, analytics goes one stage further. Keita’s pass accuracy might be low, but by analysing the rest of the data you gain an understanding as to why and whether overall that players overall contribution is positive or negative to the team.
 
I saw yesterday that after pundits (Keys and Gray) were laughing at Liverpool for bringing a throw in specialist on board, they've gone from 3rd worst in the PL to 2nd best in Europe at retaining possession from throw ins. The first is Midtjylland, the specialists other club.

Here's a clip of one of their data analysts talking through some relatively basic stuff but interesting how they go back and look at decision making and then use it in the transfer market:

https://streamable.com/ay4qq

As you might be able to tell, Reddit is just full of Liverpool stuff atm
 
It’s all well and good throwing money at things that matter.

We’ve funded Liverpools ability to pay these people by buying three of their rejects.

Stats decide what’s happened after the fact, a bit like our criticism of buying Liverpool players.

If you perform on the day and win the game, it doesn’t matter what the stats say.

Stats are another reason to try and explain what’s already happened.

A bit like doing a spreadsheet on the reasons why your missus ran off with the milkman.
 
How much money do you really think the throw in specialist costs? I suspect they could pay for him for about 100 years with Ibe's money alone. Saying we've funded it is ridiculous. It's not the expense, anyone like that will be so small compared to a players salary. It's more that most managers wouldn't see the value in something as insignificant as throw ins.
 
dragonm8 said:
Stats are another reason to try and explain what’s already happened.
Click to expand...

Partially right.

Stats are another reason to try and explain what’s already happened so the players can be better informed what decision in a given situation is likely to bring a better result next time they encounter that scenario.

It won't always work as that's the nature of probability. If the research has been done corrctly and is followed correctly by the players then it will bring a better result more times than not. Except allowing for variable change, which could be the other team's analytics team coming up with a way to combat your method. In which case, it's back to the stats and analytics to work on the next method.
 
Agreed, but football happens at a million miles an hour. It’s all well and good for Steve the stat man to slow it down and analyse it after the fact, but I don’t think telling Jordan Ibe he had 3 touches in Zone B when he would have been better being in position 2 with his left testicle out would compute.
 
You must log in or register to reply here.

;