[FM24] The Traipsing of the Shrew

Shrewnaldo · October 22, 2023

Leoni del Garda - FeralpiSalò

Those of you who are, shall we say, of a certain vintage may remember that I once blogged about Football Manager and a tiny team from the bank of Lake Garda. Back in FM13, I had one of my favourite ever saves with little FeralpiSalò - taking them from the depth of Serie C to a Champions League victory (Lukas Holub you hero), collecting 5 Serie A titles and 3 Coppas Italia along the way.

In real life, the club has been progressing nicely and are playing in their first ever Serie B campaign and, because I no longer have the time or patience to start any lower than the second tier, this felt like the perfect opportunity to re-visit Italy and a team I've always kept one eye out for.

So first, why FeralpiSalò? Well perhaps I can just quote the original blog from 2012:

Quote

So who the hell are Feralpi Salò and why have I decided to take over this club? Salò is a small town on the western edge of Lake Garda in Italy, a place that I know well from holidays with Mrs Naldo and quite probably where our son was conceived – which explains his excellent dress sense, footballing ability and propensity for changing allegiance without warning...

...They are, in any terms, true minnows. Salò has a population of just 10,000 although the surrounding area is fairly well populated; and they play in the tiny Lino Turina, a ground of just 2364 seats which I have actually passed a couple of times… I took this picture just a couple of miles up the shore from Salò.

... This picture serves no purpose other than, you know, it’s nice and it underlines one of the reasons for picking this club. I always like to have a little bit of personal attachment, however small, to the teams I manage

The stadium is so small that it fails Serie B's minimum attendance standard and we will need to play out home games at Piacenza's Leonardo Garilli stadium - 75 miles down the road. Perhaps more on that later.

For now, the story is clear - a tiny club making steady progress but taking its first steps towards the big time... perhaps too soon? Well probably not, but I'm hoping that it should provide a decent challenge. Knowing that the club would/should struggle at this level, I decided to holiday the game-world up to the date I started the save in real life - 20 October 2023. I wanted to introduce some sort of narrative or realism perhaps. It's just not realistic that, having got the club its first-ever promotion to Serie B, Stefano Vecchi would be sacked in pre-season. But 9 games in, rock-bottom of the league and looking like he's out of his depth? That's a different story.

And so this is where I found myself on day one - with no pre-season, no first window transfers and less-than-no money (again, more on that later perhaps).

And that is just fine with me. I'm hoping this save provides a bit of a challenge and to help that along, I'm going to try some other aspects of the game a little differently - specifically tactically?

So what can you expect from this thread? I tend to use these threads to problem solve - so expect rambling streams of consciousness about tactical issues, finances and recruitment. Definitely recruitment. Whilst my interest in other areas of FM has dwindled over the years, I've found myself more and more interested in setting up interesting recruitment strategies - focusing as much as I can on statistics and finding value in non-traditional markets or clubs.

Ideally I'd like to do this without numerical attributes and am hoping that some skinning genius can release a graphical attribute skin soon. For now, I've started with the vanilla skin and just focusing on my own players and opponents - avoiding any sort of recruitment analysis whatsoever. Which is probably just as well given we have zero transfer budget, are over-spending on wages and are currently £2.1m in the red... some more of that challenge I was mentioning.

Anyway, I think that's enough for a brief introduction to the save - I'll be back later on some tactical principles and a bit on recruitment strategies.

Forza Feralpi!

Edited December 17, 2023 by Shrewnaldo

_Ben_ · October 22, 2023

Shall be following along Shrew! Good luck.

Sizeman21 · October 22, 2023

I was a silent observer of your blog way back. You bring a unique touch to your FM experience that resonates in me somewhere. Good luck in Italy, Shrew!

Shrewnaldo · October 22, 2023

2 hours ago, _Ben_ said:

Shall be following along Shrew! Good luck.

Cheers Ben. When are you starting yours? Waiting until after 'early access'? I'm looking to steal some ideas...

1 hour ago, Sizeman21 said:

I was a silent observer of your blog way back. You bring a unique touch to your FM experience that resonates in me somewhere. Good luck in Italy, Shrew!

Thanks! I'm always surprised when people remember the old stuff - even more so when they have positive things to say. All seems like so very long ago...

_Ben_ · October 22, 2023

1 hour ago, Shrewnaldo said:

Cheers Ben. When are you starting yours? Waiting until after 'early access'? I'm looking to steal some ideas...

Yes. First post is ready but skinning and playing until the full game is out and I can dig very, very deeply into it!

SteinkelssonFM · October 22, 2023

Very much looking forward to following along with your FM24 journey Shrew!

Shrewnaldo · October 22, 2023

Identity

Leoni del Garda - FeralpiSalò

Even committed monoglots should be able to get the gist of 'Leoni del Garda' - the Lions of Garda. FeralpiSalò's nickname and embedded within the club's badge, I want to extend the metaphor to the team's identity on the pitch - part of the different approach I mentioned in the opening post. Not unusually, I typically like my FM teams, regardless of the defensive shape, to move into a 3-2-5 or 2-3-5 with the ball - dominating possession as a defensive tactic and then try to get at least one creative player into the '10' position / slot 14. Whilst this has always been easy to do in FM, FM24 offers a few more options - particularly at the 3-2 / 2-3 side of the attacking shape - and some more reactivity to the positioning of team-mates... but I want to ignore all that, at least for now.

Coming into possibly the weakest side in Serie B and sitting bottom after just one win in nine, it feels like we want to focus on basics first. Our predecessor Stefano Vecchi had favoured a 4-4-2 diamond, playing out from the back and trying to keep the ball on the deck, then defending with a high-line and an aggressive press in what looked like a mid-block. So a relatively fashionable approach, shall we say? But it clearly wasn't working - his only victory coming in the local derby with Brescia ("form out of the window" etc) and his team scoring only 5 in 9.

Remember that, in this scenario, the FeralpiSalò board has dispensed with the services of Signor Vecchi because they considered him "out of his depth", that his principles just wouldn't cut it in Serie B. So I like the idea that they're looking for a change of style - an acknowledgement that the squad just doesn't have the ability to play the open, dominant game that the old coach favoured. Instead, we need to be realistic. To be pragmatic and dogged, hard to beat, aggressive in the challenge and direct in attack.

And that short-term pragmatism is going to align nicely with what I want our long-term identity to be - the Lions of Garda. Tough, aggressive, physically dominant and ruthless on the break.

Having taken a short hiatus from FM in the lead-up to 24, I did a spot of football reading in search of inspiration. Mostly I was interested in the analytics and recruitment side of the game, so I read the usual suspects - Moneyball, Money Hackers, The Numbers Game, etc. And whilst there was plenty of recruitment info in there, a few tactical pieces also stood out.

I'm sure this one will be familiar to most people - the value of the clean sheet. Not conceding will bring you, on average, 2.4 points per game. Old news, but worth re-iterating.

The next come as a pair really, the first is talking about the quality of finishing chances:

Quote

... the differences between players at the top level are surprisingly small. English football analyst Omar Chaudhari... has taken a closer look at Cristiano Ronaldo's numbers to prove the point. Part of the British consultancy firm 21st Club, he analysed 1490 shots taken by the Portuguese forward from 2010 to 2017 for Real Madrid in the Spanish league, penalties excluded. A total of 13.3% of Ronaldo's attempts were successful, an increase of just over 2 per cent on the league average, 11.1%. Chaudhari concluded that one of the best strikers of his generation was scoring between one and two goals more per season due to his special talent.

Cristiano Ronaldo can therefore not be said to be an amazingly good finisher; he's extraordinary in a different sense. He shoots much more frequently on goal than all other strikers; on average almost seven times per game, which is an incredibly high figure. More importantly, still, he often shoots from promising positions on the pitch, comprising high xG values.

from Football Hackers

and what sort of attacks tend to lead to shots with high-xG value...

And lastly, another quote from Football Hackers (which I cannot recommend highly enough, fantastic book):

Quote

In football, the team creating the better quality of chances only end up winning the game two thirds of the time. Better teams are this served best by games with many goals; 21st Club's analysis showed that superior sides won 75% of games with more than 2.5 goals. Games with fewer than 2.5 goals were only won in half of all cases by the better teams

All of which brings me nicely back to that identity. We're going to be a low-block side, looking to protect our goal first-and-foremost to maximise those clean sheets and limit the number of goals overall; whilst our primary means of attacking is going to be the counter-attack, with some set-piece prowess thrown in for good measure.

