AI Squad Management - 10 year analysis

[Dave C]

One thing I always test is how well squad strengths hold up over time. That in mind, I've run a game through 10 season on holiday mode to compare the stats for Premier League, and Blue Square Premier, clubs.

I've compared the figures from when the game starts, to those at the 1st September 2021. Both "Add Staff" and "Add Players" were unticked.

I've not made any allowance for loanees, so there'll be some duplication as players are counted twice in some cases. But the methodology is consistent over both games, so shouldn't skew the figures.

Squad Size is all full-time players for PL clubs, and all FT/PT/NC players for BSP clubs.
Youth Team stats are based on all players on a YTH contract.
Average number of keepers is ALL keepers at the club, of any status.
Average number of coaches is based on manager, AM, coaches, including youth coaches.


Premier League 2011/12

Average Squad Size = 45.4
Total players:
20-29 = 0
30-34 = 1
35-39 = 5
40-44 = 3
45-49 = 6
50+ = 5
Total FT players in PL = 908

Average Youth Squad = 22
Total players
10-19 = 4
20-29 = 15
30-34 = 1

Average number of goalkeepers = 7.2
2 x 4
4 x 5
1 x 6
2 x 7
4 x 8
6 x 9
1 x 10

Total number keepers in PL = 144

Average number of coaches = 11.95
Range = 5-18

Premier League 2021/22

Average Squad Size = 36.7
Total players:
20-29 = 2
30-34 = 3
35-39 = 9
40-44 = 6
45-49 = 0
50+ = 0
Total FT players in PL = 734

Average Youth Squad = 23.95
Total players
10-19 = 0
20-29 = 20
30-34 = 0

Average number of goalkeepers = 5.8
4 x 4
2 x 5
9 x 6
4 x 7
1 x 8
0 x 9
0 x 10

Total number keepers in PL = 116

Average number of coaches = 10.8
Range = 7-14

Blue Square Bet Premier League 2011/12

Number of professional sides = 15

Average Squad Size = 23.75
Total players:
<15 = 0
16-19 = 2
20-24 = 14
25-29 = 6
30+ = 2
Total players in BSP = 570

Average Youth Squad = 9.58
Total players
0 = 7
<15 = 10
16-19 = 4
20-24 = 2
25-29 = 1
30-34 = 1

Blue Square Bet Premier League 2021/22

Number of professional sides = 18

Average Squad Size = 23.29
Total players:
<15 = 2
16-19 = 3
20-24 = 10
25-29 = 6
30+ = 3
Total players in BSP = 559

Average Youth Squad = 20.7
Total players:
0 = 1
<15 = 1
16-19 = 1
20-24 = 12
25-29 = 9

[Dave C]

Some thoughts on that data.

The most obvious issue is the dwindling of PL squads, down by nearly 10 players on average. There's a normalising to around 35-44 players, and none of the 45+ super-squads" we see IRL. There's also a five-fold increase in the number of clubs running smaller, <35 player, squads. So squad stability is not holding up at all. This is reflected in the loss of keepers from the game.

PL youth squads hold up fairly well, although there is a lack of smaller/larger squads, everyone plotting in the same area.

Staff drops a little on average, and this is mainly due to teams reducing numbers of youth coaches.


In the BSP, there's a slight increase in the number of pro sides, which I'm not to worried about (I'd be more concerned if that was going the other way).

Outwardly, average senior squad size holds up well. But there are concerns. There are 5 full-time sides there that have squads of 18 or fewer players, including two that don't even have the 16 players they'd need for a match of 11 + 5 subs. That needs looking at.

Average youth squad jumps hugely, mainly because almost every team now has a functional youth team (17 teams in 11/12 have <15 youth players, down to 2 in 21/22). So there's maybe an issue there.

[nigelquelch]

EDIT: You answered me at the same time as I posted

[pikeal]

Good. But I think what would be better to see is the squad spread over all the positions (particularly in the positions of the formation currently used).

[Dave C]

Left the staff stats out, hang on...

[Dave C]

EDIT: You answered me at the same time as I posted

To see if things are stable over time. It's important, as if the AI isn't building strong squads, then the game gets easier.

[Dave C]

Good. But I think what would be better to see is the squad spread over all the positions (particularly in the positions of the formation currently used).

I really can't be bothered doing that, it'd take ages! Plus, the formation used changes over time and is subject to numerous variables.

[Sheffield Sunday]

I'm afraid you cannot prove anything just by running one ten year soak test.

To get a feel for things statistically I'd recommend doing at least 20 x 10 year saves and giving us the statistical range in each of the categories given. That way we can 'normalise' some of the in-game events that may effect single saves. - i.e. administration/windfall events and the effect on squad building.

[Barside]

There is one piece of information that would be interesting, average age of senior squad.

[Sheffield Sunday]

Forgot to add time consuming and wouldn't blame you for not signing up to it! Well done, starts a debate at least!

[MSCCG]

Numbers look like they are not maintained, but without statistical significance I cannot make a conclusion. Squad size is noticeably lower in the PL though, I wonder what the reason is? Most likely it is that the characteristics of the AI are encoded differently to how clubs behave in real life, but due to what stimuli? Is it the debt/lack of transfer funds, or the tendency for the AI to buy less players for more money and concentrate on quality over quantity?

[Riz Remes]

Average youth squad jumps hugely, mainly because almost every team now has a functional youth team (17 teams in 11/12 have <15 youth players, down to 2 in 21/22). So there's maybe an issue there.

The jump in BSP youth squad size is likely due to the fact that at the start of the game quite a few BSP teams don't have fully modelled youth teams in the DB, even if they might have enough youth players in real life. And since the game was started with the "Add Players" unticked, at the game start only the actual players from the DB would be present causing the small sizes of the youth teams.

[Dave C]

I'm afraid you cannot prove anything just by running one ten year soak test.

To get a feel for things statistically I'd recommend doing at least 20 x 10 year saves and giving us the statistical range in each of the categories given. That way we can 'normalise' some of the in-game events that may effect single saves. - i.e. administration/windfall events and the effect on squad building.

Over the years I've done this multiple times over. I don't need to do it again, because it's *always* an issue. this was just to test whether anything has changed. it hasn't.

[Dave C]

There is one piece of information that would be interesting, average age of senior squad.

I can probably pull that one out pretty easily, when I have time.

[Dave C]

The jump in BSP youth squad size is likely due to the fact that at the start of the game quite a few BSP teams don't have fully modelled youth teams in the DB, even if they might have enough youth players in real life. And since the game was started with the "Add Players" unticked, at the game start only the actual players from the DB would be present causing the small sizes of the youth teams.

Yeah, I'm not unduly concerned by that one, it's too badly affected by initial data.

[PhatsUK]

To see if things are stable over time. It's important, as if the AI isn't building strong squads, then the game gets easier.

I would have expected the size of premier league squads to continue to drop with the recent introduction of the 26 man squads, mid 30s would seem a reasonable size to include a few up and coming u21's and some dead wood. What I would be more interested in would be whether the AI Managers are able to build balanced and competitive squads this time around.

A) Do they buy players to fit their desired formations and gaps in their squad balance or do they still buy players they clearly don't need frequently
B) Is the AI able to identify and use a non preferred formation where possible that would suit their current players better and employ it where appropriate
C) What quality of players are the human and AI managers able to keep happy sitting on the bench constantly with very little game time and how long before these players want to play first team football
D) Is the AI able to identify potential talents within its squad and give them game time appropriately to help them fulfill their potential where possible

There were some good suggestions about ways this could be done over the summer on these forums and I had hoped some information would have come from the developers in blog series about these issues as they would make those long term saves more fun

[Sheffield Sunday]

Dave C - Fair enough! Didn't realise that this was an ongoing project. My apologies if I came across short - I see a fair number of *proofs* in the fourm that aren't based on any kind of statistical analysis and it annoys me greatly, so thought I'd shout up on this occassion.

