Instead of me clogging up other threads, especially @RTB 's, I thought i'd start my own football data thread. I will post anything data related in here throughout the season so if anyone is interested in this please keep an eye out for this thread. As some on here know, I work within football. Specifically in scouting and recruitment departments. Recently left a Championship club but I have also worked for another EFL club and 2 non league clubs in similar positions. My work has always involved data so my main goal of this thread is to dispel some myths around data and try to explain things better than I have done previously. If you think data is pointless then that is fine, however Football has embraced it and the clubs that use it the best are more often than not very successful. As I said, I will post anything I find interesting within the football data world here, with the focus being SAFC and the Championship of course. A lot of this thread will be me rambling on so I apologise for that in advance. If anyone has any questions please fire away.
Good idea for a thread. I'll be watching with interest. Here's a question I have. I understand that clubs typically develop their own recruitment databases, combining in-house scouting reports with performance data. However, when it comes to analysing opposition squads, is there an industry-standard provider for performance data that most clubs rely on? If so, do clubs need to subscribe to these services, and how big an investment is it? I guess, ultimately, I’m wondering to what extent access to performance data varies between clubs based on their willingness to invest and how big an advantage that might be?
Good idea for a thread and I apologise in advance for going off on a tangent already but since I've been watching the grandson's football matches and seen some scouts there, I've always wondered how you actually get picked up by clubs to go and scout players either kids or above
There's a load of companies that clubs pay to use their data. Also a load of companies that clubs pay to make data related decisions for them, instead of employing in house data scientists. Then you have the likes of Brighton and Brentford that have owners that have built their own data. So to answer the question, it depends on the club really. Some will build and rely on their own people whereas some will pay a fee for a company to do the work for them. A company like StatsBomb are probably the industry leaders, their website claims they work with over 300 clubs around the world. There are some big clubs that have fallen behind in recent years and in my opinion I think a lot of it is reluctance to invest in this side of the game. Man Utd for example, they are getting their act together in the last year or so, however they have a long way to go. You look at the money they have spent in the last 10 years or so compared to Liverpool for example, who are a club who rely heavily on data, and then compare the trophy hauls and it's pretty obvious Liverpool have gained an advantage. Someone I know in the industry constantly says that PL clubs will chuck £10m on a fee + £5m in wages on a bang average footballer, but if they spent half of that on a data setup they could reap rewards for decades.
My personal route in was my Uni lecturer. It's very much a who you know not what you know industry. A lot of voluntary work in non league and constantly applying for paid work within the football league once your CV looks good. Plenty courses to take as well, the FA have a free level 1 that is pretty much required for most roles and then you can pay to do the level 2 or pay for courses with other companies.
As a stand alone metric, no not really. Think it can be misleading. You can have one game where you win 5-0 with 5 tap ins with a total xG of 5 but then not score in the next 3 games and not create any chances but still top the xG league table. You could also theoretically have 20 shots from distance that measure at 0.1 xG and finish the game with an xG of 2, whereas the opposition could create 3 big chances worth 0.5xG each and win the game easily but look like they've took a battering on the stat sheet. A bit wordy but some places give you Open Play xG per shot and I like that. Takes away teams that shoot a lot from distance and rewards the teams that create big chances. xGOT (expected goals on target) is a good metric for finishing ability, it measures where in the goal the shot crosses the line and how likely it is to go in.
Very interesting, mate. I wonder if anyone's done the analysis on the value of the analysis, would assume they have, given the money involved. Very hard to measure but is anyone measuring to what extent a club's level of investment in data influences how well they are prepared for the opposition, accounting for obvious variables like the overall talent of the squad and therefore the expected outcome, or is that primarily down to the talent and ability of the gaffa. Probably an impossible question to fully answer but it crosses my mind regularly.
Majority of clubs have opposition analysts, that goes right down to non league. I guess that some clubs will have multiple of them which would allow for better preparation for upcoming opponents. A lot of it is down to the manager to use what the analyst has seen and create a game plan though. I'd imagine someone out there knows individual clubs levels of investment, would be interested to see how the level of investment equates to success.
Good idea for a thread, I'll definitely be keeping an eye on it. What do you make of our data structure? We seemed to be doing it in house with the appointment of James Young but then he left and we seem to be out sourcing it now to serios and possibly another company. Do you think Young was brought in to develop a model of criteria which would suit our player profiles and then out source it when he left ( people in his position only ever seem to stay for a limited period) or do you think they just changed tack and decided not to do it fully in house?
I think you pretty much nailed it here. My personal preference would be to keep everything in house, but the flip side of that is that the best people in the industry don't really want to be tied to individual clubs, they'd rather have their own business and do contract work. I think the summer just gone was a return to a more data driven club which I think we went away from last season. It's well known that we appointed Mowbray because a company compiled a list of suitable head coaches for us and he was number 1, I think RLB was a similar appointment. Same with the transfers we made.
Yeah I think that's probably it. The recruitment did go off piste around the same time he left, but hopefully that was just teething issues. I'd like to see it fully in-house but I doubt we have the resources atm. At least it's a local company doing it.
People might remember the introduction of the "black box" which was a hard drive, remotely updated by a scouting group that collated data on players, this is probably going back 15 year ago, I believe it cost around £80,000 a month to receive a subscription and almost all Premiership teams had it. It has obviously come much further along now.
Not that I've seen. I hope none of them do as the grandson said he'll get me a box at the SOL when he makes it
I did an online scouting course during COVID and on the back of the course the guy who ran it put me in contact with the club I scout for now. Def about who you know, most scouts do it on a voluntary basis and the paid jobs will have 100s and 100s of applicants for it To get a paid scouting job there's a lot of luck in it but the key is building a good network in the game
The following data I am going to discuss is from a piece written by Ben Mayhew on his website; https://www.experimental361.com/p/championship-scatter-graphics-november First off is a graph looking at shots for and against. Fairly self explanatory, you want to be in the bottom right of the graph ideally. Leeds are suffocating teams and you can see why Boro are starting to have an upturn in form. Next we are going to look at attacking effectiveness, measuring shots taken per goal and shots taken per match. Again, you want to be in the bottom right of this graph. We are the most clinical team in the league, I personally just wonder how sustainable that is. Now onto some defensive numbers. Top left is where you want to be here, this is a strong point of ours. Then finally an xG for vs xG against table. Bottom right is best here. Shows Millwall are a very good side and we are in for an incredibly hard game tomorrow. In conclusion, I want us to be creating a load more chances. We are good defensively but often in games we fail to create all that much which will become an issue over the course of the season. As a major caveat to all of this, gamestate is not factored in. We have only been behind for 5.4% of our minutes played this season (2nd fewest behind Sheff Utd) and ahead in 46.9% (again only 2nd behind Sheff Utd) so we haven't really been chasing games at any point which would naturally mean more shots and more chances created. We are also the best 1st half team in the league, won 9 first halfs, drawn 5 and lost just 1 (Watford away), scoring 13 and conceding 2 in first halfs. This will play a large factor into why we seem to create less, although I still want to see more.