I know that "counter-attacking sides" can mean a lot of things, but we'll be aiming for the more old-fashioned sense of the term than counter-pressing, more Moyes than Klopp.

And whilst I would usually focus on Mental attributes, the Lions of Garda are going to be all about the physicals. Height and power through the middle, pace out wide, stamina throughout. The mentals might not be entirely overlooked - but it's going to be Aggression and Work Rate, more than Composure and Decisions.

You'll probably notice that none of this has mentioned specific shapes or formations. That is deliberate. The principles are simple and do not change - low block, counter with direct attacks. The defensive shape will definitely change, potentially game-to-game. We will need to be pragmatic to both counter-act the strengths of the opposition and counter-attack into their weakest zones.

TL;DR? A summary in bullets:

Feeding into an identity of 'The Lions of Garda'
We'll be prioritising clean sheets, not through possession but in a compact, low-block
We'll be aiming for low-scoring games because these are more likely to see the weaker side produce a win
We'll be prioritising high-xG chances by playing for counter-attacks
We'll be prioritising Physical attributes to make ourselves that big, tough, aggressive team that is hard to beat
Formations may change, the principles do not
If you don't like that, try Leo del Garda instead

How long this approach lasts is uncertain - we might hit a point at which the opposition see us as the stronger side and counter-attacking low blocks become unviable. I suspect (hope) that won't happen for quite some though. Next up - some thoughts on recruitment strategy.

Forza Feralpi!

Shrewnaldo · October 22, 2023

4 hours ago, _Ben_ said:

Yes. First post is ready but skinning and playing until the full game is out and I can dig very, very deeply into it!

Understood. Looking forward to the new version of the skin. Should tie-in with my game style nicely

1 hour ago, MattyLewis11 said:

Very much looking forward to following along with your FM24 journey Shrew!

Cheers Matty, always appreciate your input to the recruitment queries I post

Matty Aqua · October 22, 2023

Nice start! I love old-fashioned counter-attacking football, I will be reading along.

Shrewnaldo · October 23, 2023

Life imitating art here. FeralpiSalò has just relieved Stefano Vecchi of his job on the 23rd October 2023, 3 days after I created that scenario for this save.

robterrace · October 23, 2023

2 minutes ago, Shrewnaldo said:

Life imitating art here. FeralpiSalò has just relieved Stefano Vecchi of his job on the 23rd October 2023, 3 days after I created that scenario for this save.

They're 19th at the moment in Serie B aren't they? Although, I know that doesn't make any difference in wanting a change of management.

Shrewnaldo · October 23, 2023

2 minutes ago, robterrace said:

They're 19th at the moment in Serie B aren't they? Although, I know that doesn't make any difference in wanting a change of management.

Indeed. Only Lecco below them - who are incidentally the only team Feralpi has beat this season (narrowly). And they had a big loss yesterday.

I reckon it's pretty close to what I've been saying above - the board just think he's a bit out of his depth at this level.

Matty Aqua · October 23, 2023

I always start my saves like this, always seems strange to me to take over a manager's job if he still has the job IRL.

SixPointer · October 23, 2023

Here we go!! Love the idea of taking it to real life with the holiday Shrew.

Jimbokav1971 · October 23, 2023

Loving the kits, but 1 & 3 seem too similar. :applause:

On 22/10/2023 at 12:31, Shrewnaldo said:

I tend to use these threads to problem solve - so expect rambling streams of consciousness about tactical issues, finances and recruitment. Definitely recruitment. Whilst my interest in other areas of FM has dwindled over the years, I've found myself more and more interested in setting up interesting recruitment strategies - focusing as much as I can on statistics and finding value in non-traditional markets or clubs.

Ideally I'd like to do this without numerical attributes and am hoping that some skinning genius can release a graphical attribute skin soon. For now, I've started with the vanilla skin and just focusing on my own players and opponents - avoiding any sort of recruitment analysis whatsoever. Which is probably just as well given we have zero transfer budget, are over-spending on wages and are currently £2.1m in the red... some more of that challenge I was mentioning.

Liking the sound of this. You saves often throw up interesting discussions. :thup:

[Edit]

ps. Wow there are some big boys in Serie C. :eek:

Edited October 23, 2023 by Jimbokav1971

Shrewnaldo · October 23, 2023

Recruitment Strategies (Part 1)

Leoni del Garda - FeralpiSalò

FM used to be all about the tactics for me but, for various reasons, my interest in that side of the game has waned. Now I tend to focus more on the squad-building and recruitment is a huge part of that. Part of my increased interest has come with a focus on analytics, rather than just judging players by their numerical attributes. Indeed, I've taken to playing without the 1-20 attributes in recent versions. Whilst some will play entirely attributeless, it's just not for me. I don't think it's realistic that you wouldn't have any idea whatsoever about the players at your disposal - just think about the people that work for you, or with you, in real life. I'm sure you could assign some strengths and weaknesses to each person according to some basic work-related attributes. Or at least, a good manager should - and this is what the attributes represent for me.

Regardless, if I use the numerical attributes then I struggle to see past them, and just tend to let the numbers dominate my thoughts. Replacing them with coloured stars (or similar) in the last few versions has really helped me focus on statistics instead. Everyone is aware that stats are becoming more prevalent, if not ubiquitous, in football and whilst FM is some way behind the curve, in both the type of stats that are collected and their accuracy, there are a number of metrics available for players to use.

Whilst I wait for a graphical attributes skin, I thought I'd outline the strategy I'll be taking to analyse our own squad, in order to identify the recruitment priorities, and then how to find the players to fill those gaps.

I really like the addition of this experience matrix - allowing the player to very quickly assess the age profile of the squad and any gaps. Up until my Bristol City save (Statman and Robins), I'd have been like most FMers and aimed for a squad that fills out the left-hand side of this matrix - bringing in youth players that we can develop and then sell on for profit. Indeed, this is one of the many strategies that has been erroneously claimed as a 'Moneyball' approach to FM. Indeed, 'Moneyball' would probably be the opposite, with the book making numerous references to Bill James' conclusions that "(older) college players are a better investment than high school players by a huge, huge, laughably huge margin".

The reasoning behind this conclusion is that high-school baseball players do not have verifiable statistics from which franchises can draw direct conclusions about the player's likely professional future. It is this theme which I have carried through to FM - purchases will be based on evidence, and that comes from statistical output. Attributes and scout reports are not statistical outputs.

So that means I'm not going to sign a raft of 18 year-old wonderkids who have played just a handful of games each. I'm looking for a track record and will be keeping saves from the end of each season so that I can go back and check detailed statistics from each campaign, before they are wiped when the game refreshes itself in June / July.

In my Robins and Telstar saves, I also abandoned youth development completely - focusing solely on the senior squad. Here I'm going to have a halfway house - jettisoning the under-18 squad at the end of the season but keeping the under-20s. Primarily this is to honour the club vision and FeralpiSalò's real-life commitment to youth development. Sadly, the facilities are pretty poor and even this lip-service may be abandoned in a couple of seasons.

So it's all starting to look a bit Big Sam - low blocks, direct football and a focus on physical, experienced players. And like the much-maligned Mr Allardyce, we'll be heavily into the analytics. Both the Data Hub and custom views such as the below will be thoroughly used.

I've got a decent idea of the profile of players that I'm after, and a combination of these statistics and some basic logic will highlight the three to four profiles / positions that require strengthening. And then it's off shopping.