Do you have data for the other divisions rather than just top and bottom? What I suspect might be happening is that Championship squads are getting larger due to relegation from the top, and teams taking a gamble on 'making the big time'. Do you have anything to back up my theory?

[Dave C]

I would have expected the size of premier league squads to continue to drop with the recent introduction of the 26 man squads, mid 30s would seem a reasonable size to include a few up and coming u21's and some dead wood.

But FM doesn't sim trends, and it's not trying to predict changes in the nature of football. it should remain stable for how things are now.

What I would be more interested in would be whether the AI Managers are able to build balanced and competitive squads this time around.

A) Do they buy players to fit their desired formations and gaps in their squad balance or do they still buy players they clearly don't need frequently
B) Is the AI able to identify and use a non preferred formation where possible that would suit their current players better and employ it where appropriate
C) What quality of players are the human and AI managers able to keep happy sitting on the bench constantly with very little game time and how long before these players want to play first team football
D) Is the AI able to identify potential talents within its squad and give them game time appropriately to help them fulfill their potential where possible

There were some good suggestions about ways this could be done over the summer on these forums and I had hoped some information would have come from the developers in blog series about these issues as they would make those long term saves more fun

You'll have to do all that yourself!

[Tommy_Styles]

Did not read the comments but it seems to make sense to me.

Players are getting more expensive -> Teams need to trim down on the amount of players -> Teams start taking in more youth to develop their own, thus cheaper, players.

Football evolves, squads evolve. Nothing stays the same, sorry.

[PhatsUK]

But FM doesn't sim trends, and it's not trying to predict changes in the nature of football. it should remain stable for how things are now.

You'll have to do all that yourself!

That is fair enough, although I have a job already so instead continue my fm11 save and I'll be waiting for a body of anecdotal evidence to start to appear in the forums once people get that far in the game. As far as the overall number of players I thought that the whole idea was to reduce squad size and stop the biggest/richest teams stockpiling the most talented players IRL.

I still think that mid 30s is reasonable and could be a leveling off of squad size to meet the recent changes in competitions squad rules, however if the soak test was done over a longer time and continued to drop to low 20s then there would be a trend that would be more worrying.

Out of interest how many active leagues are you running for the soak? (A lack of players deemed to be of sufficient caliber could also cause problems in theory)

If the data can be collated easily enough I'd also be interested how many players in each squad were 'non-registered'/ineligible to play in the league.

Has there been any significant change to the reputation of the leagues with the DLR in place? How much is the average salary by club, how much is the total salary spend etc etc

[Dave C]

Did not read the comments but it seems to make sense to me.

Players are getting more expensive -> Teams need to trim down on the amount of players -> Teams start taking in more youth to develop their own, thus cheaper, players.

Football evolves, squads evolve. Nothing stays the same, sorry.

Again, FM isn't attempting to predict and model future changes. It reflects the game world as it is now, therefore it should stay stable. It doesn't, therefore there is a problem.

[Erimus1876]

I would have expected the size of premier league squads to continue to drop with the recent introduction of the 26 man squads, mid 30s would seem a reasonable size to include a few up and coming u21's and some dead wood. What I would be more interested in would be whether the AI Managers are able to build balanced and competitive squads this time around.

A) Do they buy players to fit their desired formations and gaps in their squad balance or do they still buy players they clearly don't need frequently
B) Is the AI able to identify and use a non preferred formation where possible that would suit their current players better and employ it where appropriate
C) What quality of players are the human and AI managers able to keep happy sitting on the bench constantly with very little game time and how long before these players want to play first team football
D) Is the AI able to identify potential talents within its squad and give them game time appropriately to help them fulfill their potential where possible

There were some good suggestions about ways this could be done over the summer on these forums and I had hoped some information would have come from the developers in blog series about these issues as they would make those long term saves more fun

This is the crucial part that needs overhauling. The number of players in a squad after 10, 20 years etc is important to know, but its the quality of those players, and therefore the AI squads, and subsequent game challenge, that is the real issue. Some of us tried testing this a few months ago on FM11 as its pretty well known now that AI squad building has been very poor for the last few games in the series. I doubt FM12 will be much different since nothing was ever mentioned in regards to wether or not this crucial aspect of the game was improved, or overhauled.

Now that the add league feature is here longer saves may become the norm now as people keep playing the same save instead of restarting a new one. So I think as the months go by there'll be a lot of posts on here by people complaining just how easy the game gets as the seasons roll by due to the declining and stagnating AI gameword.

[Dave C]

That is fair enough, although I have a job already

So do I, hence I can only do something quite simple like this.

As far as the overall number of players I thought that the whole idea was to reduce squad size and stop the biggest/richest teams stockpiling the most talented players IRL.

I still think that mid 30s is reasonable and could be a leveling off of squad size to meet the recent changes in competitions squad rules, however if the soak test was done over a longer time and continued to drop to low 20s then there would be a trend that would be more worrying.

Sorry, what "whole idea"? From who? Do you mean FIFA and the squad regs? True enough, but personally I think the big clubs will continue to find ways round it.
Mid-30s is reasonable, and it may well be where things so. I don't think that's what will happen, and I think the Man Utds of this world will continue to stockpile players, then loan them out over a period of years. But that's a prediction, and that's not what FM does.

Out of interest how many active leagues are you running for the soak? (A lack of players deemed to be of sufficient caliber could also cause problems in theory)

Only the English leagues to BSN/S

If the data can be collated easily enough I'd also be interested how many players in each squad were 'non-registered'/ineligible to play in the league.

Has there been any significant change to the reputation of the leagues with the DLR in place?

No idea on either of those, I might have a look later.

[PhatsUK]

-snip-
you've covered what i was going to reply to while i was typing

[Dave C]

Right but have all the clubs managed to get rid of their dead wood and get their squad sizes down to useful levels, why would they keep players they couldn't register and/or use, it makes no financial sense in the rules that have been set for the AI in FM12 game.

As in real life, they'd be loaning those players out to other clubs. The bigger PL clubs IRL seem to manage those sizes of squad just fine, with a combination of loans and reserve football. Plus, all the younger player remain eligible whether selected for 25 man squads or not.

[Dave C]

Do you have data for the other divisions rather than just top and bottom? What I suspect might be happening is that Championship squads are getting larger due to relegation from the top, and teams taking a gamble on 'making the big time'. Do you have anything to back up my theory?

When I've done this in the past, and I see no reason to imagine it's any different now, you actually saw the total number of players over the whole database drop. Players retired early (and there's anecdotal evidence around the forums that this remains an issue), and you saw all squad sizes drop down.

[Dave C]

I still think that mid 30s is reasonable and could be a leveling off of squad size to meet the recent changes in competitions squad rules, however if the soak test was done over a longer time and continued to drop to low 20s then there would be a trend that would be more worrying.

I'd also have to point out, this has been an issue in the game for at least 5 years now. So it can't just be an effect of the squad rules. That's just a convenient excuse for a long-term issue with the game.

