Transfers based on software

Cliche Guevara

Full Member
Joined
Dec 8, 2015
Messages
3,790
Location
Inverness
I’m starting to wonder if technology and science will be the ruin of football.

Seeing cnuts sitting in the dugout with an open laptop is fecking demoralising.
 

Bulldog United

New Member
Joined
Feb 18, 2019
Messages
1,226
Location
Liverpool
Not as practical.
Using a tablet is much more convenient in that environment.

The issue is getting the software onto them, because Apple has everything locked down. A lot of football clubs have more than enough money to fund getting custom apps built for their own internal use only on iOS. Of course, they could always use something like Surface running Windows if they insist on sticking with their current software.
 

Swarlos

New Member
Newbie
Joined
Sep 25, 2018
Messages
157
Location
Oslo, Norway.
Supports
Lyn FK, Liverpool
To be honest, I think statistical analysis, or software as you call it, will be better than humans at identifying players if it isn't already. Humans have lots of inherent biases which effect how they view a player. Are they too short? Do they run funny? Do they look clumsy on the ball, or maybe a little lanky? Software using statistics will cut right through that, and will better be able to identify the players that fit your team, without human biases coming in the way.

The issue right now is collecting enough data, and applying it effectively. I am certain that as time goes on, these systems will get better and better, and anyone not using it will be left behind. In time maybe machine learning will be applied in the shape of neural networks as well.
 

Crashoutcassius

Full Member
Joined
Oct 10, 2013
Messages
10,295
Location
playa del carmen
Is that really how it happened and not just some clickbait exaggeration?
I'm sure most clubs use some kind of advanced analytics to assists (not flat out decide) their transfers these days.
No mate goal dot com says it's the software. It's like that scene from Willy wonkas chocolate factory, huge computer splutters and spits out 3 names every summer
 

Dolf

Full Member
Joined
May 27, 2016
Messages
2,887
Location
Amsterdam
Depay actually used software in his decision to go to Lyon. He hired a company that used software to analyze all available clubs for him and they decided that Lyon was by far the best option because of their playing style, league, chance of playing minutes etc. Has worked out great for him so far.
 

sullydnl

Ross Kemp's caf ID
Joined
Sep 13, 2012
Messages
34,063
It would be much more worrying if we weren't using available data to inform transfer decisions.
 

Cliche Guevara

Full Member
Joined
Dec 8, 2015
Messages
3,790
Location
Inverness
To be honest, I think statistical analysis, or software as you call it, will be better than humans at identifying players if it isn't already. Humans have lots of inherent biases which effect how they view a player. Are they too short? Do they run funny? Do they look clumsy on the ball, or maybe a little lanky? Software using statistics will cut right through that, and will better be able to identify the players that fit your team, without human biases coming in the way.

The issue right now is collecting enough data, and applying it effectively. I am certain that as time goes on, these systems will get better and better, and anyone not using it will be left behind. In time maybe machine learning will be applied in the shape of neural networks as well.
Or it will quite quickly be shown up as bollocks because football isn’t a game of statistics.
 

starman

Full Member
Joined
May 14, 2007
Messages
7,092
Location
Under a tree.
Every club uses some kind of analytic tool.

Surely not signing Boateng was the correct decision. Hes not been good for about 3 or 4 years.. Felix is also a big gamble for 120m
 

0le

Full Member
Joined
Dec 25, 2017
Messages
5,806
Location
UK
Why do people believe these stupid articles? Do you really think someone at Goal.com has access to our data analytics team and that they know the exact reasons why we did not sign a player? Even if they had a source (they don't), that source would also just be making a guess because the decisions will be made higher up the food chain.
 

Cee90

Redcafe Fantasy Football Champion 2012/13
Joined
May 4, 2009
Messages
5,019
Location
N2402
Why do people believe these stupid articles? Do you really think someone at Goal.com has access to our data analytics team and that they know the exact reasons why we did not sign a player? Even if they had a source (they don't), that source would also just be making a guess because the decisions will be made higher up the food chain.
As already mentioned in this thread, goal.com got their information from a recent report written by a reputable journalist at the Athletic.

I'm pretty sure it's legit.

I like that we're putting some emphasis on data analytics in football personally. However, what worries me is that surely you don't really need some fancy software to identify that Boateng was on his last legs?
 

adexkola

Doesn't understand sportswashing.
Joined
Mar 17, 2008
Messages
48,031
Location
The CL is a glorified FA Cup set to music
Supports
orderly disembarking on planes
Do you have more info on it doing so at the moment? I wouldn't have thought there'd be software good enough yet.
The NBA have sensors in each arena that track player and ball position.

https://blog.revolutionanalytics.com/2016/09/analyzing-nba-basketball-data-with-r.html

A casual fan was able to download the data and play with it in open source software. Imagine what's being done with it behind the scenes.

I see no reason why football can't be analyzed similarly.
 