I really, really don't like the Recruitment Focus system that was introduced in FM23. It takes away a huge part of what I want to do myself, how I want to find players and what I want my scouts to do. So whilst I'll set up the odd focus just to see what comes back, it's not the primary means of finding players. Instead, I typically go through the following steps:

I use the Players in Range screen to extract a record of statistics from all the players within our scouting package
- I used to avoid this screen like the plague, but I've recently accepted that it isn't the cheat screen I'd previously thought. It does not provide access to every player within the game. Instead, it's just those that your club would realistically know about. As FM now presents it to you, I see it as your club going to one of the data providers (Wyscout, Opta, etc) and buying access to their data across a given geographical range
- To make it less 'gamey', I de-select the "interested" tickbox (so that all players are shown) and, whilst I always use attribute masking anyway, I never use it to search by attributes
- I use custom views such as that shown below to extract the data into excel and from there manipulate it into a few metrics for the profile I'm looking for

For example, our 32 year-old on-loan targetman Andrea La Mantia is very likely to leave come the end of the season. Looking for a replacement 9, I've extracted the data for all strikers within our package who have started at least 5 games and are out of contract come the end of the season - all information that is readily available to clubs' recruitment teams
- Opening the export in excel, I then normalise a few selected statistics that I believe will help me rank the profile of player I want. Taking the targetman example again, I've normalised headers won (%), shots / 90, xG per shot, conversion rate, non-penalty xG per 90 and the net possession won versus possession lost.
- These are normalised by taking the average for all players in the export (removing those with zero returns as they will be in unplayable leagues), then comparing the player's output against this average. That comparison is effectively a percentage but I've set it to a numerical figure (100% = 1.00) for no other reason than I prefer the way 1.00 looks to 100%.
- Normalising in this way brings all the metrics into the same scale, allowing me to sum the metrics that I've chosen to be important for that profile (Column M below)

A quick sense-check against the names returned lets me know if I'm on the right track. Yussuf Poulsen, Mehdi Taremi and Duván Zapata being top of this list suggests I'm bang on the money. I can then look for names that I want my scouts to find out some more about. So Poulsen et al are clearly out - not exactly being realistic targets for FeralpiSalò. But Daniel Ciofani, coming to the end of his contract at fellow Serie B side Cremonese... that's more like it

Essentially, all I'm doing is conducting a statistical screen to produce targets for my scouts - rather than getting the output from a recruitment focus and then use stats to filter out that subset. Either option is entirely viable and I'm sure lots of people will think all this excel stuff is boring as, but who cares? This is how I like to play. Perhaps the analyst options in the Recruitment Foci will eventually become good enough that I can do this within the game. Perhaps not.

For now, this is how I want to play my game and how I'll be looking for signing targets come the summer. It let's me combine statistics into overall ranking scores, and create new statistics by combining data that FM separates - for example, identifying players who might have low goals per 90 metrics, but who score a large proportion of their team's goals. Are these good players hidden in a poor team? Perhaps. Using statistics like this will help me identify such players and ask my scouts to find out.

Specifically for my save, we've got a number of obvious targets - a 9 is key, as already mentioned. But our first choice XI features at least 6 loanees. That's not a comfortable position to be in, particularly when we don't have the money to secure any of them on permanent deals. So it could be a very busy second half of the season and an even busier summer.

Forza Feralpi!

Edited October 23, 2023 by Shrewnaldo

Shrewnaldo · October 23, 2023

4 hours ago, SixPointer said:

Here we go!! Love the idea of taking it to real life with the holiday Shrew.

The holidays came first, in this instance. I love Lake Garda. Will definitely be heading back once the kids are away.

4 hours ago, Jimbokav1971 said:

Loving the kits, but 1 & 3 seem too similar.

Liking the sound of this. You saves often throw up interesting discussions.

[Edit]

ps. Wow there are some big boys in Serie C.

I really like the third kit. So much so that I got one delivered today

robterrace · October 23, 2023

10 minutes ago, Shrewnaldo said:

Recruitment Strategies (Part 1)

Leoni del Garda - FeralpiSalò

FM used to be all about the tactics for me but, for various reasons, my interest in that side of the game has waned. Now I tend to focus more on the squad-building and recruitment is a huge part of that. Part of my increased interest has come with a focus on analytics, rather than just judging players by their numerical attributes. Indeed, I've taken to playing without the 1-20 attributes in recent versions. Whilst some will play entirely attributeless, it's just not for me. I don't think it's realistic that you wouldn't have any idea whatsoever about the players at your disposal - just think about the people that work for you, or with you, in real life. I'm sure you could assign some strengths and weaknesses to each person according to some basic work-related attributes. Or at least, a good manager should - and this is what the attributes represent for me.

Regardless, if I use the numerical attributes then I struggle to see past them, and just tend to let the numbers dominate my thoughts. Replacing them with coloured stars (or similar) in the last few versions has really helped me focus on statistics instead. Everyone is aware that stats are becoming more prevalent, if not ubiquitous, in football and whilst FM is some way behind the curve, in both the type of stats that are collected and their accuracy, there are a number of metrics available for players to use.

Whilst I wait for a graphical attributes skin, I thought I'd outline the strategy I'll be taking to analyse our own squad, in order to identify the recruitment priorities, and then how to find the players to fill those gaps.

I really like the addition of this experience matrix - allowing the player to very quickly assess the age profile of the squad and any gaps. Up until my Bristol City save (Statman and Robins), I'd have been like most FMers and aimed for a squad that fills out the left-hand side of this matrix - bringing in youth players that we can develop and then sell on for profit. Indeed, this is one of the many strategies that has been erroneously claimed as a 'Moneyball' approach to FM. Indeed, 'Moneyball' would probably be the opposite, with the book making numerous references to Bill James' conclusions that "(older) college players are a better investment than high school players by a huge, huge, laughably huge margin".

The reasoning behind this conclusion is that high-school baseball players do not have verifiable statistics from which franchises can draw direct conclusions about the player's likely professional future. It is this theme which I have carried through to FM - purchases will be based on evidence, and that comes from statistical output. Attributes and scout reports are not statistical outputs.

So that means I'm not going to sign a raft of 18 year-old wonderkids who have played just a handful of games each. I'm looking for a track record and will be keeping saves from the end of each season so that I can go back and check detailed statistics from each campaign, before they are wiped when the game refreshes itself in June / July.

In my Robins and Telstar saves, I also abandoned youth development completely - focusing solely on the senior squad. Here I'm going to have a halfway house - jettisoning the under-18 squad at the end of the season but keeping the under-20s. Primarily this is to honour the club vision and FeralpiSalò's real-life commitment to youth development. Sadly, the facilities are pretty poor and even this lip-service may be abandoned in a couple of seasons.

So it's all starting to look a bit Big Sam - low blocks, direct football and a focus on physical, experienced players. And like the much-maligned Mr Allardyce, we'll be heavily into the analytics. Both the Data Hub and custom views such as the below will be thoroughly used.

I've got a decent idea of the profile of players that I'm after, and a combination of these statistics and some basic logic will highlight the three to four profiles / positions that require strengthening. And then it's off shopping.

I really, really don't like the Recruitment Focus system that was introduced in FM23. It takes away a huge part of what I want to do myself, how I want to find players and what I want my scouts to do. So whilst I'll set up the odd focus just to see what comes back, it's not the primary means of finding players. Instead, I typically go through the following steps:

I use the Players in Range screen to extract a record of statistics from all the players within our scouting package

I used to avoid this screen like the plague, but I've recently accepted that it isn't the cheat screen I'd previously thought. It does not provide access to every player within the game. Instead, it's just those that your club would realistically know about. As FM now presents it to you, I see it as your club going to one of the data providers (Wyscout, Opta, etc) and buying access to their data across a given geographical range

To make it less 'gamey', I de-select the "interested" tickbox (so that all players are shown) and, whilst I always use attribute masking anyway, I never use it to search by attributes

I use custom views such as that shown below to extract the data into excel and from there manipulate it into a few metrics for the profile I'm looking for

For example, our 32 year-old on-loan targetman Andrea La Mantia is very likely to leave come the end of the season. Looking for a replacement 9, I've extracted the data for all strikers within our package who have started at least 5 games and are out of contract come the end of the season - all information that is readily available to clubs' recruitment teams

Opening the export in excel, I then normalise a few selected statistics that I believe will help me rank the profile of player I want. Taking the targetman example again, I've normalised headers won (%), shots / 90, xG per shot, conversion rate, non-penalty xG per 90 and the net possession won versus possession lost.

These are normalised by taking the average for all players in the export (removing those with zero returns as they will be in unplayable leagues), then comparing the player's output against this average. That comparison is effectively a percentage but I've set it to a numerical figure (100% = 1.00) for no other reason than I prefer the way 1.00 looks to 100%.

Normalising in this way brings all the metrics into the same scale, allowing me to sum the metrics that I've chosen to be important for that profile (Column M below)

A quick sense-check against the names returned lets me know if I'm on the right track. Yussuf Poulsen, Mehdi Taremi and Duván Zapata being top of this list suggests I'm bang on the money. I can then look for names that I want my scouts to find out some more about. So Poulsen et al are clearly out - not exactly being realistic targets for FeralpiSalò. But Daniel Ciofani, coming to the end of his contract at fellow Serie B side Cremonese... that's more like it

Essentially, all I'm doing is conducting a statistical screen to produce targets for my scouts - rather than getting the output from a recruitment focus and then use stats to filter out that subset. Either option is entirely viable and I'm sure lots of people will think all this excel stuff is boring as, but who cares? This is how I like to play. Perhaps the analyst options in the Recruitment Foci will eventually become good enough that I can do this within the game. Perhaps not.