[Dave C]

This is the crucial part that needs overhauling. The number of players in a squad after 10, 20 years etc is important to know, but its the quality of those players, and therefore the AI squads, and subsequent game challenge, that is the real issue. Some of us tried testing this a few months ago on FM11 as its pretty well known now that AI squad building has been very poor for the last few games in the series. I doubt FM12 will be much different since nothing was ever mentioned in regards to wether or not this crucial aspect of the game was improved, or overhauled.

This is purely anecdotal, as it's the impression I built up while compiling this data, but it looked like the squads were down to maybe 18-20 good players, often aging players, with the rest being younger players. So some questions there over squad depth. But that's not hard evidence at all, as I say.

[PhatsUK]

DaveC, please don't think I'm disagreeing with you because i'm not in theory. I agree that squad building on numerical, quality, age and positional distribution has been a problem in the past and most likely is again. I am not convinced however that a drop to an average squad size of 35 employing the rules that the AI is set out to follow (ie register 26, homegrown players, salary caps) shows a significant trend beyond the realms of plausibility yet.

That is why I asked for example about the number of ineligible players since the actual available squad size needs to deduct that, buying players they don't even register has been a problem in the past. (Taking account of your theory that bigger clubs would use the loan system to hold bigger squads, players over 21 who were loaned out would not be counted as ineligible as the AI has come up with a solution for their continued development).

This is purely anecdotal, as it's the impression I built up while compiling this data, but it looked like the squads were down to maybe 18-20 good players, often aging players, with the rest being younger players. So some questions there over squad depth. But that's not hard evidence at all, as I say.

From an anecdotal point of view, in this instance do you feel that there are players available in the game world that would improve the squads of these clubs and that would be realistic transfer targets given their values and the status of the clubs in question? Have the clubs 'wasted' their salary budget on players who are aging and no longer offer good value for money? Have they failed to develop their potential talents? No hard facts necessary, just your overall feeling of where it is going wrong is fine by me.

[Govnar1]

Top work Dave C!

In my experience (FM 2011) the main problem of AI squad building is the lack of connection between the player search criteria and preferred (chosen) formation/playing stile. Furthermore it seems that the AI does not take into account specific qualities/characteristics of the existing squad, as well as need for any new acquisitions to complement those existing qualities/characteristics.

[phnompenhandy]

Dave, I welcomed you back on your other thread and I'm loving your analyses. I'm on your side regarding the big picture, but I'd just like to comment on some issues here.

Firstly, whilst I accept and agree that SI doesn't go in for predicting the future, i do feel comfortable with patterns that can be rationalised by real-world projections. This the smaller PL squads for me can reflect the UEFA financial rules that will cause clubs to reduce their wage budgets. I like the idea that lower league clubs with have healthier youth set-ups; it implies that in the FM world there will be a strong reaction to the ugly decision this week to allow PL clubs to poach kids on the cheap.

However, in purely FM terms, I'd be very interested in the point Rez raised, which brings me to my second point. As a Lower League manager myself, I always tick the add players/add staff boxes since I don't want to be facing teams of greys. It looks like that makes a significant difference to your results, so I wonder if you'd be interested in running the tests again with that change (and maybe you can explain why you decided not to tick those boxes).

[Dave C]

However, in purely FM terms, I'd be very interested in the point Rez raised, which brings me to my second point. As a Lower League manager myself, I always tick the add players/add staff boxes since I don't want to be facing teams of greys. It looks like that makes a significant difference to your results, so I wonder if you'd be interested in running the tests again with that change (and maybe you can explain why you decided not to tick those boxes).

Got fed up which is why I didn't return to this! Didn't tick the boxes as I wanted to compare real data to FM, not have anything being boosted by the game. Fully agree that it causes some problems, especially in that YT area. I might re-run this week if I can be bothered.


Personally, for me, the most concerning aspect is that you still get pro sides in FM with an inadequate number of players down the leagues. That just shouldn't be happening. I would like to see PL sides holding slightly bigger squads still, say 40. That would account for a decent number of younger players on pro deals being loaned out for development.

[Dave C]

The reliance on loans at lower levels in FM is another thing that should be sorted. I've noted clubs that obviously identify a weakness in their squad (let's say they lack a strong centre-half). They fix this with a loan, which is fine as a temporary measure. But then they keep contracting the players that have been deemed inadequate, and just loaning in the first-teamer instead. I'd really like to see the AI looking preferentially to getting the permanent squad right, with loans used more for emergencies, or as a reluse to a signing. Y'know, like in real life...

[AytchMan]

I applaud your work, Dave, it's very interesting stuff. A couple of posts touched on the squad quality issue which is where I'd like to see more analysis. Specifically, if there was any easy way to generate some sort of overall squad quality index. I doubt that it would be an easy job (nor terribly accurate) but simply totaling up squad attributes in the aggregate and comparing averages over the ten years might be instructive. Could any of the editors somehow be used for this?

[Dave C]

This, incidentally, is even worse on the latest build. There are Premiership teams with less than 30 players between their Senior and Reserve sides, some have as few as 18-20, meaning they don't even fill a squad!

[Barside]

This, incidentally, is even worse on the latest build. There are Premiership teams with less than 30 players between their Senior and Reserve sides, some have as few as 18-20, meaning they don't even fill a squad!

This is something I have also noticed, a few seasons back on my save Newcastle won the Premier League with 16 senior outfield players & Manchester City just missed out on a top 4 place with a senior pool of 13 outfield players.

For some reason it only appears to be a problem in England & more specifically in the top flight.

[Jops14]

I think largely an issue that arises in every single league is the death of the reserve team. after 10 years (i frequently played 15+ years in my saves) almost every team has no reserve players, meaning a drop in squad size which is primarily down to youth teams not producing a sheet list of good enough players. You may end up with 1 or 2 players destined for the premiership (which is about right) but generally the rest are utter crap and dont get professional contracts, which in turn leads to smaller squads across the board.

Of course the other issue which has been noted is the AI buying players it doesnt need as well as holding on to ageing players too long, or alternatively (something that really annoys me) releasing them when they are 32 and these stars (often good enough for mid table teams) get no offers dont drop a division and thus retire. It frustrates me to see Ibrahimovic, robinho, robben etc retire at 32 when they would still be a star for most teams.

[goldtiger0]

As someone previously stated (and I apologize for not naming him) the problem lies with the mostly horrendous quality of youth players being produced. The problem is that in a 'class' of say, 15 players, only 3 will go on to persue professional careers. In real life that number would be fine due the the astonishing number of teams in football around the world, but in FM there aren't enough teams to compensate for those players. As a result, I usually just use FMRTE to mass edit all regens with a potential of 100 or below and raise it by 50. That, along with raising their CA by 20, allows most players to play professionally and make up for the lack of players in long term saves.

[hursty2]

It seems as if my save has produced one of the most realistic gameworlds, squad building has been great and the premier league has been competitive throughout.

I'm in 2025, so if anybody wants any stats or are just generally interested in what has happened over the years, fire away!

[bleventozturk]

It seems as if my save has produced one of the most realistic gameworlds, squad building has been great and the premier league has been competitive throughout.

I'm in 2025, so if anybody wants any stats or are just generally interested in what has happened over the years, fire away!

I'm glad to hear this. I just started a career game in FM2012, and reading threads like this (probably exaggerating some issues in the game) sometimes makes me stop playing.

So, I'm curious, in your game, in PL, the team with the least amount of first team players has how many players?

Maybe the DB selection at the start of the game is effective here. I wonder if DaveC selected small DB, and only PL active maybe, whereas you selected large DB and many leagues active?

