Research on players' sharp decline (with data!)

[RBKalle]

As of a late there's been a bit of a debate about players declining way too early and way too sharply in FM.

Since it's been something I've been noticing for some iterations, while the board's semi-official response was "we're vaguely aware of that, but all in all we're happy with the balance (and we'd rather have players retiring at 35 than hogging their Starting XI spot til they're 40", I decided to go past what had been dismissed as "anecdotal evidence" and I ran a dedicated test-savegame.

SETUP

 

• England (EPL + Championship)
• Italy (Serie A + Serie B)
• Spain (La Liga)
• Germany (Bundesliga)
• Brazil (Série A)
• All International players for European nations
• No additional staff
• Approx 40k players

STARTING DATA

A staggering 1493 players aged 35 and above, with a fair amount of guys pushing 40 still able to perform at a low-EPL level, and even a handful of respectable middle-aged men.

But what about their mobility? The Oldboys FC can still be technically sound, but surely they'll need a walker or a scooter to move around the pitch? Guess again...

Until age FOURTY, the majority of the players are still on the right side of the double-digits! And the distribution feels very homogeneous and consequential, with no major dips, except the weird absence of superslow 39 and 40yo.

OUTCOME

I went on holiday and simulated all the way til July 2023.... and that's what happened in five short seasons.

That's 478 players between age 35 and, well... 39!

Basically everybody retires around age 36. The two 39yo guys are, as you can see, lower leaguers (two GK from Guinea and Germany... no idea what they're still doing in the game). And the good 38yo one is none other than Cristiano Ronaldo (CA 161, Acc and Pac 9, still physically impressive)

At that stage, there's little point in adding a graph about Acceleration, but for those who wonder, only 23 players have Acc  10-14 (nobody 15 or above), and out of those, 22 are 35yo, 1 is 36 (Luis Suarez, but Pace 9).

Again, CR7 is the only older player in the Acc 6-9 bracket, with 7 36yo players still being somewhat responsive.

 

CONCLUSION

The evidence looks pretty damning IMO.

In only five years, the gameworld has lost ONE THOUSAND players (-68%) over 35, with the tail of the bell curve coming much earlier and steeper, basically a cliff.

Basically the game can't take care of its aging players, and the original players are the only ones able to have a career past age 35, while those who start younger will experience a dramatic, fast and unforgiving physical decline that'll see them out of the game much earlier.
Unless, of course, they're Cristiano Ronaldo..

I can't honestly think of a different explaination besides the decline model for physical stats being setting in too early and with no chance to slowing it down. And things won't likely improve in a newgens-only world, because if the rather well-rounded original players didn't stand a chance, how can the majority of the somewhat "flawed" newgens fare better?

 

SUGGESTION

Easy: reduce the decline rate and/or the impact of accelertion/pace/strength in the ME. Smart, but slow, players can still have a role to play, but not if they can't even get to the ball or get hold of it for long enough to work their magic.

But the decline model is the #1 priority here. It doesn't need to go to the other extreme (like in an old FM where 38yo legends were still in the Starting XI despite being very much past their prime), but there's nothing wrong with having the best trained, more focused, physically healthy and motivated over 35s to still be capable enough to enjoy a final stint somewhere. And a selected few could indeed still be going strong at a high level in time for their 40th birthday. Instead of, you know, having turned into useless coaches or scouts at 36.

 

 

[fish18ish]

Thanks for doing this. I had previously thought (based on my own anecdotal evidence) that the game wasn't far wrong with player decline, but this has forced me to re-think. 

Now there is hard data out there SI will presumably have no choice but to make some adjustments. 

[santy001]

It's very interesting stuff to look at. While this isn't critical in anyway there are two major things that stand out to me just that unfortunately can't really be tested.

- Is the first great filter of sorts around 34-36 being hit because players are without clubs, or without an offer in the last year of their contracts? That could in itself be robbing saves of players in their twilight years.
- Do ambition and professionalism which play a role in decision making in regards to retirement lead to the players who could perhaps keep on going longer, because they're professional & have high natural fitness quitting earlier than their more likely to decline counterparts?

I have no knowledge of how the system works behind the scenes, but it has been mentioned that the game does need to push the older players on their way as previously clubs held onto them for too long. It has been something I've been thinking of since my own posts on the matter and the figures do raise some other questions:

- Are some players slipping through the net on research with too high an overall CA?
- Are some players, even if their CA's are right a little too high physically?

