Mathematically assessing your squad

I always have trouble assessing a squad in a structured way when I start a new save or move jobs. I therefore thought I'd take a stab at working on a set process for assessing your squad at the start of a new save.

I've got this idea that players can be assessed by:

(a) looking at the average of the key attributes for each position they can play; and

(b) whether all their attributes cluster around that average or are spread out

I've put together an example of (a) using Nottingham Forest's Defenders so you can get a sense of what I mean here: https://docs.google.com/spreadsheets/d/1bMnWiWg7FKb-dB6SXVoMFwIcR3qROz9VO0DLTA3_lhM/edit?pli=1#gid=2050460887 (scroll right to see the averages)

My idea is that you can sort your players into five categories:

1. Players for whom the average of their positions' attributes is below divisional average

- Look to get rid of these

2. Players for whom the average of their positions' attributes is around the divisional average AND who’s individual core and mental attributes cluster around their average

- These will be solid players who will fit into their roles/duties in any given situation

3. Players for whom the average of their positions' attributes is around the divisional average AND who’s individual core and mental attributes are spread over a large range

- These may or may not be worth keeping. Check what they excel at and see if they can fit in with your tactics. Keep an eye on them in-match to make sure their strength are delivering for you and their weaknesses aren’t being constantly exploited

4. Players for whom the average of their positions' attributes is above the divisional average AND who’s individual core and mental attributes cluster around their average

- Keep these - they’re really great players

5. Players for whom the average of their positions' attributes is above the divisional average AND who’s individual core and mental attributes are spread over a large range

- Keep these - they’re really great players. Additionally, identify the exceptional attributes that are bringing their average up and look to take advantage of them in your tactics.

My suspicion is that I should be expanding (a) beyond each position's 'key' attributes to include a small number of mental stats (the same for all positions?).

Just wondered what everyone thinks?

I don't know - for example a central defender I might want to keep just for his jumping ability and positioning, even if he's slow - if I would generally play deep and force play down the flanks. would not matter ot m if the average attributes for his position are below league average. On the other hand I might just sell a central defender with bad positioning and anticipation even if his other defensive stats look great if I push up a lot.

Also bear in mind some players are worth keeping just for their personality (tutoring; performing above expectations when mentality is good) or for their specific preferred moves (either because they fit into a tactic or because you want to try and transfer them more easily to younger players via tutoring again.

and then again... league average might just not be good enough to hold down a place in the squad if I'm stakced at a position and have promising youngsters as well. I won't take away playing time from promising players to keep league avrage guys happy (unless they are first choice for the time being...)

Not all attributes are worth the same. For example, mental attributes like anticipation, concentration, positioning, etc. are arguably far more important for a central defender than heading, marking or tackling. Then, for some positions, physical stats are of prime importance (possibly for wingers, wing backs, etc.), and for some positions mental attributes are far more important. Next thing, a central defender could have 18-20 in every key attribute (except jumping reach), and still be very short and be an aerial liability. Etc. etc. etc.

Attributes should really be weighted, but how could that be done? It would be very difficult to estimate just how important is each stat.

Don't get me wrong; I applaud your efforts, and generally love to see mathematical analysis of in-game things. But this must be thought out from many angles.

Thank bieritarier and genesimmons. Going through the points you raise...

"a central defender I might want to keep just for his jumping ability and positioning, even if he's slow"

I don't think that Pace is included in the list of "key" attributes for a DC, but I take your point - that you'd be fine if a player had some low "key" attributes if those were compensated by other "key" attributes being high. I'd hoped that, by taking the average of all the key attributes, I was allowing for this? i.e. that a DC with 15 in all of his key attributes was neither better nor worse than a DC with 10 Decisions, 10 Determination, 20 Heading, 20 Positioning and 15 for the rest. - the same average.

"bear in mind some players are worth keeping just for their personality"

Totally agree. This model is just about the basics of attribute assessment. Maybe once I've got that down I could expand it to include this kind of thing.

"then again... league average might just not be good enough to hold down a place"

Totally agree. Managing a team towards the bottom of the table, you might want to be getting players in that are league average (say, to stay up) but if you're established, you'd want to set the "average" your team aspires to a little higher.