[bleventozturk]

Personally I think it is ok that teams with 45+ players in the roster no longer exist with few seasons down the road in the game. The game has 1 model for squad building which applies to all the PL teams, so it is natural that after a few seasons all the teams will have similar squad sizes. And if the squad size is 30+ I think it is good enough. So, instead of comparing the stats at the start to the stats in 2021, I would compare the stats in 2021 to stats in 2031 for example (granted that the squad sizes are reasonable after converging to an average number).

The game does not model the future, it does not predict the future, but the squad size for every team eventually converges to a number that is set by how the squad building is modeled in the game code. It does not differentiate between M.City and Swansea, hence both of these teams are likely to end up having similar squad sizes. The difference between them will be the players' quality, not quantity.

[hursty2]

I'm glad to hear this. I just started a career game in FM2012, and reading threads like this (probably exaggerating some issues in the game) sometimes makes me stop playing.

So, I'm curious, in your game, in PL, the team with the least amount of first team players has how many players?

Maybe the DB selection at the start of the game is effective here. I wonder if DaveC selected small DB, and only PL active maybe, whereas you selected large DB and many leagues active?

I started my game with a medium database with England (League 2), Spain (Adelante), Germany (Bundesliga 2.), Italy (Serie B), France (Ligue 2) and have since added Dutch (Eredivisie), Denmark (3rd Tier), Switzerland (2nd Tier) but have since removed England (League 1 and 2), Denmark (3rd Tier).

The club with the smallest SENIOR SQUAD size in the premier league is; Chelsea and Middlesbrough (21 players).

The past winners of the premier league can be seen here. It is worth noting that from 2011 to 2020 I managed Stoke City taking them up to 4th in 2019 before getting sacked a year later after a 10th place finish. I then took over Atalanta in Serie A for 2 years, then HB Koge in a newly added Danish League before returning to the Premier League Just this year to manage recently promoted Bristol City.

[bleventozturk]

This means you have pretty much the same problem actually. 21 is too small imo.

[hursty2]

Hardly, how many players does Man Utd have currently that you would class as senior squad players? I use this example because I know the team off by heart.

De Gea, Amos, Lindegaard, Jones, Rafael, Smalling, Evans, Rio, Vidic, Evra, Fabio, Scholes, Carrick, Park, Cleverley, Anderson, Nani, Young, Valencia, Rooney, Welbeck, Hernandez, Berbatov, Owen.

That's 24.

Most squads have around this amount.

[Bababui]

I'm glad to hear this. I just started a career game in FM2012, and reading threads like this (probably exaggerating some issues in the game) sometimes makes me stop playing.

Why would you claim this? The OP has provided statistical breakdowns to backup his thesis.

[Bababui]

I started my game with a medium database with England (League 2), Spain (Adelante), Germany (Bundesliga 2.), Italy (Serie B), France (Ligue 2) and have since added Dutch (Eredivisie), Denmark (3rd Tier), Switzerland (2nd Tier) but have since removed England (League 1 and 2), Denmark (3rd Tier).

The club with the smallest SENIOR SQUAD size in the premier league is; Chelsea and Middlesbrough (21 players).

The past winners of the premier league can be seen here. It is worth noting that from 2011 to 2020 I managed Stoke City taking them up to 4th in 2019 before getting sacked a year later after a 10th place finish. I then took over Atalanta in Serie A for 2 years, then HB Koge in a newly added Danish League before returning to the Premier League Just this year to manage recently promoted Bristol City.

Mr Hursty is misleading us. He added leagues thus setting off the exploding number of players bug. His game will have little in common with a player using a set number of leagues.

[hursty2]

Mr Hursty is misleading us. He added leagues thus setting off the exploding number of players bug. His game will have little in common with a player using a set number of leagues.

I think you will find that in my game player numbers have not increased dramatically, in fact genie scout currently says I have 129K not exactly a huge jump compared to some other games. It might be because I started this game as soon as it came out and that I haven't added any Brazilian leagues which seem to cause the most problems. Although at the last time of checking there were only 82K worth of players but since that time I haven't added any leagues, just removed a couple.

Also I don't think that this would matter either in the context of squad sizes and all, the players that are filling up the DB are very low CA & PA players.

Now you have mentioned it though, I hope my game doesn't break.

[achilles-the-victorious]

My Fm11 save is in 2036, so I decided to check some of the opposition's team in this regard. Usually when I scout the opposition I want to know what formation they are likely to slot into, and who the key players are. But I don't particularly check for squad size, because they only get to put 11 men on the pitch, so my focus is always on the first XI.

Well, after being a little disappointed at first with some lowish numbers I don't consider it a problem. Why? Because even though Man City had an old average squad, and only about 23 players in the first team, there were about 6 or 7 promising 20/21 y.o.s loaned out by the Reserves, which means that even though their first team looks aging, and no sign of quality replacements, their players of the future are actually off developing at another club!

Meanwhile, I checked Lazio, who have been one of the top Italian teams since a tycoon took them over. I was a little disappointed at what seemed to be a lack of squad depth, not enough covers for positions. But then I looked closer. The 'manager' of Lazio had actually been quite clever, not only buying high quality players, but players who could play in multiple positions, as well! LM that plays LB, LB that plays LM, DMC that plays AMC and vice versa. I don't know how well those players play out of position, but it strikes me as quite clever, what they have done.

Most importantly, though, teams still provide a good challenge, and there are some very formidable players that you will come up against.

[Dave C]

So, I'm curious, in your game, in PL, the team with the least amount of first team players has how many players?

Maybe the DB selection at the start of the game is effective here. I wonder if DaveC selected small DB, and only PL active maybe, whereas you selected large DB and many leagues active?

Nope, Large DB, all players retained from active countries, all players from major nations, and about 15 leagues across 6 countries. The problem is the AI, plain and simple. It needs fixing, and it's being ignored year after year.


Incidentally, since I've provided the actual numbers, can you tell me how I possible exaggerated anything? I'm not some teenage ranter complaining that my super team doesn't win. I've been a researcher and tester for this game, for many years. I know how to look at these trends and identify problems. I quit testing precisely because I was frustrated at stuff like this being ignored.

[Dave C]

Personally I think it is ok that teams with 45+ players in the roster no longer exist with few seasons down the road in the game. The game has 1 model for squad building which applies to all the PL teams, so it is natural that after a few seasons all the teams will have similar squad sizes. And if the squad size is 30+ I think it is good enough. So, instead of comparing the stats at the start to the stats in 2021, I would compare the stats in 2021 to stats in 2031 for example (granted that the squad sizes are reasonable after converging to an average number).

The game does not model the future, it does not predict the future, but the squad size for every team eventually converges to a number that is set by how the squad building is modeled in the game code. It does not differentiate between M.City and Swansea, hence both of these teams are likely to end up having similar squad sizes. The difference between them will be the players' quality, not quantity.

30+ is fine, but we're talking about sides with barely 20 in some cases, no depth, no cover, no players being loaned out, tiny squads. That's the PL.
Go down the leagues, some don't even have enough for a starting XI and subs. How anyone can possibly defend this is beyond me.

[Barside]

Go down the leagues, some don't even have enough for a starting XI and subs. How anyone can possibly defend this is beyond me.

^^This^^

I stumbled across a Reading team who had 5 players for a couple of seasons, they had failed to offloaded their higher earners after a couple of successive relegations which meant they could not afford to offer contracts to the bulk of their squad.

Currently L2 York have 13 senior players (2 are keepers), L1 Bournemouth have 16 (3 GK's), PL Man City & Newcastle have 19 (3 GK's)

[bleventozturk]

Nope, Large DB, all players retained from active countries, all players from major nations, and about 15 leagues across 6 countries. The problem is the AI, plain and simple. It needs fixing, and it's being ignored year after year.