For now, this is how I want to play my game and how I'll be looking for signing targets come the summer. It let's me combine statistics into overall ranking scores, and create new statistics by combining data that FM separates - for example, identifying players who might have low goals per 90 metrics, but who score a large proportion of their team's goals. Are these good players hidden in a poor team? Perhaps. Using statistics like this will help me identify such players and ask my scouts to find out.

Specifically for my save, we've got a number of obvious targets - a 9 is key, as already mentioned. But our first choice XI features at least 6 loanees. That's not a comfortable position to be in, particularly when we don't have the money to secure any of them on permanent deals. So it could be a very busy second half of the season and an even busier summer.

Forza Feralpi!

Really love this way of doing stuff, actually makes it far more workable as well.

Once I get my first season out of the way in the save I'm messing about in, I'm probably going to do a very similar system of recruitment for replacing players.

abulezz · October 23, 2023

I love how in-depth you get with your thought process. Great read thus far.

danielgear · October 23, 2023

13 hours ago, Shrewnaldo said:

Life imitating art here. FeralpiSalò has just relieved Stefano Vecchi of his job on the 23rd October 2023, 3 days after I created that scenario for this save.

How did you know?!

Following along Shrew, good luck with this save.

Shrewnaldo · October 23, 2023

56 minutes ago, robterrace said:

Really love this way of doing stuff, actually makes it far more workable as well.

Once I get my first season out of the way in the save I'm messing about in, I'm probably going to do a very similar system of recruitment for replacing players.

Ideal, if you've got any tips to share then I'm always happy to shamelessly steal stuff from other people's saves

51 minutes ago, abulezz said:

I love how in-depth you get with your thought process. Great read thus far.

Cheers, I appreciate that

First signing confirmed as 6'6" Norwegian striker will join on a free from Serie C club Ancona in the summer. Destined to be the back-up 9, Kristofferson scored a 7.38 on my Target Forward Rank, putting him 21st overall (of 361). Excellent aerial performance, as you'd expect for his height, and the 14th best xG/shot really stood out. He's not good enough to be first-choice, but I like him as the bench option a lot.

danielgear · October 23, 2023

1 hour ago, Shrewnaldo said:

Recruitment Strategies (Part 1)

Leoni del Garda - FeralpiSalò

FM used to be all about the tactics for me but, for various reasons, my interest in that side of the game has waned. Now I tend to focus more on the squad-building and recruitment is a huge part of that. Part of my increased interest has come with a focus on analytics, rather than just judging players by their numerical attributes. Indeed, I've taken to playing without the 1-20 attributes in recent versions. Whilst some will play entirely attributeless, it's just not for me. I don't think it's realistic that you wouldn't have any idea whatsoever about the players at your disposal - just think about the people that work for you, or with you, in real life. I'm sure you could assign some strengths and weaknesses to each person according to some basic work-related attributes. Or at least, a good manager should - and this is what the attributes represent for me.

Regardless, if I use the numerical attributes then I struggle to see past them, and just tend to let the numbers dominate my thoughts. Replacing them with coloured stars (or similar) in the last few versions has really helped me focus on statistics instead. Everyone is aware that stats are becoming more prevalent, if not ubiquitous, in football and whilst FM is some way behind the curve, in both the type of stats that are collected and their accuracy, there are a number of metrics available for players to use.

Whilst I wait for a graphical attributes skin, I thought I'd outline the strategy I'll be taking to analyse our own squad, in order to identify the recruitment priorities, and then how to find the players to fill those gaps.

I really like the addition of this experience matrix - allowing the player to very quickly assess the age profile of the squad and any gaps. Up until my Bristol City save (Statman and Robins), I'd have been like most FMers and aimed for a squad that fills out the left-hand side of this matrix - bringing in youth players that we can develop and then sell on for profit. Indeed, this is one of the many strategies that has been erroneously claimed as a 'Moneyball' approach to FM. Indeed, 'Moneyball' would probably be the opposite, with the book making numerous references to Bill James' conclusions that "(older) college players are a better investment than high school players by a huge, huge, laughably huge margin".

The reasoning behind this conclusion is that high-school baseball players do not have verifiable statistics from which franchises can draw direct conclusions about the player's likely professional future. It is this theme which I have carried through to FM - purchases will be based on evidence, and that comes from statistical output. Attributes and scout reports are not statistical outputs.

So that means I'm not going to sign a raft of 18 year-old wonderkids who have played just a handful of games each. I'm looking for a track record and will be keeping saves from the end of each season so that I can go back and check detailed statistics from each campaign, before they are wiped when the game refreshes itself in June / July.

In my Robins and Telstar saves, I also abandoned youth development completely - focusing solely on the senior squad. Here I'm going to have a halfway house - jettisoning the under-18 squad at the end of the season but keeping the under-20s. Primarily this is to honour the club vision and FeralpiSalò's real-life commitment to youth development. Sadly, the facilities are pretty poor and even this lip-service may be abandoned in a couple of seasons.

So it's all starting to look a bit Big Sam - low blocks, direct football and a focus on physical, experienced players. And like the much-maligned Mr Allardyce, we'll be heavily into the analytics. Both the Data Hub and custom views such as the below will be thoroughly used.

I've got a decent idea of the profile of players that I'm after, and a combination of these statistics and some basic logic will highlight the three to four profiles / positions that require strengthening. And then it's off shopping.

I really, really don't like the Recruitment Focus system that was introduced in FM23. It takes away a huge part of what I want to do myself, how I want to find players and what I want my scouts to do. So whilst I'll set up the odd focus just to see what comes back, it's not the primary means of finding players. Instead, I typically go through the following steps:

I use the Players in Range screen to extract a record of statistics from all the players within our scouting package

I used to avoid this screen like the plague, but I've recently accepted that it isn't the cheat screen I'd previously thought. It does not provide access to every player within the game. Instead, it's just those that your club would realistically know about. As FM now presents it to you, I see it as your club going to one of the data providers (Wyscout, Opta, etc) and buying access to their data across a given geographical range

To make it less 'gamey', I de-select the "interested" tickbox (so that all players are shown) and, whilst I always use attribute masking anyway, I never use it to search by attributes

I use custom views such as that shown below to extract the data into excel and from there manipulate it into a few metrics for the profile I'm looking for

For example, our 32 year-old on-loan targetman Andrea La Mantia is very likely to leave come the end of the season. Looking for a replacement 9, I've extracted the data for all strikers within our package who have started at least 5 games and are out of contract come the end of the season - all information that is readily available to clubs' recruitment teams

Opening the export in excel, I then normalise a few selected statistics that I believe will help me rank the profile of player I want. Taking the targetman example again, I've normalised headers won (%), shots / 90, xG per shot, conversion rate, non-penalty xG per 90 and the net possession won versus possession lost.

These are normalised by taking the average for all players in the export (removing those with zero returns as they will be in unplayable leagues), then comparing the player's output against this average. That comparison is effectively a percentage but I've set it to a numerical figure (100% = 1.00) for no other reason than I prefer the way 1.00 looks to 100%.

Normalising in this way brings all the metrics into the same scale, allowing me to sum the metrics that I've chosen to be important for that profile (Column M below)

A quick sense-check against the names returned lets me know if I'm on the right track. Yussuf Poulsen, Mehdi Taremi and Duván Zapata being top of this list suggests I'm bang on the money. I can then look for names that I want my scouts to find out some more about. So Poulsen et al are clearly out - not exactly being realistic targets for FeralpiSalò. But Daniel Ciofani, coming to the end of his contract at fellow Serie B side Cremonese... that's more like it

Essentially, all I'm doing is conducting a statistical screen to produce targets for my scouts - rather than getting the output from a recruitment focus and then use stats to filter out that subset. Either option is entirely viable and I'm sure lots of people will think all this excel stuff is boring as, but who cares? This is how I like to play. Perhaps the analyst options in the Recruitment Foci will eventually become good enough that I can do this within the game. Perhaps not.

For now, this is how I want to play my game and how I'll be looking for signing targets come the summer. It let's me combine statistics into overall ranking scores, and create new statistics by combining data that FM separates - for example, identifying players who might have low goals per 90 metrics, but who score a large proportion of their team's goals. Are these good players hidden in a poor team? Perhaps. Using statistics like this will help me identify such players and ask my scouts to find out.