"Not all attributes are worth the same. For example, mental attributes like anticipation, concentration, positioning, etc. are arguably far more important for a central defender than heading, marking or tackling"

I think that's the next stage in refining this. Basically it's two questions:

(1) Are there non-"key" attributes that should be included - either for everyone or for certain positions. i.e. can the the in-game list of key attributes be improved upon

(2) To what extent should the list of attributes for each position used in this model be weighted. Should some count double? Triple? And, if so, which ones?

To answer your questions:

(1) It depends. If you would like to form a squad with high work ethic, it would make sense to include Work Rate. If you would like to make a team that will fight to the end, including Determination makes sense. However, I don't think there are clear rules. Some attributes, like background ones (Concentration or Decisions for example), are arguably more important than the prime and seconday ones.

(2) I don't think that any should be doubled, but, a lot of them would be in the 1.1x-1.7x range, I think. How can one know!? Guys who know how the match engine actually works could tell us a lot more.

Thanks genesimmons.

I definitely agree that the choice between two players is often about the type of team you're trying to build - but I think we can all agree that there are a minimum set of attributes for a particular position for a particular division.

I mean, if I was managing Manchester United or City, Arsenal or Chelsea, whatever type of team - hardworking or fast or technically brilliant or intelligent - I'm trying to build, I'm never going to buy Danny Collins (attributes below). The system I'm looking at here is about filtering out the players that are not making the cut before you start making the more nuanced decisions on the type of team you want to build.

I don't think such an approach makes any sense. Reasons:

a) Time investment of establishing and updating such a sheet as well as calculating league averages.

b) Team gelling. Some players will perform for you even if they are punching over their weight. I've seen this happen often in saves where I have moved up from a lower league and have club icons who can put in decent performances in the champions league even if they are championship level players.

c) Attribute roles can only be seen as basic guidelines. As a manager you have to find out what works for you, particularly pace is advantageous in pretty much all positions. Me personally I want defensively solid passers in central midfield and fast fullbacks that can cross. I also don't want Centerbacks with any glaring weaknesses.

d) Variance in performance. It is my observation that players do not maintain stable performance levels over seasons but that it is not uncommon for players (particularly strikers) to have average performance ratings varying by .3 points or more between seasons, thus I try to avoid offloading players after one bad season. Likewise I try to sell players at a good price if I feel they have been overperforming for me to make a profit.

Once upon a time, before personal computers and football management games, I used to 'rate' footballers by certain key attributes, though with much less exposure than we get now (attending Liverpool games and watching MOTD). I'd play with various 'weightings' of those attributes by position. Of course, those weightings will be determined by what sort of player you want, what sort of system you prefer. I must have used thousands of pieces of paper 'rating' players and 'weighting' attributes, with no real end purpose to it, other than to confirm that Graeme Souness was a better player than Glenn Hoddle.

But FM has a lot of players. Without being able to do a CSV dump of players with attributes (I'm assuming we can't, I never really looked), it's too much info. The star rating gives a basic version of this process, though without the 'customisation' of using our own weighting preferences. The player search can focus on key attributes (and insist on X/Y of them), but again lacks a weighting. To a degree, one can glance at a player and assess the colour-coding of attributes in certain places in the profile, depending on position.

You could do this with your own squad, then plug in certain scouted players to see where they would fit within your own system and perhaps, if you should buy them. But I wouldn't see the need to compare against some 'league average', unless (via CSV dump etc) that became much easier to do and to automate.

Isn't this why we have tools like Genie Scout?

I guess I'm trying to come up with an answer to the question "How should I go about assessing all my individual players at the start of a new job or save" that goes beyond the glib:

• "Just look at key attributes and mentals"
  
• "It depends on what league you're in"
  
• "It depends on their position"
  
• "It depends on your tactics"
  
• "It's all about practise"
  

These all have some truth to them but don't actually help new players out. i.e. they all just lead to follow up questions:

• "Just look at key attributes and mentals"
  But are some more important than others? Which ones? Just take an average? Weight them? Establish a "baseline" for some of them - i.e. never sign someone with attribute y below z

  
  

• "It depends on what league you're in"
  How so? What constitutes a "good" number for attributes in the Premier/Championship/D1/D2? Should I be looking for players to have
  all
  their key and mental attributes at that number? If not how many and which ones? How low should I allow some to fall in favour of maximising other ones?