Incidentally, since I've provided the actual numbers, can you tell me how I possible exaggerated anything? I'm not some teenage ranter complaining that my super team doesn't win. I've been a researcher and tester for this game, for many years. I know how to look at these trends and identify problems. I quit testing precisely because I was frustrated at stuff like this being ignored.

I understand your frustration and I'm with you there. I quit beta testing every once in a while when I get fed-up with some improtant issues being constantly ignored. So, I assure you I have no intention to defend SI here, and I think I have been one of the biggest criticisers in these forums the last few years.

My comment on exaggeration just came out wrong. It wasn't directed towards this post actually. It sounded like that though, simply because I realized i closed the paranthesis 1 word too early. My mistake. So, although this thread is definitely not one of them, sometimes I see threads where people do exaggerate the shortcomings of the game.

Going back to the subject now: Yes, I agree with you guys. There should be at least a minimum squad size enforced in the game, as a temporary workaround until SI fixes this properly. so that AI managers will fill their squad with some lesser players from the free transfer list. It is ridiculous for a professional team to have less than ~25 senior players in the roster.

For the time being, I just want to know if this happens with many teams, or are those ones with the extremely small squad are seldom? I can live with it and continue my career game that I started with enthutiasm if 1 out of 10 team has these issues. I never use FMRTE, exploiting super tactics, player search, etc. so the game provides eough challange for me and therefore it is still entertaining.

I wish AI squad building will be improved significantly for FM2013, but I'm pretty sure SI is busy trying to come up with some bells and whistles to implement into the new version, as we speak, rather then first improving the current features. Some of those features won't be functioning on the game release day, and patches will follow in an attempt to fix those. In the meantime transfer AI and squad building AI will be ignored because 'it is not a priority'. This has been the story the past 5? years, and SI still has to prove me wrong about my expectations for the new versions.

[Jops14]

You can only assume it hasnt been fixed in these few years as a solution has not been thought of, i doubt they are ignoring the issue.

[bleventozturk]

You can only assume it hasnt been fixed in these few years as a solution has not been thought of, i doubt they are ignoring the issue.

I disagree. They are a small company and therefore they have limited manpower, as they always claim. With the limited resources, it becomes a question of what they will see as priority. They are a business, and their first priority every year is to get the new version out at more or less the same time of the year. To be able to sell the game to majority, they feel like the game has to offer new features, changes from the last version. So, coming up with those new features and implementing them into the game is their priority. The transfer system in the game is a working feature. It is not working perfectly, but it is working, enough to kepp the game enjoyable, therefore enough to sell the game. They will start using their resources on improving the already existing features only after the addition of the new features. Until then it doesn't become a priority.

People may not realize but this is a very advanced game, which shows me how capable the coders in SI are. The only reason the transfer AI in the game not improving drastically, is that it needs a major overhaul (like throwing away the reputation system), and SI simply does not feel ready to use their limited resources on this yet. I am pretty sure they will fix it at some point, but they will do it when they decide to do it.

[Dr. Hook]

^^This^^

I stumbled across a Reading team who had 5 players for a couple of seasons, they had failed to offloaded their higher earners after a couple of successive relegations which meant they could not afford to offer contracts to the bulk of their squad.

Currently L2 York have 13 senior players (2 are keepers), L1 Bournemouth have 16 (3 GK's), PL Man City & Newcastle have 19 (3 GK's)

This I have noticed as well- teams that have suffered relegations, but retain players on ridiculous wages. Results in 5-8 first teamers and the rest youth, and it's a vicious trap that sends them spiralling down even further. The fix seems to to be some code that will direct the AI, if they fail to return right back to the league they relegated from, to push off the high earners, or at least not resign them when the contracts are up.

[Validicus]

It's been a while since I played a save for longer than 4-5 seasons. It's difficult to retain the motivation to pursue any kind of long-term save with this ongoing issue. Out of curiosity, Dave C, does your annual research find the AI Squad Management issue to be gradually getting worse somehow? I could have sworn that AI squads seemed to fare significantly better as the seasons went on in FM 05/06/07 compared with the last few versions.

[bleventozturk]

It's been a while since I played a save for longer than 4-5 seasons. It's difficult to retain the motivation to pursue any kind of long-term save with this ongoing issue. Out of curiosity, Dave C, does your annual research find the AI Squad Management issue to be gradually getting worse somehow? I could have sworn that AI squads seemed to fare significantly better as the seasons went on in FM 05/06/07 compared with the last few versions.

I could have sworn it was better in CM 01/02 even. Back then the input that was used by transfer AI was a lot more simple though. More tools got added into the game, more difficuly it became to manage the AI transfers. There are more parameters now I guess. Hence SI needs to really focus on this finally in a new version, rather than adding stuff like agents to make the matter more complicated.

[Äktsjon Männ]

Yeah, layers of new code added on to outdated base mechanics have made it crumble. Both the transfer market and general AI squad building were more basic in the older versions but at least they worked to a degree. Now the same AI has to deal with agents and negotiate contracts that can involve several financially devastating clauses. It can't really cope.

[Barside]

Yeah, layers of new code added on to outdated base mechanics have made it crumble. Both the transfer market and general AI squad building were more basic in the older versions but at least they worked to a degree. Now the same AI has to deal with agents and negotiate contracts that can involve several financially devastating clauses. It can't really cope.

This is something that I also agree with, for me the AI has been left behind by the sheer complexity of the game & now it has reached the point where its decisions are based purely on the reputation system which is also an outdated & increasingly flawed system.

My biggest fear is that if the new match engine gets to see the light of day it will only advance the problem further as it will no doubt be a more complex beast so that we can have a much more realistic match experience but on the flip side the AI will not be capable of using all the extra aspects presented which in turns makes the game less challenging.

The questions faced by SI when decided where best to focus their resources is what type of players makes up the largest proportion of the consumer base?

If most players only do 3-5 season saves before starting over then spending additional time on reworking the AI to provide a tougher long-term challenge might not be a worthwhile expense.
Is the above true because people get bored by the easiness as squads fail to improve & go back to the beginning to rediscover the challenge of competing against current squad?
How much do the long-term/career (25+ season) players drive this game forward & add to/maintain its reputation?

My hope is that with SI hiring an internal marketing & PR team they are being given much more autonomy by Sega which could ease the commercial pressures placed on them, however the fear is that all the extra income from the anti-piracy measures has been allocated to the marketing & PR budget, with little additional investment in the development process.

[Validicus]

Thing is, Dynamic League League reputation was obviously introduced with long-term saves in mind. However, for anyone who plays a save long enough for DLR to come into effect, they happen to reach a point in the game where they need only take a glance at a few AI squads to see a blatant issue which far eclipses the shortcomings of the old static reputation. There's much that can be speculated from this. SI, to their credit, surely do want to cater for long-term save players, or else something as trivial and superficial (in comparison) as DLR would never have been looked at in the first place, especially considering how incredibly complex it must have been to implement. So, is it plausible that AI squad management is simply an insurmountable problem? I find it hard to believe that it would be low on the priority scale considering how glaringly obvious an issue it is, as well as how much of a bearing it has on medium to long-term saves. Nevertheless, the situation would almost seem like a sports car being increasingly unstable at high speeds, yet the manufacturers still strive to improve the car's acceleration, all the while introducing a series of flashy new additions to the car's interior which will hopefully be suitably distracting.