Specifically for my save, we've got a number of obvious targets - a 9 is key, as already mentioned. But our first choice XI features at least 6 loanees. That's not a comfortable position to be in, particularly when we don't have the money to secure any of them on permanent deals. So it could be a very busy second half of the season and an even busier summer.

Forza Feralpi!

Really enjoyed this post. Love reading about the thoughts behind certain processes. (Even though you said you hate recruitment focus how dare you! :mad: )

But the process you are doing helps you recruit the way you want to.

I do agree with what you said about attribute/attributeless. I’m not a fan and I think the system can be evolved, a graphical representation is a decent compromise.

cmason84 · October 23, 2023

1 hour ago, Shrewnaldo said:

Recruitment Strategies (Part 1)

Leoni del Garda - FeralpiSalò

FM used to be all about the tactics for me but, for various reasons, my interest in that side of the game has waned. Now I tend to focus more on the squad-building and recruitment is a huge part of that. Part of my increased interest has come with a focus on analytics, rather than just judging players by their numerical attributes. Indeed, I've taken to playing without the 1-20 attributes in recent versions. Whilst some will play entirely attributeless, it's just not for me. I don't think it's realistic that you wouldn't have any idea whatsoever about the players at your disposal - just think about the people that work for you, or with you, in real life. I'm sure you could assign some strengths and weaknesses to each person according to some basic work-related attributes. Or at least, a good manager should - and this is what the attributes represent for me.

Regardless, if I use the numerical attributes then I struggle to see past them, and just tend to let the numbers dominate my thoughts. Replacing them with coloured stars (or similar) in the last few versions has really helped me focus on statistics instead. Everyone is aware that stats are becoming more prevalent, if not ubiquitous, in football and whilst FM is some way behind the curve, in both the type of stats that are collected and their accuracy, there are a number of metrics available for players to use.

Whilst I wait for a graphical attributes skin, I thought I'd outline the strategy I'll be taking to analyse our own squad, in order to identify the recruitment priorities, and then how to find the players to fill those gaps.

I really like the addition of this experience matrix - allowing the player to very quickly assess the age profile of the squad and any gaps. Up until my Bristol City save (Statman and Robins), I'd have been like most FMers and aimed for a squad that fills out the left-hand side of this matrix - bringing in youth players that we can develop and then sell on for profit. Indeed, this is one of the many strategies that has been erroneously claimed as a 'Moneyball' approach to FM. Indeed, 'Moneyball' would probably be the opposite, with the book making numerous references to Bill James' conclusions that "(older) college players are a better investment than high school players by a huge, huge, laughably huge margin".

The reasoning behind this conclusion is that high-school baseball players do not have verifiable statistics from which franchises can draw direct conclusions about the player's likely professional future. It is this theme which I have carried through to FM - purchases will be based on evidence, and that comes from statistical output. Attributes and scout reports are not statistical outputs.

So that means I'm not going to sign a raft of 18 year-old wonderkids who have played just a handful of games each. I'm looking for a track record and will be keeping saves from the end of each season so that I can go back and check detailed statistics from each campaign, before they are wiped when the game refreshes itself in June / July.

In my Robins and Telstar saves, I also abandoned youth development completely - focusing solely on the senior squad. Here I'm going to have a halfway house - jettisoning the under-18 squad at the end of the season but keeping the under-20s. Primarily this is to honour the club vision and FeralpiSalò's real-life commitment to youth development. Sadly, the facilities are pretty poor and even this lip-service may be abandoned in a couple of seasons.

So it's all starting to look a bit Big Sam - low blocks, direct football and a focus on physical, experienced players. And like the much-maligned Mr Allardyce, we'll be heavily into the analytics. Both the Data Hub and custom views such as the below will be thoroughly used.

I've got a decent idea of the profile of players that I'm after, and a combination of these statistics and some basic logic will highlight the three to four profiles / positions that require strengthening. And then it's off shopping.

I really, really don't like the Recruitment Focus system that was introduced in FM23. It takes away a huge part of what I want to do myself, how I want to find players and what I want my scouts to do. So whilst I'll set up the odd focus just to see what comes back, it's not the primary means of finding players. Instead, I typically go through the following steps:

I use the Players in Range screen to extract a record of statistics from all the players within our scouting package

I used to avoid this screen like the plague, but I've recently accepted that it isn't the cheat screen I'd previously thought. It does not provide access to every player within the game. Instead, it's just those that your club would realistically know about. As FM now presents it to you, I see it as your club going to one of the data providers (Wyscout, Opta, etc) and buying access to their data across a given geographical range

To make it less 'gamey', I de-select the "interested" tickbox (so that all players are shown) and, whilst I always use attribute masking anyway, I never use it to search by attributes

I use custom views such as that shown below to extract the data into excel and from there manipulate it into a few metrics for the profile I'm looking for

For example, our 32 year-old on-loan targetman Andrea La Mantia is very likely to leave come the end of the season. Looking for a replacement 9, I've extracted the data for all strikers within our package who have started at least 5 games and are out of contract come the end of the season - all information that is readily available to clubs' recruitment teams

Opening the export in excel, I then normalise a few selected statistics that I believe will help me rank the profile of player I want. Taking the targetman example again, I've normalised headers won (%), shots / 90, xG per shot, conversion rate, non-penalty xG per 90 and the net possession won versus possession lost.

These are normalised by taking the average for all players in the export (removing those with zero returns as they will be in unplayable leagues), then comparing the player's output against this average. That comparison is effectively a percentage but I've set it to a numerical figure (100% = 1.00) for no other reason than I prefer the way 1.00 looks to 100%.

Normalising in this way brings all the metrics into the same scale, allowing me to sum the metrics that I've chosen to be important for that profile (Column M below)

A quick sense-check against the names returned lets me know if I'm on the right track. Yussuf Poulsen, Mehdi Taremi and Duván Zapata being top of this list suggests I'm bang on the money. I can then look for names that I want my scouts to find out some more about. So Poulsen et al are clearly out - not exactly being realistic targets for FeralpiSalò. But Daniel Ciofani, coming to the end of his contract at fellow Serie B side Cremonese... that's more like it

Essentially, all I'm doing is conducting a statistical screen to produce targets for my scouts - rather than getting the output from a recruitment focus and then use stats to filter out that subset. Either option is entirely viable and I'm sure lots of people will think all this excel stuff is boring as, but who cares? This is how I like to play. Perhaps the analyst options in the Recruitment Foci will eventually become good enough that I can do this within the game. Perhaps not.

For now, this is how I want to play my game and how I'll be looking for signing targets come the summer. It let's me combine statistics into overall ranking scores, and create new statistics by combining data that FM separates - for example, identifying players who might have low goals per 90 metrics, but who score a large proportion of their team's goals. Are these good players hidden in a poor team? Perhaps. Using statistics like this will help me identify such players and ask my scouts to find out.

Specifically for my save, we've got a number of obvious targets - a 9 is key, as already mentioned. But our first choice XI features at least 6 loanees. That's not a comfortable position to be in, particularly when we don't have the money to secure any of them on permanent deals. So it could be a very busy second half of the season and an even busier summer.

Forza Feralpi!

Always love a good spreadsheet, so will be keeping up to date on this

keeper#1 · October 23, 2023

Definitely following along.

Lestri · October 24, 2023

One of the things I've loved seeing is people's approach to statistical analysis, and really looking forward to how you approach yours, especially after the Target Forward extract above!

wils2603 · October 24, 2023

10 hours ago, Shrewnaldo said:

Recruitment Strategies (Part 1)

Leoni del Garda - FeralpiSalò

FM used to be all about the tactics for me but, for various reasons, my interest in that side of the game has waned. Now I tend to focus more on the squad-building and recruitment is a huge part of that. Part of my increased interest has come with a focus on analytics, rather than just judging players by their numerical attributes. Indeed, I've taken to playing without the 1-20 attributes in recent versions. Whilst some will play entirely attributeless, it's just not for me. I don't think it's realistic that you wouldn't have any idea whatsoever about the players at your disposal - just think about the people that work for you, or with you, in real life. I'm sure you could assign some strengths and weaknesses to each person according to some basic work-related attributes. Or at least, a good manager should - and this is what the attributes represent for me.

Regardless, if I use the numerical attributes then I struggle to see past them, and just tend to let the numbers dominate my thoughts. Replacing them with coloured stars (or similar) in the last few versions has really helped me focus on statistics instead. Everyone is aware that stats are becoming more prevalent, if not ubiquitous, in football and whilst FM is some way behind the curve, in both the type of stats that are collected and their accuracy, there are a number of metrics available for players to use.