Cloud7

Full Member
Joined
Jan 11, 2016
Messages
12,778
It’s a fecking good thing the software told us not to buy Boateng, he would have been a dreadful signing.

Move software into the DOF role?
 

Fosu-Mens

Full Member
Joined
Apr 11, 2016
Messages
4,101
Location
Fred | 2019/20 Performances
Similar to how most clubs have operated for years, the ancient traditionalistic (insert) people at the club have gotten around to get some statistical software. Would not be suprised if they cheaped out on this and are using WhoScored or Squawka Lite without adblocker.

What most clubs uses this software for is screening and identifying potential hidden gems. Given that we are among the last ones to employ this not so new tool we will be far behind most other clubs in how to use the software to identify the best indicators for each position. E.g. used for identifying and creating a list of potential targets that requires further scouting, not as the ultimate decisiontool. Then again, they might use this software similar to this:

The idea that club thinks that the fans will be appeased by this is like a 40 year old man asking for compliments because he was able to wipe his own ass.
 

SadlerMUFC

Thinks for himself
Joined
Dec 7, 2017
Messages
5,746
Location
Niagara Falls, Canada
Using data to make the right desicion is a good thing. Boateng is slowing down after the 70th minute is a good information. But desicion of not buying Felix, because he is too weak, reminds me of some american sport analytics, who formed the perfect quaterback based on data and now have a tamplet how he has to look. Perfect size, perfect wingspan...... But it does not work this way, just look at Messi!
And look at the Bills in the 90's. Had the right guy with Flutie but they said he was "too short" despite leading them to victory and victory. So they benched him for their new poster boy and lost...
 

Mark Pawelek

New Member
Joined
Apr 9, 2014
Messages
2,598
Location
Kent, near London
goal.com has an artical about United. We had the chance to sign Boateng and Felix but refused because of a new software, which has been included into the new scouting system. The analysing tool saw Boatang as too slow afer the 70th minute and Felix too weak for the PL.
United had no working scouting system over the last years, but should United sign or not sign someone based on a software? Think of Messi, a small boy, too small for football as a teen, now the best player in the world. Use a software and he would have been a no go because of his size. Life is not that easy....
What do you think?

https://www.goal.com/de/meldungen/m...oao-felix-scouting/1bwerrvnriq6c1ir1cwm2ux4p8
Software just does what it's instructed. Look through a database of stats. Score a player positively for the good things they do. Score negatively for bad things. This saves a lot of time. If any particular scout or coach objects to it they can argue for mitigating circumstances. For example - when a defender is too slow or lacks stamina, you can actually watch their match performances after the 70th minute to see whether they concede more or fewer chances to the opposition. There are tens of thousands of players to consider when scouting replacements. You can't watch every minute each of them played. You also need something to check innate bias and favouritism; stats do that.
 
Last edited:

red thru&thru

Full Member
Joined
Mar 2, 2004
Messages
7,657
It depends how the software is used, which I have no idea on. But what is for sure, nothing beats the trained eye. Look at one of the best DoF's, Luis Campos. In his recent Sky interview, it is said he was on a plane that often last year, he travelled around the world 3 times over!
 

tenpoless

Full Member
Joined
Oct 20, 2014
Messages
16,176
Location
Fabinho's forehead
I have tried to built a Neural Network transfer recommendation using three parameters only : speed, skill and age when joining. And then I trained the model using several successful and failed ManUtd transfers as examples. Result is between 0 to 1 based on the degree of success (up to me). I tried to evaluate Daniel James, the model gave me 0.67 something, I tried Fred it gave me 0.78 something.

In my mind this is the kind of model that We use. I trust this model and I trust the transfer based on software too. At least it's more reasonable and argumentative than "For me, a zim without Brazilian isn't a football zim".

Other results : Fellaini (0.21), Martial (0.89), Pogba (0.92), McTomminay (0.32).
 

Beans

Full Member
Joined
Aug 9, 2019
Messages
3,512
Location
Midwest, USA
Supports
Neutral
I like our current strategy, where we bend over and deficate on our own face, and then look in the mirror to see what name it spells out.
 

Zed 101

Full Member
Joined
Jun 27, 2014
Messages
1,391
Funny thing is with this hoo-ha is that the computer didn't watch those players and make those decisions, a scout, multiple scouts watched those players and their reports were fed into the computer, along with stats, where from errr match reports compiled by humans.... all this software does is compile the info and extrapolate it against a set of parameters.
 

Cliche Guevara

Full Member
Joined
Dec 8, 2015
Messages
3,790
Location
Inverness
I have tried to built a Neural Network transfer recommendation using three parameters only : speed, skill and age when joining. And then I trained the model using several successful and failed ManUtd transfers as examples. Result is between 0 to 1 based on the degree of success (up to me). I tried to evaluate Daniel James, the model gave me 0.67 something, I tried Fred it gave me 0.78 something.

In my mind this is the kind of model that We use. I trust this model and I trust the transfer based on software too. At least it's more reasonable and argumentative than "For me, a zim without Brazilian isn't a football zim".