  
  

• "It depends on their position"
  Are there any constants across all positions? What are the variables?

  
  

• "It depends on your tactics"
  Are there any constants across all tactics? What are the variables? Is there simple way of conceptualising types of tactics down to five or six types that would allow players to be assessed for them in a structured way?

  
  

• "It's all about practise"
  I hope the glibness of this is self-evident
  

Cleon, llama3 and wwfan have done an amazing job at outlining the underlying principles behind designing and maintaining tactics and I think there must be a similar set of principles that new players could use for assessing their players/squads. I've no doubt that it would imperfect and not apply to every situation -as I'm sure Cleon, llama3 and wwfan would also say about their threads - but it would give new players a place to start.

jjtile - I am excited to see how you work this out. I think that some sort of attribute guide to assess a squad for a new player would be a great addition to the community here.

The averaging process is a good choice. I remember asking about attributes for lower level leagues and after asking a few times in different threads, someone ended up telling me that is is not so much about the VALUES of the attributes but HOW MANY are high enough for the position. This seems to tell me that having 2-3 excellent (say 16+) attributes could keep a players average high enough for League 1 and able to play a limited role, but make that player unsuitable for Championship and above. If those attributes were Jumping Reach, Heading, and Finishing on a Striker, which are very sexy attributes for a Stirker, a new player might attempt to play that player on a BPL club and never get the quality expected out of them. This could be because the Off the Ball, Pace, and Stamina were so low that defenders were able to easily mark the striker out of the match.

Disclaimer -- all examples are completely made up to prove a point and may be suitably absurd to make this process easier.

Also interested to see how jjtile will work this out. I think he has a point in post #10, and the topics he spoke about are usually neglected.

Some theory: when looking at players, I identify 4 main responsibilities: defending, passing, running with the ball, finishing. There are probably more basic elements of the game, but I have no formal knowledge about the game of football, so someone could help out with this.

Anyway, when looking at a player, I am trying to determine what he can do well. So, for example, I need one FB who can defend, and one who can pass and run with the ball. For each one, I look at the blocks of attributes that are needed for a given duty. If I need him to run with the ball, I look at acceleration, pace, balance, stamina, dribbling, technique. Now, a completely different thing is to rate which one of those attributes are relatively more important than the others.

Some theory: when looking at players, I identify 4 main responsibilities: defending, passing, running with the ball, finishing. There are probably more basic elements of the game, but I have no formal knowledge about the game of football, so someone could help out with this.

Anyway, when looking at a player, I am trying to determine what he can do well. So, for example, I need one FB who can defend, and one who can pass and run with the ball. For each one, I look at the blocks of attributes that are needed for a given duty. If I need him to run with the ball, I look at acceleration, pace, balance, stamina, dribbling, technique. Now, a completely different thing is to rate which one of those attributes are relatively more important than the others.

Thanks. This is really useful. I've been thinking about a system whereby you quickly assess groups of players in terms of attributes requires to play in certain groups of roles - and your idea around groups of skills is in a similar vein. Definitely food for thought.

I would like to share something I remembered today, which is semi-related to this topic.

Whenever I develop my CMs, I always make sure they can play either DM, or AM. There is one basic distinction I make between central midfielders - are they defensively oriented, or are they attacking oriented.

When it comes to defensive ones, I like to have guys with high mental stats, such as Aggression, Anticipation, Bravery, Positioning, Teamwork and Work Rate. Next priority are physical stats: Agility, Balance, Stamina and Strength. Tackling and Marking are also needed, but for some reason, I am more willing to go with guys that are mentally and physically capable, then with guys who are Technically proficient, but lack in either of those areas.