Whilst I wait for a graphical attributes skin, I thought I'd outline the strategy I'll be taking to analyse our own squad, in order to identify the recruitment priorities, and then how to find the players to fill those gaps.

I really like the addition of this experience matrix - allowing the player to very quickly assess the age profile of the squad and any gaps. Up until my Bristol City save (Statman and Robins), I'd have been like most FMers and aimed for a squad that fills out the left-hand side of this matrix - bringing in youth players that we can develop and then sell on for profit. Indeed, this is one of the many strategies that has been erroneously claimed as a 'Moneyball' approach to FM. Indeed, 'Moneyball' would probably be the opposite, with the book making numerous references to Bill James' conclusions that "(older) college players are a better investment than high school players by a huge, huge, laughably huge margin".

The reasoning behind this conclusion is that high-school baseball players do not have verifiable statistics from which franchises can draw direct conclusions about the player's likely professional future. It is this theme which I have carried through to FM - purchases will be based on evidence, and that comes from statistical output. Attributes and scout reports are not statistical outputs.

So that means I'm not going to sign a raft of 18 year-old wonderkids who have played just a handful of games each. I'm looking for a track record and will be keeping saves from the end of each season so that I can go back and check detailed statistics from each campaign, before they are wiped when the game refreshes itself in June / July.

In my Robins and Telstar saves, I also abandoned youth development completely - focusing solely on the senior squad. Here I'm going to have a halfway house - jettisoning the under-18 squad at the end of the season but keeping the under-20s. Primarily this is to honour the club vision and FeralpiSalò's real-life commitment to youth development. Sadly, the facilities are pretty poor and even this lip-service may be abandoned in a couple of seasons.

So it's all starting to look a bit Big Sam - low blocks, direct football and a focus on physical, experienced players. And like the much-maligned Mr Allardyce, we'll be heavily into the analytics. Both the Data Hub and custom views such as the below will be thoroughly used.

I've got a decent idea of the profile of players that I'm after, and a combination of these statistics and some basic logic will highlight the three to four profiles / positions that require strengthening. And then it's off shopping.

I really, really don't like the Recruitment Focus system that was introduced in FM23. It takes away a huge part of what I want to do myself, how I want to find players and what I want my scouts to do. So whilst I'll set up the odd focus just to see what comes back, it's not the primary means of finding players. Instead, I typically go through the following steps:

I use the Players in Range screen to extract a record of statistics from all the players within our scouting package

I used to avoid this screen like the plague, but I've recently accepted that it isn't the cheat screen I'd previously thought. It does not provide access to every player within the game. Instead, it's just those that your club would realistically know about. As FM now presents it to you, I see it as your club going to one of the data providers (Wyscout, Opta, etc) and buying access to their data across a given geographical range

To make it less 'gamey', I de-select the "interested" tickbox (so that all players are shown) and, whilst I always use attribute masking anyway, I never use it to search by attributes

I use custom views such as that shown below to extract the data into excel and from there manipulate it into a few metrics for the profile I'm looking for

For example, our 32 year-old on-loan targetman Andrea La Mantia is very likely to leave come the end of the season. Looking for a replacement 9, I've extracted the data for all strikers within our package who have started at least 5 games and are out of contract come the end of the season - all information that is readily available to clubs' recruitment teams

Opening the export in excel, I then normalise a few selected statistics that I believe will help me rank the profile of player I want. Taking the targetman example again, I've normalised headers won (%), shots / 90, xG per shot, conversion rate, non-penalty xG per 90 and the net possession won versus possession lost.

These are normalised by taking the average for all players in the export (removing those with zero returns as they will be in unplayable leagues), then comparing the player's output against this average. That comparison is effectively a percentage but I've set it to a numerical figure (100% = 1.00) for no other reason than I prefer the way 1.00 looks to 100%.

Normalising in this way brings all the metrics into the same scale, allowing me to sum the metrics that I've chosen to be important for that profile (Column M below)

A quick sense-check against the names returned lets me know if I'm on the right track. Yussuf Poulsen, Mehdi Taremi and Duván Zapata being top of this list suggests I'm bang on the money. I can then look for names that I want my scouts to find out some more about. So Poulsen et al are clearly out - not exactly being realistic targets for FeralpiSalò. But Daniel Ciofani, coming to the end of his contract at fellow Serie B side Cremonese... that's more like it

Essentially, all I'm doing is conducting a statistical screen to produce targets for my scouts - rather than getting the output from a recruitment focus and then use stats to filter out that subset. Either option is entirely viable and I'm sure lots of people will think all this excel stuff is boring as, but who cares? This is how I like to play. Perhaps the analyst options in the Recruitment Foci will eventually become good enough that I can do this within the game. Perhaps not.

For now, this is how I want to play my game and how I'll be looking for signing targets come the summer. It let's me combine statistics into overall ranking scores, and create new statistics by combining data that FM separates - for example, identifying players who might have low goals per 90 metrics, but who score a large proportion of their team's goals. Are these good players hidden in a poor team? Perhaps. Using statistics like this will help me identify such players and ask my scouts to find out.

Specifically for my save, we've got a number of obvious targets - a 9 is key, as already mentioned. But our first choice XI features at least 6 loanees. That's not a comfortable position to be in, particularly when we don't have the money to secure any of them on permanent deals. So it could be a very busy second half of the season and an even busier summer.

Forza Feralpi!

Superb. Love this approach. Always enjoyed your blogs in the past. I would be interested to know in your experience how successful and I guess transferable promising statistical player metrics translate to in game performances. It’s something I like to look at but not in any massive detail, more of a good sense check for me. I mean we’ve all had players in saves who’s attributes might not be the strongest but they always perform “above their attributes” and I wonder if “in game” this approach is more vindicated with every release as IRL how important these metrics are. Like you say FM is behind the curve, but im interested to hear what you would deem successful in terms of player performance and would you give more weighting to the overall team performance and result rather than say a player who hit the right score in your statistical metrics and came into the team and although the team played and team results were successful but that individual player had a poor or average average rating etc?

Shrewnaldo · October 24, 2023

10 hours ago, danielgear said:

Really enjoyed this post. Love reading about the thoughts behind certain processes. (Even though you said you hate recruitment focus how dare you!)

But the process you are doing helps you recruit the way you want to.

I do agree with what you said about attribute/attributeless. I’m not a fan and I think the system can be evolved, a graphical representation is a decent compromise.

I hope they never get rid of the 1-20 attributes - but I like the idea that SI provide various options. Something like the coloured stars I've been using, or the bars that others have used, would be a nice alternative to have available.

10 hours ago, cmason84 said:

Always love a good spreadsheet, so will be keeping up to date on this

10 hours ago, keeper#1 said:

Definitely following along.

4 hours ago, Lestri said:

One of the things I've loved seeing is people's approach to statistical analysis, and really looking forward to how you approach yours, especially after the Target Forward extract above!

Thanks all, the more the merrier

51 minutes ago, wils2603 said:

Superb. Love this approach. Always enjoyed your blogs in the past. I would be interested to know in your experience how successful and I guess transferable promising statistical player metrics translate to in game performances. It’s something I like to look at but not in any massive detail, more of a good sense check for me. I mean we’ve all had players in saves who’s attributes might not be the strongest but they always perform “above their attributes” and I wonder if “in game” this approach is more vindicated with every release as IRL how important these metrics are. Like you say FM is behind the curve, but im interested to hear what you would deem successful in terms of player performance and would you give more weighting to the overall team performance and result rather than say a player who hit the right score in your statistical metrics and came into the team and although the team played and team results were successful but that individual player had a poor or average average rating etc?

Regarding how transferable statistics are... watch this space. I've got a couple of thoughts.

I love those players that you have when they're output far outweighs what their attributes would suggest - think they're my favourite sort of players. I still remember Antonino La Gumina that I had in my old Samp save (funnily enough, he's since joined irl). His attributes were average but he'd always produce - no matter where I played him.