Other results : Fellaini (0.21), Martial (0.89), Pogba (0.92), McTomminay (0.32).
I’ve no idea what you’ve done or how you’ve done it but those results look like utter bollocks.

I’m probably misunderstanding your post.
 

Schneckerl

Full Member
Joined
Jun 17, 2016
Messages
2,704
No mate goal dot com says it's the software. It's like that scene from Willy wonkas chocolate factory, huge computer splutters and spits out 3 names every summer
:lol:
Goal.com and the likes always dumbs down anything involving numbers/analytics like it's meant for middle schoolers.
It's like an article about the Belgian research paper was titled "Supercomputer solves the Ronaldo vs Messi debate"
 

Mr. Meeseeks

Full Member
Joined
Apr 12, 2011
Messages
714
Location
In a box
This is a rather ignorant thread. Software is one of the major forces behind the modern world.

They use software to model the interactions of the many forces on buildings, bridges, aircraft and vehicles thereby ensuring their safety and efficiency. They use software at Cern to analyse the complex interactions of countless particles in order to determine which particles form after their collisions. No human or team of humans can do those things in anything approximating a reasonable timeframe without those tools.

They have software so sophisticated that it could beat the best humans in the world at Chess, Go and Jeopardy (google’s AI can probably do all three of those things).

If a player is good, the data will bear that out. A knowledgeable scout equipped with the proper statistical analysis software and the relevant statistics will find better players than one relying on ‘gut feelings’ alone.
 

Cliche Guevara

Full Member
Joined
Dec 8, 2015
Messages
3,790
Location
Inverness
This is a rather ignorant thread. Software is one of the major forces behind the modern world.

They use software to model the interactions of the many forces on buildings, bridges, aircraft and vehicles thereby ensuring their safety and efficiency. They use software at Cern to analyse the complex interactions of countless particles in order to determine which particles form after their collisions. No human or team of humans can do those things in anything approximating a reasonable timeframe without those tools.

They have software so sophisticated that it could beat the best humans in the world at Chess, Go and Jeopardy (google’s AI can probably do all three of those things).

If a player is good, the data will bear that out. A knowledgeable scout equipped with the proper statistical analysis software and the relevant statistics will find better players than one relying on ‘gut feelings’ alone.
Based on what?
 

Tom Cato

Godt nyttår!
Joined
Jan 3, 2019
Messages
7,565
goal.com has an artical about United. We had the chance to sign Boateng and Felix but refused because of a new software, which has been included into the new scouting system. The analysing tool saw Boatang as too slow afer the 70th minute and Felix too weak for the PL.
United had no working scouting system over the last years, but should United sign or not sign someone based on a software? Think of Messi, a small boy, too small for football as a teen, now the best player in the world. Use a software and he would have been a no go because of his size. Life is not that easy....
What do you think?

https://www.goal.com/de/meldungen/m...oao-felix-scouting/1bwerrvnriq6c1ir1cwm2ux4p8
Manchester United employs over 50 scouts and has since 2016. There is very much an extensive scouting network in place. This software is used as a complimentary piece to evaluate important datasets that determines anything from statistical chances of injury, injury history, current conditioning, ball ability, shooting %, pass %, duels won/lost, aerial stats. All of it. Think of it as a scouts failsafe. The scouts will recommend a player, and the club uses both the scout and the analytical tools to determine if this is worth a shot.
 

Mr. Meeseeks

Full Member
Joined
Apr 12, 2011
Messages
714
Location
In a box
I’d say it’s the opposite. Your examples bear no relation to football and you’re removing the two most relevant elements from the equation. Those being the human element and how the player actually plays football.
It is not actually the software that is doing the work, it is statistical analysis which is simple mathematics. The software is essentially acting like a complex calculator for them to compare many different players at once.

I am not saying that software will ever replace scouts just like the software used in engineering will never replace actual engineers. You still need competent professionals in the field who are able to interpret what the data says.

But the fact that data makes a scout more effective is undeniable isn’t it? How else can they reliably make an apples to apples comparison between hundreds of different players?
 

giorno

boob novice
Joined
Jul 20, 2016
Messages
26,144
Supports
Real Madrid
Using data to make the right desicion is a good thing. Boateng is slowing down after the 70th minute is a good information. But desicion of not buying Felix, because he is too weak, reminds me of some american sport analytics, who formed the perfect quaterback based on data and now have a tamplet how he has to look. Perfect size, perfect wingspan...... But it does not work this way, just look at Messi!
Messi, who is the same size as Pelé, Maradona, Garrincha, Puskas, Charlton, Romario, Zico, Muller, Baggio, Dalglish, Keegan, Owen, do i have to keep going?

Messi *is* the perfect size for a creative forward/midfielder

Now good statistical models can be of great help to scouting. Rejecting a 19 year old on the basis of being too weak is bonkers, but i doubt it was the software making the decision...