(Also, arguably, Concentration, Decisions and Determination are actually more important than any of the mentioned attributes)

So, once I have a midfielder that has a strong defensive game, which is by the way very important for me, I am looking at what extras he can bring to the table. Ideally, if he is a real good player, he will be a creative force. So: Passing, Technique, Creativity, Flair, possibly First Touch and Off the Ball, are the key attributes for this sort of a player, in my opinion. He is capable of playing Regista or DLP (def). Having this type of a player is a goal of any long term save I'm playing, because I believe that a player like this can really turn things around for a club, allowing it to play a much more defensively solid tactic, while still being able to contribute a lot to attacking phase.

Now, not every player has a potential to be that good, so if my DM is not suited to being a creator like this, I still train him to have a solid passing game. So, First Touch, Passing and Technique are the desired attributes for my regular DM/CM. I will usually play him as DM (def/sup), CM (def), BWM (sup), and maybe as a CM (sup), depending on his stats.

When it comes to attacking CMs, I always train them to be competent as AMCs, both in positional ability and attributes-wise. I make a simple distinction - is he able to create, or is he able to run with the ball and finish.

The first guy, would be a playmaker. I don't care about his defensive capability, but I want him to have these attributes: First Touch, Passing, Technique, Creativity, Flair, Off the Ball, Teamwork, Work Rate, Agility. He would play as an advanced playmaker for me or as a CM (sup), possibly as a CM (att). If he could Finish, it would be a big plus. Also, some Positioning wouldn't hurt, if he should play at MC, rather than AM position, even if I'm not expecting him to defend.

The second guy, let's call him a Runner, would need: Dribbling, Finishing, First Touch, Passing, Technique, Off the Ball, Acceleration, Agility. Some Creativity and Flair wouldn't hurt. This is the guy that I want to penetrate the defense, to support the attack and score when possible. Usual roles are: CM (att), AMC (sup/att), maybe a Shadow Striker.

Now granted, it would be the best to have a player that can do both - run with the ball and finish, while still be a creative force - and I do develop my top prospects to be that way. But the regular ones, I train to be either a Runner or a Creator.

Once in a while, I also find a player that can do it all, but usually doesn't excel in anything. Box to Box midfielder comes to mind. I haven't figured out how to train those players the best, but, for now, I think I like them to be good at Finishing, First Touch and Passing, Off the Ball, Positioning, Teamwork, Work Rate, Agility, Balance, Stamina and Strength. Finishing is an interesting choice, maybe. I find the likes of B2B mids do get into a goalscoring situations far more often than I imagined before.

So, there are some thoughts for you. Not perfect, I know. Perhaps some of the more experienced members could share their own thoughts.

When it comes to the wingers, the thing is simple. I train all of my left footed wingers to be able to play both ML and AML positions, and my right footed guys should be able to play MR and AMR. I might also train them to play AMC position, but I don't train them to be stikers. I like my strikers to be tall and strong, and it's not often that you find those qualities in a winger.

It is commonly said that fast wingers dominate the lower leagues. I find that it is also true for European football. If you can find a player that technically and mentally good, but not excellent, and is very fast (say 18+ for Acceleration and Pace), he will bang in goals for you like crazy.

So what I look for in a winger is primarily physical: Acceleration, Pace, Agility, Balance, in order of importance. Dribbling, Finishing, First Touch, Passing, Technique are the important technical attributes. Perhaps, Dribbling is the most important. Finishing might seem like the most important one, at least for me, because I rely on wingers to score. But I've had guys with only decent Finishing scoring 30+ goals for me, so I don't obsess about it. Key mental attributes would be Composure and Off the Ball. Work Rate is nice, but not so common in wingers. Creativity, Flair and Teamwork would be nice, if you like free flowing passing.

This guy would be equally suited to play as a Winger, or as an Inside Forward. Both will get into goalscoring chances. As I said, I like him to be able to play both in M and AM strata, but if I played him in midfield, I would still play him as a winger. If he is able to play as a WM (in other words - if he has some sort of a defensive game), that is fine, but I would never play him as a Defensive Winger; I would just use my Wing Back for that. Wide Target Man is something I didn't really play with yet, so no comments on that, but I assume that the type of a player I described wouldn't be suited to this role.