But I like your point about individual versus team performances, it's often hard to see the wider picture. That's why I add "team conceded per 90", "team scored per 90" and "points per 90" to all my statistics view. This means I can check to see if the team is performing unusually well / poorly when a particular player is on the pitch. That then prompts me to go look deeper at the individual stats to see if I'm missing anything.

danielgear · October 24, 2023

Just now, Shrewnaldo said:

hope they never get rid of the 1-20 attributes - but I like the idea that SI provide various options. Something like the coloured stars I've been using, or the bars that others have used, would be a nice alternative to have available

Yeah they won’t and shouldn’t remove them, but they used to have bars and for some reason removed them

Shrewnaldo · October 24, 2023

Recruitment Strategies (Part 2)

Leoni del Garda - FeralpiSalò

It's been mentioned above, but one of the difficulties with the statistical approach has always been comparing statistics across different leagues. How comparable is a striker with 0.7 xG/90 in the Ekstraklasa to a striker with 0.5 xG/90 in Liga NOS? The latter is the stronger league but by how much?

I wanted to come up with a quantifiable way of comparing the leagues so I spent the first day of FM24 creating a manager in almost every playable league in the game. Then, using the Comparison tables (under Squad Planner > Report) I manually recorded the average attributes, wage and transfer value. I was then able to produce average values for the Physical, Mental, Technical and Overall attributes in every single top-tier, plus a few lower tiers. Like this:

The first thing to notice is that the 'average value' is absolutely useless. I think the value ranges that were introduced in FM22 (?) have borked this completely and no-one has noticed. South Korea has the highest average value in the game at £36m, which is clearly just wrong. So I'm going to ignore that metric completely.

The next moderately interesting conclusion is the overall average:

Just highlighting the top few leagues, you can see that Spain is 'the strongest' with the highest average attributes whilst Brazil is in fifth with Mexico seventh - all three benefitting from higher technical averages than any other league. So for those managing big clubs, it seems to make a lot of sense to go shopping here.

Making statistics comparable

However, my objective with this task was to quantify the differences between the leagues. Having these raw numbers I could then normalise the results to compare league to league, and country to country. Again, I've quantified this as 1.00 cf 100% purely because of personal aesthetic preference and have produced the figures for the recorded top-tiers below:

Ok, so what?

Well I hope this means that I can adjust the statistical output between countries and leagues, to provide make the metrics comparable, albeit accepting that it's never going to be perfect. The obvious conclusion I'm trying to draw here is that playing against opposition with average attributes of 10.46 in Switzerland is slightly easier than playing against opponents with average attributes of 11.56 in Italy. So when I'm producing the normalised ranking from the statistical models in Part 1, I can adjust these figures by multiplying by the corresponding Deviation figure. Making up an example to illustrate:

Kaly Sene, playing for Lausanne, has a statistical output which gives him a 6.5 false 9 ranking from my chosen metrics. Adjusting by the Deviation for the Swiss league, his adjusted score would be (6.5*1.016) = 6.604
Marko Djuricin, playing for Spartak Trnava, has a 6.8 false 9 ranking. Adjusting for the Deviation from the Slovak league, his adjusted score would be (6.8*0.938) = 6.38

Is this fair? Is it accurate? Will it work? I haven't got the slightest scoob but, like I've said a few times, I just want to try something different and am willing to give it a shot.

Finding markets with value

The second conclusion I hope to be able to draw from this data extract is where transfer value can be found. We've already seen that the transfer value averages are broken, so we can't use that to find leagues with players who are cheap relative to their attributes. However, we can compare the actual attribute averages to the reputation of the league - reputation being one of the key factors in driving transfer value in Football Manager.

To do this, I compared the 'FM Ranking' of each league (World > Competitions > Leagues) to their ranking by average attributes. I've ordered the below by the offset in their ranking, i.e. how much their attribute ranking deviates from their reputation ranking. Positive figures means their attributes are better than their reputation would suggest.

This one is going to come with some heavy caveats:

some of the lower rep leagues will have artificially high offsets because higher reputation leagues are unplayable and therefore I have not captured average attribute data
some of the non-European leagues will be affected by the bias towards Europe in terms of reputation
reputation is not the only show in town when it comes to valuation - therefore I have also included the average wage from each league which will provide some indication of the league's financial strength

So what, if anything, can we draw from this? In my view, there are a couple of theories we can test from it:

There may be value in shopping in the second tiers of big leagues. Spain, Germany, France and Italy all perform well whilst maintaining a solid attribute average that is higher than the global average (10.31)
However, the second tiers in smaller leagues, whilst having a solid offset, have average attribute levels which make finding value unlikely
Romania, Colombia and Uruguay look like fertile shopping grounds - solid average attributes above the global average, comparatively low reputation and low average wages
Denmark and Mexico may also present an opportunity but likely at higher cost, given their higher starting reputation and higher average wages
Despite having solid average attribute values, the likes of Norway, Portugal, Belgium, the Netherlands and Austria may not offer the best value. Of course, if you're a really big club then who cares about value. Just splash the cash and these leagues definitely offer fertile talent

For my own game, then, I'll be looking at Romania primarily. Serie B clubs can't sign non-EU players from outside Italy so Uruguay and Colombia are out (for now). And I'll be looking to compare the statistical output from these players using the Deviation offset.

I'd welcome any thoughts on the above. Have I missed something obvious? Am I barking up the wrong tree? Have I bored you massively? Are you a walrus?

All important questions.

Forza Feralpi!

Edited October 24, 2023 by Shrewnaldo

Lestri · October 24, 2023

This is fantastic and baselining statistical outputs weighted by league rating has arguably been my biggest blocker for going full statistical nerd with FM. Mainly because I run it on a piece of crap laptop so I haven't been able to fully deep dive, so massive plaudits for doing this!

I use a score out of 100 for role suitability based on player's attribute, but I've always been stuck on the statistical side of things outside of using the likes of FMStag's for baselining what is good. So this gives me plenty to ponder about, considering you've made the leap of linking stats with specific player roles.

Long winded way of saying what a bloody great job.

robterrace · October 24, 2023

Love the way this is beginning to look., As long as you stick to your own rules when running the numbers, especially when it involves an unplayable league, then, things should work.

As for the 'weighting', I take it that you're using the overall average to work that out, although, does the division of players you're looking at add another variable to look into as well?

vkastanas · October 24, 2023

2 hours ago, Shrewnaldo said:

I'd welcome any thoughts on the above. Have I missed something obvious? Am I barking up the wrong tree? Have I bored you massively? Are you a walrus?

I don't really have any thoughts cause you've blown me away really! Am I bored? Not at all!! I'm just loving all of this!

Any thoughts on sharing those views and those excels?

ifinnem · October 24, 2023

Always love to follow your threads Shrew.

Noy new, but the big question for me is how do you plan to use stats to find good defenders and DMs given that things like interceptions, tackles are a function of role and team setup as much as individuals skills.

I guess you also miss out on big retraining opportunities that I think are great ways to find value, eg move a CD to DM, DM to FB, AM to ST etc

Shrewnaldo · October 24, 2023

6 hours ago, Lestri said:

This is fantastic and baselining statistical outputs weighted by league rating has arguably been my biggest blocker for going full statistical nerd with FM. Mainly because I run it on a piece of crap laptop so I haven't been able to fully deep dive, so massive plaudits for doing this!

I use a score out of 100 for role suitability based on player's attribute, but I've always been stuck on the statistical side of things outside of using the likes of FMStag's for baselining what is good. So this gives me plenty to ponder about, considering you've made the leap of linking stats with specific player roles.

Long winded way of saying what a bloody great job.

Thanks, much appreciated.

4 hours ago, robterrace said:

Love the way this is beginning to look., As long as you stick to your own rules when running the numbers, especially when it involves an unplayable league, then, things should work.

As for the 'weighting', I take it that you're using the overall average to work that out, although, does the division of players you're looking at add another variable to look into as well?

You mean dividing the attribute splits between the positions? Thinking perhaps that certain leagues will tend to produce better defenders and therefore it's better a better market?

It's possible to do that but it does add workload to creating the baseline. Even just splitting by 'keepers, defenders, midfielders and strikers quadruples the effort. Splitting by sides adds again etc etc. For me, the effort would outweigh the benefit.

4 hours ago, vkastanas said:

I don't really have any thoughts cause you've blown me away really! Am I bored? Not at all!! I'm just loving all of this!

Any thoughts on sharing those views and those excels?

I'm happy to share the data export and the views, if folk think they'd be useful. But I'd encourage people to use the data in their own way rather than copy my logic.

57 minutes ago, ifinnem said:

Always love to follow your threads Shrew.

Noy new, but the big question for me is how do you plan to use stats to find good defenders and DMs given that things like interceptions, tackles are a function of role and team setup as much as individuals skills.

I guess you also miss out on big retraining opportunities that I think are great ways to find value, eg move a CD to DM, DM to FB, AM to ST etc

Indeed, I'm not likely to be doing much retraining. With the older squad, it's less likely anyway - but it is much more difficult to identify opportunities for re-training without the numerical attributes.