[bleventozturk]

The situation would almost seem like a sports car being increasingly unstable at high speeds, yet the manufacturers still strive to improve the car's acceleration, all the while introducing a series of flashy new additions to the car's interior which will hopefully be suitably distracting.

Excellent analogy mate

This is exactly what SI has been doing the past few years, not sure if they are aware or not, by adding so many new features and not improving the game's essential features nearly enough.

[Barside]

I forgot about it as it seems to have very little influence in my game but he point about DLR is a good one, the car analogy isn't.

[Dr. Hook]

The thing is, it isn't a uniform disaster. Some AI squads do quite well, and most seem at least okay. We tend to remember the egregious problems because they are so noticeable. To say SI doesn't build for the long-term saves is a bit unfair, I think. They test the games to many years out and the intent is to keep the game world stable over the long haul. It just doesn't work out in all situations. The squad-building AI needs work, simple as, but I don't think it has a thing to do with SI not caring about long term gamers.

[bleventozturk]

The thing is, it isn't a uniform disaster. Some AI squads do quite well, and most seem at least okay. We tend to remember the egregious problems because they are so noticeable. To say SI doesn't build for the long-term saves is a bit unfair, I think. They test the games to many years out and the intent is to keep the game world stable over the long haul. It just doesn't work out in all situations. The squad-building AI needs work, simple as, but I don't think it has a thing to do with SI not caring about long term gamers.

I am sure they do care about long term gamer too. But it is not their priority. This is very visible from the fact that at every version they introduce many new features, that's the priority, then they release the game, which is an unfinished product most of the time, and then they start working on fixing the obvious bugs, which there are plenty, and typically only after the second patch they work on improving the existing features, for a month or two, and then we get the dreadfull announcement, saying 'this will be something to look at in the next versions of the game'. It is always the same.

[Dave C]

Going back to the subject now: Yes, I agree with you guys. There should be at least a minimum squad size enforced in the game, as a temporary workaround until SI fixes this properly. so that AI managers will fill their squad with some lesser players from the free transfer list. It is ridiculous for a professional team to have less than ~25 senior players in the roster.

The problem is, an enforced minimum squad size would create another problem. The issue, I think, is that a combination of contract clauses, big wage rises in new contracts, and difficulty shifting the highest paid players, means teams cannot afford as many players as they need. They get stuck with half-a-dozen "Key Players" who get 80% of the wage budget. Force sides to sign players anyway, we'll just setup a pattern of collapsing teams.

For the time being, I just want to know if this happens with many teams, or are those ones with the extremely small squad are seldom? I can live with it and continue my career game that I started with enthutiasm if 1 out of 10 team has these issues. I never use FMRTE, exploiting super tactics, player search, etc. so the game provides eough challange for me and therefore it is still entertaining.

It's pretty common, especially at smaller sides. It's not isolated.

I still like the game, it just frustrates me. Makes it no real challenge to turn a side around, because you go in and fix all the stuff the AI doesn't.

I wish AI squad building will be improved significantly for FM2013, but I'm pretty sure SI is busy trying to come up with some bells and whistles to implement into the new version, as we speak, rather then first improving the current features. Some of those features won't be functioning on the game release day, and patches will follow in an attempt to fix those. In the meantime transfer AI and squad building AI will be ignored because 'it is not a priority'. This has been the story the past 5? years, and SI still has to prove me wrong about my expectations for the new versions.

Yes, reluctantly that's the conclusion I reached too.

[Dave C]

It's been a while since I played a save for longer than 4-5 seasons. It's difficult to retain the motivation to pursue any kind of long-term save with this ongoing issue. Out of curiosity, Dave C, does your annual research find the AI Squad Management issue to be gradually getting worse somehow? I could have sworn that AI squads seemed to fare significantly better as the seasons went on in FM 05/06/07 compared with the last few versions.

It's roughly the same every year, maybe a slight decline. Usually, somewhere along the line a patch seems to improve it slightly. Then it gets bad again.
It's as bad in the most recent version of FM12 as it's ever been.

[Katarian]

I still like the game, it just frustrates me. Makes it no real challenge to turn a side around, because you go in and fix all the stuff the AI doesn't.

There is a challenge in keeping your job whilst you try to sort out the crap job the previous manager has done. Especially if you take over a top club which has basically managed to maintain it's results only due to having a high reputation and being in a league that isn't properly simulated. If you make it through the first season, or have a massive transfer budget, then managing anywhere is unfortunately pretty easy.

[StormenDK]

I am sure they do care about long term gamer too. But it is not their priority. This is very visible from the fact that at every version they introduce many new features, that's the priority, then they release the game, which is an unfinished product most of the time, and then they start working on fixing the obvious bugs, which there are plenty, and typically only after the second patch they work on improving the existing features, for a month or two, and then we get the dreadfull announcement, saying 'this will be something to look at in the next versions of the game'. It is always the same.

Not really surprising ... they need new visible features to be introduced every year to justify all the gamers buying yet another version. If all they did was update the database, fix all the non-visible problems (of which there are many) and slightly fix the UI, not many would buy it at full price. And the lack of proper development cycle between that last patch and the next version, means that there will be no "major leaps" in gameplay.

I still advocate that new major features are only introduced every two years, and the other year it is a data update with fixes to existing problems and ironing out issues in gameplay that only become apparent after longer game time. But that won't happen as it is bad for the bottom line. They would not sell many "update" versions at full price.

[RBKalle]

I would pay TWICE the usual price for a version of FM that doesn't allow us to overachieve in every single season due to glaring AI squad-building and management flaws...

Or for a version of FM where I can't sign the likes of Peter Cech on a free transfer in the third season, at a club with almost no tradition (ie. reputation) in the top-flight, just because AI clubs aren't smart enough to see a bargain...


Let's be honest, the biggest share of SI's market consists of die-hard, long-time and fidelized customers... Do they (we...) really need flashy new features to be convinced to shell out 30 quids (or 50 euros...) every year?
FM is hardly an appealing game to "outsiders"... even many avid football fans will steer clear of it due to the overwhelming complexity of a football managing game and they'll gladly stick with FIFA/PES instead.

So I don't really get the need of striving to add new superflous and often frivolous features while the very key aspect of the game are either painfully showing their age or are just not working.

It's not as if the "general public" will suddenly realize football managing is cool because FM13 has interactive fan days and training camps...

[themadsheep2001]

I would pay TWICE the usual price for a version of FM that doesn't allow us to overachieve in every single season due to glaring AI squad-building and management flaws...

Or for a version of FM where I can't sign the likes of Peter Cech on a free transfer in the third season, at a club with almost no tradition (ie. reputation) in the top-flight, just because AI clubs aren't smart enough to see a bargain...


Let's be honest, the biggest share of SI's market consists of die-hard, long-time and fidelized customers... Do they (we...) really need flashy new features to be convinced to shell out 30 quids (or 50 euros...) every year?
FM is hardly an appealing game to "outsiders"... even many avid football fans will steer clear of it due to the overwhelming complexity of a football managing game and they'll gladly stick with FIFA/PES instead.

So I don't really get the need of striving to add new superflous and often frivolous features while the very key aspect of the game are either painfully showing their age or are just not working.

It's not as if the "general public" will suddenly realize football managing is cool because FM13 has interactive fan days and training camps...

We don't know that the biggest share market is all about us long term gamers who aren't fussed, and even if it was they cannot make it all about us. As for the flashy new features, there were plenty of complaints on this forum that there weren't enough new features in the game, so its a lose-lose situation for them between those who want more, and those of us (I put myself in this group) who want less, and are more interested in fixing and working on what's already there. Even then, new features are a selling point to many, so whether some of us like it or not, they are very much valid and part of the commercial process.

My biggest worry isn't that the are issues in the AI, but whether these issues are actually fixable. Because it is getting seriously left behind.

[The Hand of God]

The game desperately needs to get back to basics. Too many variables were added before the underlying, core system had been properly balanced, and, at this point, trying to patch all these bugs on a one-by-one basis is like trying to destroy an ant colony by stepping on it. Even the database (the one thing that should be relatively easy to maintain from version to version) is an incoherent mess at this point and desperately needs someone to go in and clean it up with actual gameplay in mind.

This also wouldn't scare off casual players. The "new features every year" approach is actually a strategy aimed at hardcore players and it's a trap that other developers have been consciously avoiding in efforts to revitalize sales. In recent years, EA Sports has implemented a strategy of simplification for several of its key franchises that emphasizes removing unnecessary, obtrusive features and marketing the redesigned games as being simpler and more fluid. This occurred in response to declining sales figures largely attributed to bloated, broken games that were too complex to draw in new players. The result has been rapidly recovering sales figures and, in the case of EA's golf franchise, record-setting sales.

[achilles-the-victorious]

What I have noticed is that the strongest teams are still strong. The weaker teams are now weaker, so there is a bigger gap between the strong teams and the rest, which means the top teams (mine incl.) will take a lot of points. My team takes a lot of points, so I don't see why the top 2 or 3 or 4 AI teams shouldn't, as well. They build more for experience, mental attributes, and reputation, while I build more for youth, physicality, determination, in other characteristics I aim to avoid glaring weaknesses.

I just had a 17 game winning streak in La Liga with Sevilla, drew with Valencia (also a top team) and am only top by 2 points, above Real Madrid. Simultaneously in Italy, I am managing Juventus and have done quite well, 15 wins 1 draw and 3 defeats (due to complacency, and possibly playing too many youngsters), but am 2 points off the top, AC Milan, who have gotten a really generous schedule, and made the most of it. This is in 2036. So yeah, the weak teams are weaker relative to when the game started, but the top teams have strengthened alongside my squad(s).

The thing is, it isn't a uniform disaster. Some AI squads do quite well, and most seem at least okay. We tend to remember the egregious problems because they are so noticeable. To say SI doesn't build for the long-term saves is a bit unfair, I think. They test the games to many years out and the intent is to keep the game world stable over the long haul. It just doesn't work out in all situations. The squad-building AI needs work, simple as, but I don't think it has a thing to do with SI not caring about long term gamers.

[achilles-the-victorious]

This is a FM11 save though, keep in mind.

[Validicus]

I would pay TWICE the usual price for a version of FM that doesn't allow us to overachieve in every single season due to glaring AI squad-building and management flaws...

Or for a version of FM where I can't sign the likes of Peter Cech on a free transfer in the third season, at a club with almost no tradition (ie. reputation) in the top-flight, just because AI clubs aren't smart enough to see a bargain...


Let's be honest, the biggest share of SI's market consists of die-hard, long-time and fidelized customers... Do they (we...) really need flashy new features to be convinced to shell out 30 quids (or 50 euros...) every year?
FM is hardly an appealing game to "outsiders"... even many avid football fans will steer clear of it due to the overwhelming complexity of a football managing game and they'll gladly stick with FIFA/PES instead.

So I don't really get the need of striving to add new superflous and often frivolous features while the very key aspect of the game are either painfully showing their age or are just not working.

It's not as if the "general public" will suddenly realize football managing is cool because FM13 has interactive fan days and training camps...

Would have to agree strongly with the sentiments here, especially the part in bold. AI squad management would be presumably extremely difficult to fix entirly, but if it's been this bad for several year and shown no improvement, surely this means attempts haven't been made to address it as yet? It's difficult to speculate about what is and isn't accurate with regards to SI's market because we couldn't possible know the stats and facts, but it would seem likely by this point that every football fan who knows about FIFA/PES also knows about Football Manager. With that in mind, should there be a bit less emphasis on new features and more attention given to the game's long-term problems? We all have friends who are football fans, and I'm sure some of us have failed miserably to get some friends interested in the game. Are new features ever going to get those people to want to play the game? No, it would seem more likely that we'll stop playing than of them starting to. Obviously a balance needs to be struck between attracting new customers and hanging on to long-term ones. The gradual increase in sales would suggest that SI have been shrewd in that regard, but there will come a point eventually where more long-term players become disillusioned with issues as the AI's limitations and the general issues with game difficulty. There was that poll on here not too long about difficulty where something like 65% said that they were content with the game's level of difficulty and didn't want it to be more challenging. You have to wonder though, if the AI continues to be left behind and more people get bored of constantly overachieving, might that same poll be closer to 50/50 a couple of years from now? Should SI wait until that point before acting, or will they have already lost a chunk of long-term customers? Hopefully not, but time will tell.

[Krald]

Squad building/management isnt an easy thing to fix though, it's a deep-rooted thing and there's no one single thing that can magically make it better.

[haunted1234]

Squad building/management isnt an easy thing to fix though, it's a deep-rooted thing and there's no one single thing that can magically make it better.

if you think about it longer, cant be that "complicated", the way i see how it works atm, the AI just checks his own club reputation, goes to the transfermarkt, and pretty much picks players with a good reputation and buys them, sometimes even little random, the problem is the AI doesnt seem to care that the position the player hes about to purchase is pretty good stuffed already, and he "should" be looking to improve his weak spots in the team... thats just not what hes doing

[RazvanRat]

Dave, any chance you could do a survey of this sort for the players' ages. In my experience the smallness of the squads is compounded by a reliance on ageing players and a lack of rotation to give younger players the playing time they need to boost their CA to the level where they can contribute properly (or their reputation to the level where they get selected).

[bleventozturk]

It is not just AI squad building but also lack of AI using the tools in the game that are available to us. One of the most annoying things (imo) right now, which is another factor that makes the game too easy, started happening after the introduction of agents. Every once in a while the agents offer me some hidden gems, like an 18 years old player from Bolivia, who is a 4-5* PA player for my team. There are a few very serious flaws here:
1- Most of these players are for free, because their local club has released them. Why can his former team not see his PA??? When I go check on that team's roster, all I see is a lot of grey players, and 1-2 high potential newgens, who for sure will be also available for free when their initial contract will run out, because that's exactly what happened with the player that is being offered to me.
2- I am managing a small team in Turkey, what are the chances of agents coming all the way from Bolivia and offer a local gem to me, to a small Turkish club? I shouldn't be even aware of this player's existence, unless I send scouts to Bolivia. Agents make it too easy for us to become aware of such gems.
3- Now the most important one: Agents make it too easy ONLY FOR US. How do I know? Because most of the time when a hidden gem is offered to me for free, there is no any other single AI controlled team that is interested in this player. Why? Did the agent offer that player only and only to me? If yes, why? How is that realistic? If he offered it to all the teams in Europe, why are none of them interested? Can they not see his PA? If it is purely dependent on the reputation, are there no clubs in all the divisions of all the leagues in Europe who have a matching reputation? Impossible. So this means, either AI has absolutely no clue of bargains (high PA newgens being available for free), or a bug in the game causes these offers to be made only to human managers. Either way it is a game breaker. The agents feature doesn't work, so SI never should have included it, until they get it to work properly. Feature like this one have a significant effect to the whole gameplay, they should be implemented only if it works correctly. SI includes it in the game nonetheless, because all they care about is to be able to say 'we have such and such new features' when they announce the new version. Hence every new version has a few new features that don't work, and acutally make the game worse than it is without them.

[Dave C]

For the first time, I've noticed messages about players being unhappy as they feel the squad needs thinning. Checking the sides in question, both have less than 30 players between first and reserve squads. So there's an issue that FM doesn't even seem to see squad depth as a good thing.

[bleventozturk]

For the first time, I've noticed messages about players being unhappy as they feel the squad needs thinning. Checking the sides in question, both have less than 30 players between first and reserve squads. So there's an issue that FM doesn't even seem to see squad depth as a good thing.

I actually see this as a good thing because this will prevent the human managers from having a very deep roster to a certain degree and hence an unfair advantage over AI controlled teams. And if you think about it, this may be even somewhat realistic too. Right now I have 24 players in my 1st team, and around 10 in my reserve team. Out of the 24, I would say only 15 of them get regular playing time. Another 5 get very limited PT, and the rest almost never play, unless the squad is hit by long term injuries. So, these few guys (who almost never play in the first team) are 'always available for reserve team', and there is another 5 from first team who are 'available for reserve team until they are match fit'. This set-up gives me 6-7 players on average who fill the starting 11 of the reserve team. Therefore, only 4-5 players in my reserve team roster gets significant playing time in the reserve league. So, if I go ahead and reduce my roster from 34 to 28-30 everybody would get some playing time. Considering that there are always 2-3 players in the U18 team that are good enough to play in the reserves, I may even reduce this number to 25-27 easily.

I know some people like to have very deep rosters in the game, and in real life also some teams have 40+ players, in the game AI controlled teams tend to keep smaller rosters, and if your players become unhappy when you have a deep roster (because they are worried about not getting enough playing time), this will at least help to keep things balanced between human managed teams and AI controlled teams. So, not a bad thing necessarily.

[Dave C]

I actually see this as a good thing because this will prevent the human managers from having a very deep roster to a certain degree and hence an unfair advantage over AI controlled teams. And if you think about it, this may be even somewhat realistic too. Right now I have 24 players in my 1st team, and around 10 in my reserve team. Out of the 24, I would say only 15 of them get regular playing time. Another 5 get very limited PT, and the rest almost never play, unless the squad is hit by long term injuries. So, these few guys (who almost never play in the first team) are 'always available for reserve team', and there is another 5 from first team who are 'available for reserve team until they are match fit'. This set-up gives me 6-7 players on average who fill the starting 11 of the reserve team. Therefore, only 4-5 players in my reserve team roster gets significant playing time in the reserve league. So, if I go ahead and reduce my roster from 34 to 28-30 everybody would get some playing time. Considering that there are always 2-3 players in the U18 team that are good enough to play in the reserves, I may even reduce this number to 25-27 easily.

I know some people like to have very deep rosters in the game, and in real life also some teams have 40+ players, in the game AI controlled teams tend to keep smaller rosters, and if your players become unhappy when you have a deep roster (because they are worried about not getting enough playing time), this will at least help to keep things balanced between human managed teams and AI controlled teams. So, not a bad thing necessarily.

I don't agree that it's there to balance things out; I have probably the biggest squad in the PL now, I've heard nothing. I've only seen it at AI teams that have what I consider an inadequate number in the first place.

[brightonrock]

The issue seems to me (without any data to support my theory I should add, this is purely observational!) that AI teams sign players based on their CA and/or reputation, rather than necessarily a need for that player within their squad. Plenty of times in longer term games I've seen, for example, a team sign a high quality striker when they already have 5 strikers, so they never play the guy they've just signed for 20m, before selling him 3 seasons on for 2.5m having played 4 games. It seems the AI spots a player who is available for below their market value, has a lot of room for improvement (i.e. CA>PA difference) or is on a free, and signs them whether they need the player or not. Almost as if the AI can only truly assess a player's worth to the squad once he's signed, rather than scouting and making that decision before making a bid. Teams like Man City and Barca do that a lot in my games; I once had a game where City signed 10 attacking midfielders/wingers, and then played a 451 formation so only two could ever be selected. From memory the cheapest player in those ten was about 14m. Now we can joke about City doing that IRL, but the reality is that if you're going to be playing two wingers, the most you'd really need is 4, perhaps 5 tops. To sign 10 just because the players are available is ridiculous. That's why human teams are easy to convert into dominant sides with stronger squads; the AI will either over-fill positions or not address it's weaknesses, whereas we can pick up the players the AI has ruined on the cheap and build more balanced squads.

[achilles-the-victorious]

There is a real preference for 27 y.o. + players in AI squads, but seeing as that is when players are at their peak, I don't see a real problem with it. I think the wages are very high for those players, a lot of the time, esp. if you sign them once they are already established stars. Hence, the overall wage budget can only handle so many players at their peak.

I have watched my opposition in La Liga and Serie A rebuild now for many seasons (into 2036, managing two clubs for the last 12 years, so that is what, 36 game years), and it never goes quite as badly for them when they need to replace a star as I would have hoped

Like Inter, they had a dominant fullback for the last 10 years, but he is now 33. Well, they go and buy a 26 year old, top notch (maybe 10% worse than previous world class fullback) fullback from Russia for something like $26m. If they get 5 years out of him, then thats quite good business, as they spent no time developing the player whatsoever. But I myself try to pursue that strategy as little as possible, because of the budget implications. I can pay a 26 year old MUCH less if I had him on contract since 22 or 23, than if I just bought him.

Dave, any chance you could do a survey of this sort for the players' ages. In my experience the smallness of the squads is compounded by a reliance on ageing players and a lack of rotation to give younger players the playing time they need to boost their CA to the level where they can contribute properly (or their reputation to the level where they get selected).

[bleventozturk]

I don't agree that it's there to balance things out; I have probably the biggest squad in the PL now, I've heard nothing. I've only seen it at AI teams that have what I consider an inadequate number in the first place.

I didn't say SI implemented it to balance things out. I said it could work that way, but from what you are saying now it is not what I thought it was.

I've never seen it myself, but does it happen to the teams who get relegated and still have their best players under high wages contracts? If so, maybe the problem is that those teams budgets are less than what they are paying to their already undersized squad, and maybe the players are nervous about their contracts being terminated, it's just that the warning message the game creates doesn't reflect the problem correctly. Definitely an issue, but hopefully happening with only small number of teams that already have budget issues and get relegated on top of that and it gets worse for them.

For whatever is worth, I'm entering 4th season now, I have 5 biggest leagues and Turkish league active, almost all the other leagues are 'view only', so far I haven't seen a major problem. In Turkish league all the teams remain competitive so far, and they make good transfers. Not quite as logical as mine, but close. As I said before, the thing that bother me most is those rare gems that are offered to me by their agents, from inactive countries like Bolivia, Panama, etc. It almost looks like their agents hide these players from everybody else and offer them to me only. I just refuse to sign them to avoid an unfair advantage over AI. It is so tempting, because they are for free, but I managed to resist so far

I got curious and investigated a little bit. The game creates a few high PA players from odd countries every season, and their original clubs just release them at the end of their initial contracts, so they are out there to be signed for free, however no AI team notices them for a long time, sometimes for more than a season. Probably something with their reputation is completely messed up, that should be the reason nobody cares about them, since the whole AI transfer system is based on reputation. Eventually somebody picks them up (although not playing for any club, they keep playing for their U21 national team, so their reputation slowly grows I think and eventually they become noticed by AI teams) but if I only wanted I could create a team for free with these super talents from odd countries.