It's definitely worth finding out whether the starting point is an issue, as well as several years down the line being an issue because ultimately almost no development model would produce similar results if we're too far off the mark collectively as researchers.

- - -

There's things in the game already such as the player preferred actions "Will retire at the top" and "Will play football as long as possible" - but they are very difficult PPM's for researchers to utilise and I don't know even if the use of this PPM will bypass the game's seeming hard filter of 39 @Seb Wassell maybe you could clarify? It would feel somewhat redundant if it didn't do anything for say goalkeepers who can keep going to 39 easily. 

It isn't something you can search for in the game, not all player traits are able to be searched up so it's not sure to know how many players have either but I've only used will play football as long as possible for Rory Delap who wanted to play longer, but a couple of bad injuries in the final couple of years of his career put an end to that for him. It's difficult to say what the usage of this is among ageing players who are still there in 2019/2020 but those who are then coming into the age bracket by 2023. It wouldn't impact development as much, but again it could impact the types of players still being there and whether they are suited to ageing well. It wouldn't surprise me if Ronaldo has this selected, as I frequently see him carry on and he has expressed a desire to keep playing until he's 40.

- - -

In fact, if you do have the data @RBKalle ambition, professionalism and natural fitness scores would be very interesting to see. It's only a thought/theory but based on how ambition and professionalism impart a proclivity to retire out of a sense of not dragging their name/legacy down it seems like maybe the majority of those best suited to perhaps age better past 35 are calling it a day at around 34/35. That's based on nothing other than the write-ups on the attributes and it therefore sounding plausible.

[RBKalle]

21 minuti fa, santy001 ha scritto:

In fact, if you do have the data @RBKalle ambition, professionalism and natural fitness scores would be very interesting to see. It's only a thought/theory but based on how ambition and professionalism impart a proclivity to retire out of a sense of not dragging their name/legacy down it seems like maybe the majority of those best suited to perhaps age better past 35 are calling it a day at around 34/35. That's based on nothing other than the write-ups on the attributes and it therefore sounding plausible.

I may extract the data, but to be completely honest, I think it'd be just an exercise in futulity (and quite a chore) at this stage.

In only five years, the game wiped out around 70% of the players over 35... In that gameworld there are just as many active footballers aged 37+ as the original db had players aged 40+ still able to perform at a Championship level and above! That doesn't look like a "survival of the fittest" scenario, it looks like an almost indiscriminated "agecide". After all, if CR7 is the only top player who has survived the cut and didn't fall victim of the "37 Club", maaaaaybe the bar has been set a bit too high?

The personality+fitness idea could be reasonable had the drop been less dramatic. But unless the majority of the original db players born from 1984-1988 (who hit the 35-39 age bracket during the simulated seasons) have low professionalism, ambition and natural fitness, I don't really buy that explaination.
 

For debate's and completeness' sake, I'll crunch some more data. Even though the sample from 2023 looks really depressing.

[XaW]

This is a very interesting look at things. The one thing I would like to know, is how high is the goalkeeper percentage of both those numbers. I suspect that the percentage is much higher in the "real world"-number than the simulated ones. I suspect this based on my own anecdotal experience, but it would be neat to see it in such a test as this.

I also wonder if the figure will change 10-15 years on, when most (all?) real players have retired and see if the newgens will bring the number back up again. I mean, this could be a transitional issue when going from "all real" to "all newgens"?

[santy001]

@RBKalle one of the remits under the attributes Ambition & Professionalism is that when scoring highly, they should increase the proclivity a player has to retire.

Yet the way in which they work also mean that they're more likely to be the players best suited to ageing well. There's an awful lot of nuance to this so on one hand, I wouldn't suggest reading too much into it because the section covering the mental characteristics in the researcher guide is around 2000 words in total (just as an aside in case anyone reads this and asks for it in full - SI have said its fine to mention researcher definitions for attributes, but they're constantly iterated on year upon year, so I don't ever post the content on attributes in full because otherwise it potentially ends up creating a legacy of outdated information)

This makes me think there could well be quite a high drop out rate at 34 or 35.

[Seb Wassell]

I appreciate and respect the OP here.

I would however like to clarify that, 1. the original response was not to dismiss it, and 2. that any changes made are not "easy", far from it.

We have extensive internal data on this and many other progression related indices, I'd like to assure you that it is something we are constantly tweaking, but there is no "quick fix" for anything in this module. We are working towards a balanced representation of reality for the next version, as we do for each version.

P.s. This sort of thing belongs in the Bugs section, ideally with attached saves/data. We can then kick off any investigation that is required. 

[RBKalle]

Il 21/2/2019 in 10:39 , XaW ha scritto:

how high is the goalkeeper percentage

(Un)surprisingly, our common perception of more keepers sticking around, especially the original ones, is off by a good margin.

 

 

The original db has, maybe unexpectedly, much fewer aging GK than outfield players.

As usual, year 2023 is a graveyard, where indeed keepers fare much better than other players, managing to survive in numbers for one, short extra year at age 36, where the cut is slightly less severe.

Still, I maintain it's kinda moot trying to find patterns in a gameworld where turning 37 is a death sentence, Children of the Corn-style, than only a handful of players can postpone.

[XaW]

1 hour ago, RBKalle said:

(Un)surprisingly, our common perception of more keepers sticking around, especially the original ones, is off by a good margin.

(...)

Huh, that actually surprised me since it seem to contradict my own experience where there is lacking anything resembling aging goalkeepers. Your numbers show a very massive decline on older players, but for goalkeepers the decline is much less than outfield players.

Might I ask how you extracted the data? If it's easy enough I would like to see how the rate is in my own game. Manually counting it would be a terribly dull job.

[enigmatic]

On ‎21‎/‎02‎/‎2019 at 10:04, santy001 said:

@RBKalle one of the remits under the attributes Ambition & Professionalism is that when scoring highly, they should increase the proclivity a player has to retire.

Yet the way in which they work also mean that they're more likely to be the players best suited to ageing well. There's an awful lot of nuance to this so on one hand, I wouldn't suggest reading too much into it because the section covering the mental characteristics in the researcher guide is around 2000 words in total (just as an aside in case anyone reads this and asks for it in full - SI have said its fine to mention researcher definitions for attributes, but they're constantly iterated on year upon year, so I don't ever post the content on attributes in full because otherwise it potentially ends up creating a legacy of outdated information)

This makes me think there could well be quite a high drop out rate at 34 or 35.

How justified are these assumptions about ambitious and professional players retiring in their prime anyway?

You see very ambitious and professional elite players go for seasons abroad in lesser leagues (Lampard, Iniesta, Xavi, Schweinsteiger, Ibrahimovic), drop down the divisions (less common, perhaps, but Terry springs to mind) or accepting roles as bit part players (Totti, Giggs and of course the legend that is Peter Crouch) when all of them could have comfortably retired much earlier. 

Sure, you also get Zidane retiring midway through the World Cup final, but I'm not sure there's any systematic pattern here. If I had to guess, I'd imagine Messi will retire when he stops making football look stupidly easy, and Ronaldo will carry on trying to be a supersub poacher after his legs stop working. Both are incredibly professional and ambitious 

If you look at players in their 30s with PA>150 and CA<100 you've got good pros like Kranjcar and Malbranque as well as Cassanos.

There also aren't many of those players in the starting database, fewer even than 33+ year olds with 150+ CA, so I think the rapidity of early 30s decline in FM is the bigger issue. I can think of one notable exception in-game where my 33 year old centre back (acceleration 9, pace 11) lost just 0.2 attribute points all season, having been a regular player and not had any serious injuries (and appears not to have lost any the low-detail season before) and also being a professional type.  Would like to see this be a lot less rare in FM though...

[santy001]

19 minutes ago, enigmatic said:

How justified are these assumptions about ambitious and professional players retiring in their prime anyway?

They're pieces of a jigsaw. I can't speak for their impact with any degree of expertise I'm just thinking with that bit of knowledge in the context of in-game anecdotal information in regards to players I know the stats for. 

Personally I hadn't interpreted it as retiring at 31-33 because there is also a bit in there that highlights these players shouldn't retire prematurely either, that you can already see in professionalism and this one in ambition:

There is enormous nuance to these attributes, as I said, the section on these attributes comes in at around 2k words in total. So I probably misrepresented my point initially.

I work under the assumption that 34/35 is the "sweet spot" of going out at a high point, again anecdotally looking at how players decline with age in the game (when they have natural fitness around 15+) so I was thinking perhaps the players best suited to keep going are heading out the door early on.

That's just a theory, on the most flimsy of foundations I concede it's just me sharing the thoughts that pop into my head on the matter. 

- - -

As I mentioned as well, I'd be surprised if Cristiano Ronaldo does not have the "will play football as long as possible" preferred action because based on his interviews and current mindset it makes sense. That ought to, in my mind at least, override the normal point at which retirement would be decided upon. Again, its anecdotal but in all my games in FM19 Ronaldo has continued until 39. 

[RBKalle]

Il 21/2/2019 in 11:04 , santy001 ha scritto:

one of the remits under the attributes Ambition & Professionalism is that when scoring highly, they should increase the proclivity a player has to retire.

Those are the 2018 data for AMBITION and PROFESSIONALISM

This time the outcome was more in line with the general perception. Amb and especially PRO tend to be higher among those who stick around longer, but the most prominent bracket is the 10-14, not the 15-20, which is rarer, but still a key factor for the >40s.

Professionalism seems to play a bigger role in longer careers, but again, that's the original db here, so older players being rated as professional is a reflection of a real-life scenario. Nothing really conclusive since NONE of those hundreds of ambitious and professional 35yo has made it to 2023.

(P.S. I simulated one extra year and in 2024 Cristiano Ronaldo is a terrible and unemployed Assistent Manager, despite his >18 attributes and a >150CA)

Anyway... Since the sample size for 2023 is much smaller and less diverse in age (35-36, then there are just a few survivors), I decided it wasn't worth it going into painstaking details.

However, I searched for players with Ambition, Professionalism and Natural Fitness >10.

Indeed FM19 seems to heavily favour players with above-average Ambition, Professionalism and Natural Fitness. That combination is a player's only hope to be around at age 35. It's still around the same ratio at 36 and 37 but it doesn't really prevent the game from (arbitrary?) retiring a disproportionate amount of players anyway.
As the tail of the curve approaches, the gap gets smaller and likely completely random... One of the "good" 38yo is CR7, but the 2 39yo players are the GK of Dutch club Excelsior and the GK of an unknown German team (Grimma). So that kinda ruins the picture.

 

In 2018 we got a very similar distribuition in terms of %, but obviously with much higher figures

 

So, the "good mentality vs poor mentality" % is very much the same, which means the game DOES consider Ambition, Professionalism and Nat Fit as important features for a player's career.

The bad thing is that doesn't seem to mean much once the gameworld starts to "butcher" physical attributes and has an astronomical retirement rate around age 36, whereas the original db (and the real world, I'd say) has been treating aging footballers in a much more forgiving way, especially when they are dedicated professionals.

I really don't think all those additional data change much my original conclusion.

FM19's aging dynamics COULD work fine, if the game just didn't put almost every 36yo player on a bus to a retirement home.

[RBKalle]

3 ore fa, XaW ha scritto:

Might I ask how you extracted the data? If it's easy enough I would like to see how the rate is in my own game.

Genie scout.

Filters work like a charm. The most boring part is changing them for age, attributes.

And of course the time needed to simulate through several years in the game, but if you have a save with a reasonable setup, you can just take a quick peak at what you've created while playing.

[enigmatic]

4 hours ago, santy001 said:

They're pieces of a jigsaw. I can't speak for their impact with any degree of expertise I'm just thinking with that bit of knowledge in the context of in-game anecdotal information in regards to players I know the stats for. 

tbh I was mostly thinking from the point of view of whether SI's assumption about professionalism making people more likely to retire at their prime was justified than whether your guess was a reasonable one. But as you said, there's a lot more text qualifying that assumption that I haven't read  

[grahamf]

I was pondering this thread in the shower this morning and would it maybe make sense for players to have a "love for the game" hidden attribute as well?

 

As an example I like my job well enough but as soon as my pension looks suitable I'm more than happy to retire and spend more time with my sofa. My boss on the other hand says he never wants to retire. In football examples you could compare, say Kevin Prince Boateng who says he doesn't really care about football with somebody who lives and breathes the game. The former might still be a professional, work hard, trains properly, ensures he is prepped for matches but does not feel the need to be turning out for Wrexham aged 37 when he has millions in the bank.