Re defensive stats - it is tough. I'd really like to be able to make possession-adjusted stats but FM doesn't record the possession average when a player is on the pitch. Instead, what I can do is use the overall numbers to produce a longlist that I want the scouts to investigate. This then allows me to whittle the list down to 5-10 targets, and that's then very easy to go and find their team's average possession manually - and thereby produce possession adjusted stats.

Obviously this still isn't perfect, but it's better.

robterrace · October 24, 2023

7 minutes ago, Shrewnaldo said:

You mean dividing the attribute splits between the positions? Thinking perhaps that certain leagues will tend to produce better defenders and therefore it's better a better market?

It's possible to do that but it does add workload to creating the baseline. Even just splitting by 'keepers, defenders, midfielders and strikers quadruples the effort. Splitting by sides adds again etc etc. For me, the effort would outweigh the benefit.

I was thinking more in terms of comparing a Serie B player to a Serie A player, and a Premier League player to a Championship one (or similar). How much of a difference does the level of a play have in the overall weighting compared to the national weightings?

robterrace · October 24, 2023

So, assuming a PL Striker has a weighting of 1.00, and a Championship one, a .90 (for example), do you take that into consideration before applying the national weighting when comparing to a Spanish based player.

Shrewnaldo · October 24, 2023

1 minute ago, robterrace said:

I was thinking more in terms of comparing a Serie B player to a Serie A player, and a Premier League player to a Championship one (or similar). How much of a difference does the level of a play have in the overall weighting compared to the national weightings?

Ah ok, well I have the second tier covered for the major nations and a few selected others; and I've also covered Serie C. So I can also adjust the rankings for any of these leagues. And if I stumble across someone in a competition that I haven't covered, then it should be easy enough to add them to the baseline retrospectively.

robterrace · October 24, 2023

Just now, Shrewnaldo said:

Ah ok, well I have the second tier covered for the major nations and a few selected others; and I've also covered Serie C. So I can also adjust the rankings for any of these leagues. And if I stumble across someone in a competition that I haven't covered, then it should be easy enough to add them to the baseline retrospectively.

Thanks, thats what I was thinking, especially with you commenting earlier that you weren't going to go for the highest level of player available.

vkastanas · October 24, 2023

1 hour ago, Shrewnaldo said:

But I'd encourage people to use the data in their own way rather than copy my logic.

That's fair enough for me!

Shrewnaldo · October 24, 2023

2 hours ago, vkastanas said:

That's fair enough for me!

The raw data is attached, to be done with as you will.

FM24 data export.xlsx

vkastanas · October 25, 2023

7 hours ago, Shrewnaldo said:

The raw data is attached, to be done with as you will.

Thanks a lot! May I ask?

What is the Average Value that is the hidden Column "C"?

Why Italy 3 has the characteristic excel sign that something is different from the other data?

Shrewnaldo · October 25, 2023

1 hour ago, vkastanas said:

Thanks a lot! May I ask?

What is the Average Value that is the hidden Column "C"?

Why Italy 3 has the characteristic excel sign that something is different from the other data?

The "average value" is the average transfer value in that league. As I referred to above, this just appears to be broken. I don't think it's able to deal with the value ranges that were introduced in fm22. South Korea has the highest average transfer value in the game.

Because it doesn't work, I just hid the column

Italy 3 is different because I've had to capture the three sub-leagues of Serie C separately. I've then averaged them into the line that I've left visible so there's only one record for the Italian thirdly tier

vkastanas · October 25, 2023

5 minutes ago, Shrewnaldo said:

The "average value" is the average transfer value in that league. As I referred to above, this just appears to be broken. I don't think it's able to deal with the value ranges that were introduced in fm22. South Korea has the highest average transfer value in the game.

Because it doesn't work, I just hid the column

Italy 3 is different because I've had to capture the three sub-leagues of Serie C separately. I've then averaged them into the line that I've left visible so there's only one record for the Italian thirdly tier

Nice thanks!

And what about those views and search filters! Will you share those! I know I can make them if you don't bother sharing...

vkastanas · October 25, 2023

14 minutes ago, Shrewnaldo said:

Italy 3 is different because I've had to capture the three sub-leagues of Serie C separately. I've then averaged them into the line that I've left visible so there's only one record for the Italian thirdly tier

But look, if I select cell I33 (Acceleration on Italy 3avg) the formula says that is the average of cells (I30:I33, South Korea, Poland, Uruguay). Or am I seeing somethig wrong?

Shrewnaldo · October 25, 2023

7 minutes ago, vkastanas said:

But look, if I select cell I33 (Acceleration on Italy 3avg) the formula says that is the average of cells (I30:I33, South Korea, Poland, Uruguay). Or am I seeing somethig wrong?

Ah that's weird. It should be the average of the three individual Serie C cells. Looks like something has gone wrong in transit there. I'll take a look later today

SteinkelssonFM · October 25, 2023

On 24/10/2023 at 11:24, Shrewnaldo said:

you can see that Spain is 'the strongest' with the highest average attributes

Would have been interesting to also see the average age of the league's, to gain a bit of context behind the numbers. My gut feeling is that the Spanish league holds a higher average age and therefore the reason as to why technicals are higher due to more time to perfect attributes.

I know this wasn't the reason why you did the deep dive, the data is really useful... you need to make an infographic and post it on Twitter.

nugatti · October 25, 2023

On 24/10/2023 at 12:24, Shrewnaldo said:

Romania, Colombia and Uruguay look like fertile shopping grounds - solid average attributes above the global average, comparatively low reputation and low average wages

Just a shout out for Poland, Sweden, Croatia as good markets. How do you rate them?

Edited October 25, 2023 by nugatti
Typo

Shrewnaldo · October 26, 2023

20 hours ago, MattyLewis11 said:

Would have been interesting to also see the average age of the league's, to gain a bit of context behind the numbers. My gut feeling is that the Spanish league holds a higher average age and therefore the reason as to why technicals are higher due to more time to perfect attributes.

I know this wasn't the reason why you did the deep dive, the data is really useful... you need to make an infographic and post it on Twitter.

Not sure about the infographic... have you seen my graphical skills?!

Re the age profile, that'd be fairly quick to check - could just start up a new game then add a manager to any club in the league and use the comparison page. Pretty sure it has the average age to 2 decimal points? I'll check later

18 hours ago, nugatti said:

Just a shout out for Poland, Sweden, Croatia as good markets. How do you rate them?

I'm sure they'll all have some great talents - I've certainly found some brilliant young players in Croatia in the past.

But, just using this method, they're not as likely to have value as the other leagues. Poland seems the most likely of three with a slightly higher than global attribute average and a slightly higher attribute rank than reputation rank; Sweden is over-ranked and Croatia is well over-ranked.

*But* this method is very broadbrush - it's talking about the leagues as averages. If you limited Croatia to just Dinamo Zagreb, then you know that you're much more likely to find a high-quality newgen or fourteen. Same with Brommapojkarna. As I'm a Serie B club with less than no money, I need to cast my net as wide as possible to pick up bargains and free transfers, so I wanted to get away from just targeting the well-known talent factories.

Shrewnaldo · October 26, 2023

On 25/10/2023 at 19:55, MattyLewis11 said:

Would have been interesting to also see the average age of the league's, to gain a bit of context behind the numbers. My gut feeling is that the Spanish league holds a higher average age and therefore the reason as to why technicals are higher due to more time to perfect attributes.

I know this wasn't the reason why you did the deep dive, the data is really useful... you need to make an infographic and post it on Twitter.

Average age in La Liga is 26.67; in the Premier League it's 26.11. So about half a year's difference, which may have some difference on the attributes. The Bundesliga, just for a comparison, is 25.64.

Shrewnaldo · October 27, 2023

On 25/10/2023 at 08:53, vkastanas said:

But look, if I select cell I33 (Acceleration on Italy 3avg) the formula says that is the average of cells (I30:I33, South Korea, Poland, Uruguay). Or am I seeing somethig wrong?

So it looks like I've just messed up for that line. It is supposed to be the average of the three Serie C rows, but the formula has swapped out. If you unhide between rows 4 and 10, then you'll reveal the Serie C lines and then be able to correct it again. No idea how that happened but good spot

Sign In

[FM24] The Traipsing of the Shrew

Recommended Posts

Link to post

Share on other sites

Top Posters In This Topic

Top Posters In This Topic

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites