Potential Ability

[Guest]

Nope, there doesn't have to be an explicit limit in the same way that a good racing engine doesn't hard-code the top speeds of its cars.

A good racing engine should hard-code the top speed of its cars otherwise it runs the risk of conflicting with the laws of physics & aerodynamics.

Stop trying to shoehorn that one as an appropriate comparative model.

Most racing simulators cater for unexpected scenarios. Imagine a Golf GTI being hit from behind by an F1 car in Gran Turismo. The Golf GTI's speed will doubtless increase.

Truly awful example & missing the point of my post by some considerable margin, go away & think about what you've just said.

[x42bn6]

A good racing engine should hard-code the top speed of its cars otherwise it runs the risk of conflicting with the laws of physics & aerodynamics.

Nope! A great racing simulator's physics are so awesome that the top speed of a virtual Golf GTI under ideal conditions is the same as a real Golf GTI under ideal conditions.

Think about it. Do you think SI would consider adding a "maximum score" hard coded scoreline to prevent it conflicting with the laws of physics and aerodynamics? How about a maximum age? Or maximum price. Or maximum number of players in a team ("Nobody has that many players, right?").

You don't always need to add an explicit maximum value. If anything, that's an assumption, and every simulator is better with fewer assumptions.

Truly awful example & missing the point of my post by some considerable margin, go away & think about what you've just said.

It's not a terrible example. A Golf GTI's maximum speed in reality is measured under ideal conditions (air speed, track temperature, type composition, humidity, etc.). This goes out of the window if, say, a Golf GTI goes downhill, or is hit from behind by an F1 car.

The "ideal conditions" are comparable to the scenarios that a researcher thinks of when assigning PA. However, there could well be some scenarios they didn't think plausible (i.e. a modified Golf GTI/a player hits an average rating of 8.50 over a full season in the first-team), or some strong external circumstances that fundamentally change it (i.e. a different track type/lower-league player moves to Premier League side), that affect the top speed/"PA" both positively and negatively.

This is why hard-coding in a top speed is not necessary nor recommended for a good racing simulator. All you need to do is get the physics correct, and the correct top speed will directly result from these physics.

[bleventozturk]

The same goes for the youth intake where I can tell straight away that most won't even be professional footballers in a year or two let alone at any decent level. I then discard most if not all of my intake. I understand that unlimited PA is unrealistic and I understand that dynamic PA is probably very difficult to achieve and may end up producing too many top players. If the current model is the best and most realistic that can be achieved thats fine but maybe we could have PA ratings removed, at least for players that are say under 21. The number would still exist but be hidden from the player and AI meaning that not all high PA youngsters would be spotted so easily and will move to top clubs later when their CA has increased to a reasonable level. If you have a bit of money and a good scouting knowledge you can fill your reserve and youth teams with awesome youngsters pretty quickly as it stands.

A scout report with no PA rating could then tell you a player has 1 star CA with the ability to improve considerably along with strengths and weaknesses and then you make choices based on personality and match ratings. There would be a much larger chance for transfer mistakes involving youngsters then. It would also stop top AI teams snatching quality youngsters and never giving them football and, ultimately, ruining them.

I agree that the scouts could be made less accurate than they are currently in the game.

I am not going to be bothered more with other posts that say 'such and such players have their PA sat on 130, then in real life they have a good season and in next version of the game their PA is set higher, so you see, PA should be dynamic'. They don't even realize that all that is happening is that the particular researcher guessed the PA of that player wrong, just like a scout guessing a player's PA wrong. When you first get a player, your scout can say he is 4*, a year later he can say he is only 2.5*, start giving playing time to him and if he performs your scout can say he is 3.5*. But the bottomline is, every player does have a limited PA in real life, and in the game too.

To reflect the real life better, the scouts can be made less accurate, but that's all. Players should still have a defined PA, and we wouldn't know what that value is, and it would reveal itself as the player gets older.

[bleventozturk]

Dont most researches use negative PA's for players now unless they have a very good idea of the player? So there is already a random element to most PA's in the starting database. From that point onwards in regards to regens, FM is god, it cant be wrong, it makes the players, so it already knows exactly how good they can be. I've always said this is more of an issue with the starting database, for regen players the system works almost perfectly. Scouting is the bigger issue, its too easy to know too much about a player very quickly.

+1

This post summarizes the situation perfectly, and that's all there is to it.

[x42bn6]

To reflect the real life better, the scouts can be made less accurate, but that's all. Players should still have a defined PA, and we wouldn't know what that value is, and it would reveal itself as the player gets older.

Do managers really know the potential of their players as they get older? Late bloomers are an exception, where nobody expects them to suddenly develop.

[x42bn6]

Dont most researches use negative PA's for players now unless they have a very good idea of the player? So there is already a random element to most PA's in the starting database. From that point onwards in regards to regens, FM is god, it cant be wrong, it makes the players, so it already knows exactly how good they can be. I've always said this is more of an issue with the starting database, for regen players the system works almost perfectly. Scouting is the bigger issue, its too easy to know too much about a player very quickly.

It still misses the point that some maximum limit is defined (the upper limit of that PA range). What if a player's real PA is raised above this? Imagine, say, a 16-year-old at -7, but later hitting 180?

In addition, the "PA problem" this applies to every single player with a fixed PA in the database, not just a youngster. Late bloomers, for example, buck the trend.

And regens have the same issue, as they have PAs. Take, say, Aaron Ramsey the real player, and a regen who is an exact clone of Aaron Ramsey. If Aaron Ramsey's PA is raised in a future FM version, then it will be because Aaron Ramsey the real player has performed well in reality. So if Aaron Ramsey the virtual in-game player performs comparatively well to reality, then his PA should correspondingly rise. The same rule applies to regen clone Aaron Ramsey.

In other words, this happens:

PA 140 ==> exceptional real-life season, luck with injuries ==> researcher raises PA ==> PA 150

I want to see something like this:

PA 140 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 150

Where there is a 1-1 correspondence between "exceptional real-life season" and "average rating 8.50".

And note that in this example, nowhere do I mention "real players". It happens for every player, be it real or regen.

This also has the advantage that the engine is not dependent on the data (i.e. the game is agnostic towards the type of data - regen data or real data), which encourages reuse and flexibility. Coding a system dependent on data will create issues further down the line.

Certainly, some form of randomisation in the form of negative PA ranges is possible, to mimic the measurement errors of the researcher, but this can be used in tandem with fixing PA itself.

[postal postie]

It still misses the point that some maximum limit is defined (the upper limit of that PA range). What if a player's real PA is raised above this? Imagine, say, a 16-year-old at -7, but later hitting 180?

In addition, the "PA problem" this applies to every single player with a fixed PA in the database, not just a youngster. Late bloomers, for example, buck the trend.

And regens have the same issue, as they have PAs. Take, say, Aaron Ramsey the real player, and a regen who is an exact clone of Aaron Ramsey. If Aaron Ramsey's PA is raised in a future FM version, then it will be because Aaron Ramsey the real player has performed well in reality. So if Aaron Ramsey the virtual in-game player performs comparatively well to reality, then his PA should correspondingly rise. The same rule applies to regen clone Aaron Ramsey.

In other words, this happens:

PA 140 ==> exceptional real-life season, luck with injuries ==> researcher raises PA ==> PA 150

I want to see something like this:

PA 140 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 150

Where there is a 1-1 correspondence between "exceptional real-life season" and "average rating 8.50".

And note that in this example, nowhere do I mention "real players". It happens for every player, be it real or regen.

This also has the advantage that the engine is not dependent on the data (i.e. the game is agnostic towards the type of data - regen data or real data), which encourages reuse and flexibility. Coding a system dependent on data will create issues further down the line.

Certainly, some form of randomisation in the form of negative PA ranges is possible, to mimic the measurement errors of the researcher, but this can be used in tandem with fixing PA itself.

your seeing it from the wrong angle. there's no reason to re-rate the PA up if a player does well. the PA automatically assumes the best possible scenario already.

if the player struggles with injuries/game time, they are less likely to reach their PA. so the end product of your example and the way the game currently works is the same.

[Cougar2010]

I want to see something like this:

PA 140 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 150

PA 150 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 160

PA 160 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 170

PA 170 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 180

PA 180 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 190

PA 190 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 200

\o/ Winner

and x42 thinks this is better than the current CA/PA system.........

[milnerpoint]

There also needs to be a distinction between class and form when it comes to attributes. Players do not need 180PA's to perform like the worlds best, they can do it with a much lower PA and better distribution of attributes, and as importantly, good hidden mental attributes. If a player is playing well, doesnt mean he has become better, like above in the example of a player scoring 50 goals with finishing 13, why does this player suddenly need finishing 17 or whatever it could rise too, whats the problem with the player continuing to score that number of goals with finishing 13? There has to be a distinction between form and class otherwise we will never be scouting for the best players, just the players in the best form.

[Cougar2010]

Lets look at the flip side as well:

PA 140 ==> average rating 5.50, bad luck with injuries ==> game re-rates PA ==> PA 130

PA 130 ==> average rating 5.50, bad luck with injuries ==> game re-rates PA ==> PA 120

PA 120 ==> average rating 5.50, bad luck with injuries ==> game re-rates PA ==> PA 110

PA 110 ==> average rating 5.50, bad luck with injuries ==> game re-rates PA ==> PA 100

PA 100 ==> average rating 5.50, bad luck with injuries ==> game re-rates PA ==> PA 90

PA 90 ==> average rating 5.50, bad luck with injuries ==> game re-rates PA ==> PA 80

etc

etc

\o/ Double Winner

and thats before we consider the flaws in how players are rated in matches.

EDIT

Torres would be a great example 170odd PA to 200PA to 50PA in two seasons, I think even the researchers do a better job!

[marty78]

In addition, the "PA problem" this applies to every single player with a fixed PA in the database, not just a youngster. Late bloomers, for example, buck the trend.

PA 140 ==> exceptional real-life season, luck with injuries ==> researcher raises PA ==> PA 150

PA 140 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 150

The players actual potential does not increase though just the researchers view has changed from judging CA.

This is where the visible PA is a problem for me, if that was hidden all you would know is that you have a good player playing well, who, because of his age, has the ability to become better. It should not matter what that figure is just that it is unknown to user and AI. I know a star rating is not a figure and is a bit more vague than that but in principle it is the same thing just represented more loosely. I don't think it is necessary to change the PA just hide it and rate players on CA. Then if you had a 3* CA player of 20 years, old he could be a 5* who continues to improve or a 3* who has peaked early but the key is you wouldn't know.

Take a wonderkid like Neymar, for all we know how good he is now might be almost as good as he ever gets or he may be the next player of Messi's stature. All we should know in game is that he is a star with age on his side and more to learn.

[JamesB]

I must say, I think everyone has been pretty cordial in this thread so far, good job guys

I really don't like the idea of PA changing depending on ratings. That in itself suggests that the PA of the player was wrong to begin with and will undoubtedly be wrong again in the future. That just isn't how my view of PA works, although I respect that others have a different understanding and they are equally valid.

PAs of real players will be wrong. They'll be wrong almost all the time. The researcher has a 1/200 chance of getting the PA 'correct'. I don't understand why this bothers people so much. It is unlikely that the researchers got a player's CA or individual attributes 'correct' either, especially relative to another player's. It is entirely possible that every single attribute a particular player is accorded is not a true reflection on a 1-20 scale of their ability in real life (especially outside the big leagues). No one seems upset by that.

[x42bn6]

your seeing it from the wrong angle. there's no reason to re-rate the PA up if a player does well. the PA automatically assumes the best possible scenario already.

Sadly, it doesn't. Which is why PA can be re-rated upwards by researchers over time.

The PA isn't the game's assumption - it's the researcher's assumption for real players, and a sample of some randomvariable for regens.

if the player struggles with injuries/game time, they are less likely to reach their PA. so the end product of your example and the way the game currently works is the same.

PA works to a certain extent for underachieving players but not overachieving players.

A player at 120/120 would never develop even he averages 10.00 in the Premier League. If a player averaged 10.00 in reality, he would have his PA raised.

I want to see something like this:

PA 140 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 150

PA 150 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 160

PA 160 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 170

PA 170 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 180

PA 180 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 190

PA 190 ==> average rating 8.50, luck with injuries ==> game re-rates PA ==> PA 200

\o/ Winner

and x42 thinks this is better than the current CA/PA system.........

If this is over time, I don't see it as a problem. Messi probably averages around 8.50, and if a player performs exactly like Messi for many seasons, why wouldn't he turn out like Messi?

What tends to happen in reality is that the performances taper off due to form and the fact that a player is eventually found out, as the opposition adapts to this outstanding player. However, if a player still defies the odds, then that is the next Lionel Messi, regardless of what a researcher thought of him at age 16!

There also needs to be a distinction between class and form when it comes to attributes. Players do not need 180PA's to perform like the worlds best, they can do it with a much lower PA and better distribution of attributes, and as importantly, good hidden mental attributes. If a player is playing well, doesnt mean he has become better, like above in the example of a player scoring 50 goals with finishing 13, why does this player suddenly need finishing 17 or whatever it could rise too, whats the problem with the player continuing to score that number of goals with finishing 13? There has to be a distinction between form and class otherwise we will never be scouting for the best players, just the players in the best form.

Form is obviously only going to be one part of the determination of "new PA". I believe we will need something like "talent" in the equation, for example.

Form is only good if it can be sustained, too. Second-season syndrome shows this on a team level, where opposition teams are now wary of a team that surprised them before, and play more defensively/smartly. Most teams crumble when this happens. But if the team still manages to show excellent form, then why shouldn't they develop well?

The same applies for players - you can sign a player in good form, but he is only going to develop if he actually sustains that form. If he does, then you should be rewarded. If he doesn't, you have Carroll and/or D®owning.

Lets look at the flip side as well:

PA 140 ==> average rating 5.50, bad luck with injuries ==> game re-rates PA ==> PA 130

PA 130 ==> average rating 5.50, bad luck with injuries ==> game re-rates PA ==> PA 120

PA 120 ==> average rating 5.50, bad luck with injuries ==> game re-rates PA ==> PA 110

PA 110 ==> average rating 5.50, bad luck with injuries ==> game re-rates PA ==> PA 100

PA 100 ==> average rating 5.50, bad luck with injuries ==> game re-rates PA ==> PA 90

PA 90 ==> average rating 5.50, bad luck with injuries ==> game re-rates PA ==> PA 80

etc

etc

\o/ Double Winner

and thats before we consider the flaws in how players are rated in matches.

EDIT

Torres would be a great example 170odd PA to 200PA to 50PA in two seasons, I think even the researchers do a better job!

The scale of the drops are just an example. And the fact that it is linear is taking my idea a little too far. For example, it could easily be a 5-year average or something. Or the drops could be logarithmic.

What you are really implying is that PA is becoming useless. Which is really my true thoughts - we don't need PA. We just need something that mimic's a player's talents. However, putting this across is understandably controversial and I get bogged down proposing a model that nobody downloads (yes, you can monitor your own Rapidshare downloads, and nobody did in the first few days of me uploading one). Which is why I'm targeting a lighter solution for now, with the knowledge that larger ideas are subsequently less controversial and hence more easy to explain.

[x42bn6]

The players actual potential does not increase though just the researchers view has changed from judging CA.

The actual limit doesn't change - but PA isn't really the limit. It's the researcher's opinion of the limit. So I agree - and the system is still wrong, because the system assumes the researcher's guess is actually the real limit. When it's not - it's a guess!

This is where the visible PA is a problem for me, if that was hidden all you would know is that you have a good player playing well, who, because of his age, has the ability to become better. It should not matter what that figure is just that it is unknown to user and AI. I know a star rating is not a figure and is a bit more vague than that but in principle it is the same thing just represented more loosely. I don't think it is necessary to change the PA just hide it and rate players on CA. Then if you had a 3* CA player of 20 years, old he could be a 5* who continues to improve or a 3* who has peaked early but the key is you wouldn't know.

What if a 120/120 player averages 8.50 and doesn't develop? You'd see that in-game, while mediocre players in rubbish form bizarrely surpass him due to the fact that their PAs are higher.

That is the PA issue.

Take a wonderkid like Neymar, for all we know how good he is now might be almost as good as he ever gets or he may be the next player of Messi's stature. All we should know in game is that he is a star with age on his side and more to learn.

And if you gave him a PA of 200, it would be wrong by the researchers' guidelines, despite the fact that he could well surpass Messi.

More importantly, if he had a PA of 180, say, and he outperforms Messi in-game, then he wouldn't rise to Messi's level. He'd be stuck at CA 180.

Because some random number generator said so (I'm assuming he has a negative PA).

[x42bn6]

I must say, I think everyone has been pretty cordial in this thread so far, good job guys

I really don't like the idea of PA changing depending on ratings. That in itself suggests that the PA of the player was wrong to begin with

But it probably is. This is why PAs go up and down between game versions.

and will undoubtedly be wrong again in the future

But it probably is. This is why PAs go up and down between game versions.

PAs of real players will be wrong. They'll be wrong almost all the time. The researcher has a 1/200 chance of getting the PA 'correct'. I don't understand why this bothers people so much.

Because a 120/120 player averaging 8.50 won't develop, while a 80/150 player will surpass this player as long as his personality is sensible.

Because a player who averages 8.50 consistently over many seasons would realistically be compared to Messi and Ronaldo, and would quite rightly rival them in ability. Yet he will be stuck at CA 120.

It is unlikely that the researchers got a player's CA or individual attributes 'correct' either, especially relative to another player's. It is entirely possible that every single attribute a particular player is accorded is not a true reflection on a 1-20 scale of their ability in real life (especially outside the big leagues). No one seems upset by that.

I am, and anyone who has tried to fiddle with Walcott's physical attributes is. Or two-footedness. Or weird attribute weightings...

[postal postie]

Because a 120/120 player averaging 8.50 won't develop, while a 80/150 player will surpass this player as long as his personality is sensible.

so whats the issue with a player - in real life or in FM - not being able to develop further than their current ability?

a 120/120 player averaging 8.5 a season is not likely to be playing in the top leagues so will not be classed in the same way as Messi.

[JamesB]

This is why PAs go up and down between game versions.

Yes, and that is fine. Almost every rating in the game is 'wrong' if we're holding them to perfect standards. Your rating system would not address this, it would only ensure that the PA changed to a wrong number more frequently. Oh look, Speedy McGee averaged an 8.5 rating season after season because he was playing as a lone poacher with the whole team trying to throughball to him the entire game every game, and he ended up scoring a goal a game (and goals jack your rating ridiculously). He should eventually have a PA of 200. It doesn't matter that he's one dimensional and not particularly talented.

My view of PA is that it is what you refer to as a player's "limit" and, as such, a changing PA makes no sense to me. Is it likely that the researcher 100% accurately predicted a player's PA? No. Will they probably have got it in the ballpark? Yes. If not? It will be changed as new information comes to hand.

Frankly, if you're changing player's attributes then you're fighting a losing battle. There are 100s of thousands in the database, and you're making (biased?) changes to a select few. At some point you have to accept that the game is not a mirror image of real life.

[postal postie]

Yes, and that is fine. Almost every rating in the game is 'wrong' if we're holding them to perfect standards. Your rating system would not address this, it would only ensure that the PA changed to a wrong number more frequently. Oh look, Speedy McGee averaged an 8.5 rating season after season because he was playing as a lone poacher with the whole team trying to throughball to him the entire game every game, and he ended up scoring a goal a game (and goals jack your rating ridiculously). He should eventually have a PA of 200. It doesn't matter that he's one dimensional and not particularly talented.

well put if a player is getting a high average rating but has a *relatively* low CA/PA, then the issue is in the rating of that player and not the CA/PA.

"messing up" the CA/PA so that it works within the framework of a poorly implemented rating system is not the way to go.

[bigdunk]

Nope, there doesn't have to be an explicit limit in the same way that a good racing engine doesn't hard-code the top speeds of its cars.

It actually makes it easier for reseachers since they are never put on the spot to be absolutely correct.

There's still a hard limit though, even if the researcher doesn't explicitly say what it is. If the researcher gives a player some kind of estimated PA and there's a bunch of other factors that can cause him to rise above it (his determination, his training/coaching, whatever) then his estimated PA + the maximum he can gain from each of those other factors is a hard limit on his PA. In other words, if the player was coached perfectly, had the perfect attitude etc he would reach his this hard limit on his PA. Which is exactly the siutation we should have now, but instead of a researcher being able to sit down and say this player could potentially become this good, he has to consider a ton of other factors as well, which just make the whole thing more complicated.

The same applies to a racing sim. You might not have an explicity coded top speed, but you have an implicit one, which can be derived from the cars mass, acceleration, drag, and so on. You might not have a value in the code that says car X's max speed is 200mph, but no matter what you do that car will never exceed that speed because it's physical properties do not allow it to.

[Cougar2010]

What you are really implying is that PA is becoming useless. Which is really my true thoughts - we don't need PA.

Despite not wanting to get into another discussion with you I can't ignore this point.

I am in no way, shape or form implying that PA is useless so please don't words in my mouth.

We just need something that mimic's a player's talents. However, putting this across is understandably controversial and I get bogged down proposing a model that nobody downloads (yes, you can monitor your own Rapidshare downloads, and nobody did in the first few days of me uploading one). Which is why I'm targeting a lighter solution for now, with the knowledge that larger ideas are subsequently less controversial and hence more easy to explain.

Or it could be that most people aren't interested in your proposal.

TBF If FM was being developed from scratch now your more detailed proposal would be worth considering along with the CA/PA system we have now. However because the CA/PA system has been in place and working well for nearly 20 years, your proposal offers very little, if any, tangible benefit over the current system.

You also still seem to have an issue understanding the terms, CA, Max CA & PA and the relationship between them despite having them explained to you literally hundreds of times by everyone from first time posters to SI coders over the last couple of years, especially the concept of the game being god.

Personally I think the CA/PA system is a good one but I would like to see improvements in the following areas:

A) Only a very small % of players should reach their PA within a save (say 1%) even given ideal conditions. It should also be more difficult for the player to improve the closer he is to his PA (This is possibly already included, I'm not sure).

B) Staff estimate PA via stars far too easily, there should be an obvious difference between staff with good, average & bad scouting attributes and even the best staff should be wrong far more often. I would even extend this to maybe getting rid of PA stars altogether and just showing CA stars along with maybe a comment, "Reached limit", "Can improve", "Can improve significantly".

C) More "Fog of War", this has been suggested in the past but an extension would see attributes shown first as arrows, then as a range before being narrowed down to an actual number as you scout a player over time.

[x42bn6]

so whats the issue with a player - in real life or in FM - not being able to develop further than their current ability?

a 120/120 player averaging 8.5 a season is not likely to be playing in the top leagues so will not be classed in the same way as Messi.

Let's assume that this player, however, is playing at a top club (maybe not the best club in the world).

This is possible, depending on things like form and attitude. Unlikely perhaps, but possible.

The first season he hits 8.50, there will be a lot of interest. That's something like a goal and half an assist per game, so a player who hits 40 goals, 20 assists in a season would be of course looked at. Flash in the pan? Nope - he does it again. And again. And again.

A player who gets 40 goals and 20 assists consistently per season while playing in a top league for a top team would be compared with Messi and Ronaldo, full-stop.

Such a player would be learning a lot over time. After the first season, defenders would be trying to destroy him before he waltzes the ball into the net. Yet his ratings stay constant despite the additional defensive scrutiny. So he is learning - so his CA must be rising (in the same way that first-team football for youngsters aids their development and CA).

In other words, a player who is consistently doing well "beyond what is expected of them" in a competitive environment will be learning, and hence their CA must rise in tandem.

This can't happen if they have peaked.

A player who is playing very average might not develop as fast, if at all. They're not "learning" - they are performing as expected. In this case, 120/120 might be appropriate and his CA might be stuck there.

A player who is playing utterly rubbish might even stagnate. In a season, he might drop to 110/120.

But none of these cases cover the case where a player does exceptionally well.

[marty78]

If you had a player averaging 8.5 in a league then does it matter what his CA/PA actually is though? If another player comes along and performs better does it matter what his CA/PA is? Values for CA/PA are not always reflected in performances and you would not know the second one will be better than the first if their PA values were hidden. The only way you'd know is when you'd spent time developing him and he reached a similar level and was still improving.

[postal postie]

Let's assume that this player, however, is playing at a top club (maybe not the best club in the world).

This is possible, depending on things like form and attitude. Unlikely perhaps, but possible.

The first season he hits 8.50, there will be a lot of interest. That's something like a goal and half an assist per game, so a player who hits 40 goals, 20 assists in a season would be of course looked at. Flash in the pan? Nope - he does it again. And again. And again.

A player who gets 40 goals and 20 assists consistently per season while playing in a top league for a top team would be compared with Messi and Ronaldo, full-stop.

Such a player would be learning a lot over time. After the first season, defenders would be trying to destroy him before he waltzes the ball into the net. Yet his ratings stay constant despite the additional defensive scrutiny. So he is learning - so his CA must be rising (in the same way that first-team football for youngsters aids their development and CA).

In other words, a player who is consistently doing well "beyond what is expected of them" in a competitive environment will be learning, and hence their CA must rise in tandem.

This can't happen if they have peaked.

A player who is playing very average might not develop as fast, if at all. They're not "learning" - they are performing as expected. In this case, 120/120 might be appropriate and his CA might be stuck there.

A player who is playing utterly rubbish might even stagnate. In a season, he might drop to 110/120.

But none of these cases cover the case where a player does exceptionally well.

i still believe the issue (if it actually happens, which i dont believe it does - in game) is an over performing, low CA/PA, player. as i say in my previous post, i dont believe the CA/PA should be modified (broken) to accomodate a poor implementation of the ratings system or match engine.

you say 'lets assume' the player is at a top club. your also assuming a lot of other info which i dont believe would happen. gettings 40 goals/20 assists a season (yes i accept you may have been exagerating but even 20goals/20assists a season is alot) for a player with such low CA/PA would not happen in game or real life and, if it does, and i'm repeating my self, then that is the problem and not the implementation of CA/PA

[JamesB]

Such a player would be learning a lot over time. After the first season, defenders would be trying to destroy him before he waltzes the ball into the net. Yet his ratings stay constant despite the additional defensive scrutiny. So he is learning - so his CA must be rising (in the same way that first-team football for youngsters aids their development and CA).

No. If a player with 120/120 was averaging 8.5 over many seasons, that clearly illustrates that what you are saying is not happening. The defenders aren't adjusting and trying to "destroy" him, because he is maintaining his 8.5 rating WITHOUT his attributes having changed. A 120/120 player will have not developed at all over the course of seasons (under the current system, you know that), and yet he is able to maintain his rating. What has he learned? Nothing, that is concrete.

I suppose you could have been talking about a player in real life, and somehow imposing a hypothetical 8.5 rating on them.

[marty78]

Personally I think the CA/PA system is a good one but I would like to see improvements in the following areas:

A) Only a very small % of players should reach their PA within a save (say 1%) even given ideal conditions. It should also be more difficult for the player to improve the closer he is to his PA (This is possibly already included, I'm not sure).

B) Staff estimate PA via stars far too easily, there should be an obvious difference between staff with good, average & bad scouting attributes and even the best staff should be wrong far more often. I would even extend this to maybe getting rid of PA stars altogether and just showing CA stars along with maybe a comment, "Reached limit", "Can improve", "Can improve significantly".

C) More "Fog of War", this has been suggested in the past but an extension would see attributes shown first as arrows, then as a range before being narrowed down to an actual number as you scout a player over time.

That is how I would like to see it improved also. I think more players could receive a higher PA but it should be almost impossible to reach. Staff judging PA - I've already expressed my opinion on this in this thread and the fog of war is a great idea.

The "FM is god" and potential abilty point, I think, ties into you first point about being harder to achieve PA for players. If PA is the absolute best a player can be with perfect conditions he should never reach it as it would be impossible to be perfect, very good, maybe, but perfect, impossible.

[x42bn6]

Yes, and that is fine. Almost every rating in the game is 'wrong' if we're holding them to perfect standards. Your rating system would not address this, it would only ensure that the PA changed to a wrong number more frequently. Oh look, Speedy McGee averaged an 8.5 rating season after season because he was playing as a lone poacher with the whole team trying to throughball to him the entire game every game, and he ended up scoring a goal a game (and goals jack your rating ridiculously). He should eventually have a PA of 200. It doesn't matter that he's one dimensional and not particularly talented.

Not true. If a player does really well, defenses start to tighten up against him and the opposition takes him more seriously.

In an ideal world, too, a one-dimensional player's rating might need to be lowered in-game. But that's really a different story.

My view of PA is that it is what you refer to as a player's "limit" and, as such, a changing PA makes no sense to me. Is it likely that the researcher 100% accurately predicted a player's PA? No. Will they probably have got it in the ballpark? Yes. If not? It will be changed as new information comes to hand.

What if this new information comes into hand in the game itself? i.e. A player is 70/70 but is thrown into the first-team and suddenly the first-team finds it can't live without him as he's so good?

Frankly, if you're changing player's attributes then you're fighting a losing battle. There are 100s of thousands in the database, and you're making (biased?) changes to a select few. At some point you have to accept that the game is not a mirror image of real life.

It's not a mirror image. But it's something the game should strive for. After all, as SI claim, it's a simulation.

Nor do I have an issue with changing the way other attributes are defined, if that is necessary. I'd go one further and state that this should be encouraged!

well put if a player is getting a high average rating but has a *relatively* low CA/PA, then the issue is in the rating of that player and not the CA/PA.

Wouldn't it be cool if the game could realise this and actually re-rate the player, as if it were a researcher?

"messing up" the CA/PA so that it works within the framework of a poorly implemented rating system is not the way to go.

Being too cosy with the idea of a flawed framework like CA/PA is not necessarily any better.

There's still a hard limit though, even if the researcher doesn't explicitly say what it is. If the researcher gives a player some kind of estimated PA and there's a bunch of other factors that can cause him to rise above it (his determination, his training/coaching, whatever) then his estimated PA + the maximum he can gain from each of those other factors is a hard limit on his PA. In other words, if the player was coached perfectly, had the perfect attitude etc he would reach his this hard limit on his PA. Which is exactly the siutation we should have now, but instead of a researcher being able to sit down and say this player could potentially become this good, he has to consider a ton of other factors as well, which just make the whole thing more complicated.

It's actually easier for a researcher since they only need to consider "realistic" scenarios and "take an average". They don't care about the case where a player hits 10.00 every season. They just look at Tom Cleverley and say, "He's probably going to become a solid Premier League player in the future. 'EPA' 155 sounds sensible." Next player. It can be designed to be wholly-independent of things like training facilities. Entirely possible.

The bold bit is nearly what I'm after! I want something where a researcher gives some form of input (an "average"), and the game simply takes this input into consideration (along with many other factors, like coaching facilities and personality) when determining the player's development.

I say "nearly" because I believe that you don't actually care what the limit is. If I need to perform some complex calculation in order to determine the limit, what does that give me? Not much. What really matters is that the average is used in player development - not the complex calculation of the limit.

The same applies to a racing sim. You might not have an explicity coded top speed, but you have an implicit one, which can be derived from the cars mass, acceleration, drag, and so on. You might not have a value in the code that says car X's max speed is 200mph, but no matter what you do that car will never exceed that speed because it's physical properties do not allow it to.

I calculate the maximum speed of a car to be 115.53mph on a track temperature of 35°C and wind speed 3mph at 20.3° away from the vehicle.

What use is this 115.53mph value to me?

Who cares if the limit can be implicitly calculated? I know there's a limit in my simulation, why do I care?

All I care about in my simulation is that if I have a track with track temperature of 35°C and wind speed 3mph at 20.3° away from the vehicle, then its maximum speed is 115.53mph. I can use this as justification on how awesome my simulation is. I don't care about the limit - what good does it do in my simulation? All I care about are the physics associated with my simulation.

Or it could be that most people aren't interested in your proposal.

Or I get the typical arguments that "You can't ignore the fact that there is a limit" (I'm not, just that I don't think we need to store it) or "The game is God, so it is right" (circular reasoning: The game is right... Because it is right). Again and again...

TBF If FM was being developed from scratch now your more detailed proposal would be worth considering along with the CA/PA system we have now. However because the CA/PA system has been in place and working well for nearly 20 years, your proposal offers very little, if any, tangible benefit over the current system.

It offers a small benefit in that 120/120 players can develop if they perform deserving of that development.

A small benefit perhaps, but I don't actually care about the magnitude. What I care about is that it is an improvement and hence should be thought about.

You also still seem to have an issue understanding the terms, CA, Max CA & PA and the relationship between them despite having them explained to you literally hundreds of times by everyone from first time posters to SI coders over the last couple of years, especially the concept of the game being god.

I have an issue with the terms themselves! I understand them perfectly. And I think they are wrong.

The game isn't God. I wish you'd call CA and PA for what it is - an assumption. A simplification. Something that makes SI's job easy. Why worry about how good a player is (comparing 50 attributes) when we have CA (one attribute!) to compare a left-back with a right-winger and realise they are somewhat equal. Why worry about balancing player development when you have an explicit limit.

You cannot use the argument that "the game is God" to justify the game's correctness - that is circular reasoning. I may as well claim that my posts are God, so they are right! If the game is perfect, you should be able to justify it.

Personally I think the CA/PA system is a good one but I would like to see improvements in the following areas:

A) Only a very small % of players should reach their PA within a save (say 1%) even given ideal conditions. It should also be more difficult for the player to improve the closer he is to his PA (This is possibly already included, I'm not sure).

How would you decide which 1% are deserving to reach their PA?

[postal postie]

Wouldn't it be cool if the game could realise this and actually re-rate the player, as if it were a researcher?

no, because that wouldn't be solving the actual problem.

[JamesB]

The more one thinks about your proposed system, the more flawed it becomes in my opinion.

A player's rating is only partially a product of their ability. Perhaps it is even only so to a minor degree. How a manager uses a player has more of an effect on the rating the game assigns them. I have seen screenshots of people using Mr Hough's tactic which has garnered their player an average rating in the 9.xxs over the course of many seasons. Under your system, that player, no matter their real life ability, would have a PA of around 250 in-game: the best there ever was. Perhaps more. That's far more of a fantasy land than a researcher's educated guess.

Conversely, do you not think it's possible for me to play Rooney, for example, out of position and have him average in the low 6.xxs over the course of seasons. I think I could. Is than an indictment on his ability? No, it should be on the manager's. Your system would then have him regress to a mediocre championship player by the time he retires.

I don't think people are shunning your system for any reason other than it is likely to lead to bizarre outcomes.

[x42bn6]

If you had a player averaging 8.5 in a league then does it matter what his CA/PA actually is though?

It shouldn't, but for 120/120 players, it does...

i still believe the issue (if it actually happens, which i dont believe it does - in game) is an over performing, low CA/PA, player. as i say in my previous post, i dont believe the CA/PA should be modified (broken) to accomodate a poor implementation of the ratings system or match engine.

What if the player isn't exploiting the ratings system or match engine? What if the player is simply playing very well?

This is a hypothetical scenario. Assume the player is just a very good player (maybe his attributes are very well-balanced and his mental attributes are through the roof). Assume not match-engine exploits are being made.

you say 'lets assume' the player is at a top club. your also assuming a lot of other info which i dont believe would happen. gettings 40 goals/20 assists a season (yes i accept you may have been exagerating but even 20goals/20assists a season is alot) for a player with such low CA/PA would not happen in game or real life and, if it does, and i'm repeating my self, then that is the problem and not the implementation of CA/PA

This is a rare scenario that doesn't happen a lot. More importantly, it's unexpected (hence the low PA).

Nobody saw, say, Luca Toni late-blooming in the later parts of his career. But he did nevertheless.

Feel free to knock it down to 20 goals and 20 assists. To me, that's a more creative version of Darren Bent at Spurs and I can live with that. The point is that it's still exceeding expectations and a 120/120 player would still need to develop.

No. If a player with 120/120 was averaging 8.5 over many seasons, that clearly illustrates that what you are saying is not happening. The defenders aren't adjusting and trying to "destroy" him, because he is maintaining his 8.5 rating WITHOUT his attributes having changed. A 120/120 player will have not developed at all over the course of seasons (under the current system, you know that), and yet he is able to maintain his rating. What has he learned? Nothing, that is concrete.

I suppose you could have been talking about a player in real life, and somehow imposing a hypothetical 8.5 rating on them.

It's real-life. A real-life player consistently hitting 8.50 every season (where 8.50 is calculated using SI's match-engine, based on passes, shots, goals, assists, fouls, cards, overall team performance, etc.). You could of course calculate it to be 10.00 using the rose-tinted glasses rating engine, but it's just a number. I just picked 8.50 as it makes most sense in the context of Football Manager.

A player who hits 8.50 in his first season may struggle with second-season syndrome as the opposition tries to deal with him better. If he continues to hit 8.50, then he has clearly learnt something! He's become a better player, as he now deals with tougher opposition. So realistically, his attributes must rise (he's become a better player).

Oh - he's hit his PA. Oops.

[x42bn6]

no, because that wouldn't be solving the actual problem.

The issue being the player isn't rated properly. So if the game could re-rate a player, how would that not solve the issue?

The more one thinks about your proposed system, the more flawed it becomes in my opinion.

A player's rating is only partially a product of their ability. Perhaps it is even only so to a minor degree. How a manager uses a player has more of an effect on the rating the game assigns them. I have seen screenshots of people using Mr Hough's tactic which has garnered their player an average rating in the 9.xxs over the course of many seasons. Under your system, that player, no matter their real life ability, would have a PA of around 250 in-game: the best there ever was. Perhaps more. That's far more of a fantasy land than a researcher's educated guess.

I'd rather not use exploit tactics as an example because, well, the issue lies with the match engine, the definition of "exploit tactic"!

Conversely, do you not think it's possible for me to play Rooney, for example, out of position and have him average in the low 6.xxs over the course of seasons. I think I could. Is than an indictment on his ability? No, it should be on the manager's. Your system would then have him regress to a mediocre championship player by the time he retires.

An interesting point but we could solve this by the game finding out if the user is playing him really sub-optimally and thus ensure that any drops are mitigated somewhat.

Having said that, if I stuck Rooney in goal for a whole season and he averaged in the 5.00s, then I would not be surprised if it has a relatively-permanent negative effect on his finishing and creativity, say, justifying an attribute drop.

If he somehow averaged in the 7.00s, on the other hand, maybe we could boost his goalkeeper attributes...

I don't think people are shunning your system for any reason other than it is likely to lead to bizarre outcomes.

If my system is balanced properly, I think it will lead to ~99% of the game being the same as the old system, with 1% being "bizarre", but most importantly, explainable scenarios. If Joe Average becomes the next Lionel Messi, it will be because he has averaged ~8.50 every season for a long time in a top team in a top league. Because - let's face it - 40 goals, 20 assists per season is approaching Lionel Messi levels of brilliance - regardless of what some researcher thought. And most importantly - a manager is responsible for it - and that manager could be you.

If someone exploits the match engine to do it, then the match engine is the issue, not this system.

[JamesB]

It's real-life. A real-life player consistently hitting 8.50 every season (where 8.50 is calculated using SI's match-engine, based on passes, shots, goals, assists, fouls, cards, overall team performance, etc.). You could of course calculate it to be 10.00 using the rose-tinted glasses rating engine, but it's just a number. I just picked 8.50 as it makes most sense in the context of Football Manager.

A player who hits 8.50 in his first season may struggle with second-season syndrome as the opposition tries to deal with him better. If he continues to hit 8.50, then he has clearly learnt something! He's become a better player, as he now deals with tougher opposition. So realistically, his attributes must rise (he's become a better player).

Oh - he's hit his PA. Oops.

So his PA was wrong? It will be changed in the next update, or you could do it yourself in the editor - as you said you already do.

The problem is that FM isn't real life. In FM, you can see strikers have 150 goal seasons. This is in no way a reflection of their real life ability (which I gather you want the game to reflect), but you would have the game bump up their PA, leading to even further ridiculousness.

[TheBetterHalf]

So if I understand this correctly, the main issue people are having with the PA being rigid is that once you have visited the forum and read à few threads, having followed football in lets say England and know what player are considered to be prospects and having already played 2 saves Of each 6 seasons, you know( without using scouts or cheating tools) that in your newly started save with Hudds, there is no player with à chance Of becomming your leading premiership player in the future.

Is this the problem, the actuall knowledge Of it, that you know that no researcher have set à -10 PA Of any player in league One or in the Swedish league as this player would be in tabloids in real life for you to read about if it had been the case?

[milnerpoint]

A) Only a very small % of players should reach their PA within a save (say 1%) even given ideal conditions. It should also be more difficult for the player to improve the closer he is to his PA (This is possibly already included, I'm not sure).

B) Staff estimate PA via stars far too easily, there should be an obvious difference between staff with good, average & bad scouting attributes and even the best staff should be wrong far more often. I would even extend this to maybe getting rid of PA stars altogether and just showing CA stars along with maybe a comment, "Reached limit", "Can improve", "Can improve significantly".

C) More "Fog of War", this has been suggested in the past but an extension would see attributes shown first as arrows, then as a range before being narrowed down to an actual number as you scout a player over time.

That hits the nail on the head really. It needs to be harder for both players to reach their potential and for the user to know what this potential is, removing PA is not the solution, hiding how much we know about the PA is.

[x42bn6]

So his PA was wrong? It will be changed in the next update, or you could do it yourself in the editor - as you said you already do.

1) That does not help my current running save. I guess I could boost his PA using FMRTE, but it is of course better if the game did it (and more importantly, applied the same fair treatment to all players).

2) What if the player's boost is purely down to what I am doing in my game? It doesn't make sense to change it in the editor. The editor contains opinions. My opinion of that player might not have changed.

As an example - say a manager plays some player out of position and I am smart enough to realise that his attributes suggest he has a better position. However, this player has played in the wrong position for so long he has dulled his attributes.

That opinion isn't wrong. But I can fix that virtual reality in my game.

The problem is that FM isn't real life. In FM, you can see strikers have 150 goal seasons. This is in no way a reflection of their real life ability (which I gather you want the game to reflect), but you would have the game bump up their PA, leading to even further ridiculousness.

I'm not suggesting we just take the brakes off and let the game go with my solution. Any solution has to actually be fixed with the match engine, and undergo its own balancing. I believe that if my solution were balanced, it would be better.

[x42bn6]

So if I understand this correctly, the main issue people are having with the PA being rigid is that once you have visited the forum and read à few threads, having followed football in lets say England and know what player are considered to be prospects and having already played 2 saves Of each 6 seasons, you know( without using scouts or cheating tools) that in your newly started save with Hudds, there is no player with à chance Of becomming your leading premiership player in the future.

Is this the problem, the actuall knowledge Of it, that you know that no researcher have set à -10 PA Of any player in league One or in the Swedish league as this player would be in tabloids in real life for you to read about if it had been the case?

Nothing to do with that. I suggest you read the thread properly.

My issue is that PA is an opinion that can be underestimated but the game doesn't make up for this possible underestimation.

That hits the nail on the head really. It needs to be harder for both players to reach their potential and for the user to know what this potential is, removing PA is not the solution, hiding how much we know about the PA is.

We can do both. Remove PA and hide what differentiates a talented player from an untalented one in an appropriate way...

[JamesB]

My issue is that PA is an opinion that can be underestimated but the game doesn't make up for this possible underestimation.

I guess this is one reason that I'm not as enthused about your system as you are. This just doesn't bother me. The game world becomes my reality even if, in the future, it is not in fact reality.

Thank you for taking the time to explain your system to me.

[TheBetterHalf]

Nothing to do with that. I suggest you read the thread properly.

My issue is that PA is an opinion that can be underestimated but the game doesn't make up for this possible underestimation.

We can do both. Remove PA and hide what differentiates a talented player from an untalented one in an appropriate way...

I have read the thread and was wondering if the real problem the Topic starter and people in common( not you as such) have with à PA set by SI was this.

[Cougar2010]

Nothing to do with that. I suggest you read the thread properly.

My issue is that PA is an opinion that can be underestimated but the game doesn't make up for this possible underestimation.

This only happens with real players, its a total non-issue with newgens.

and in reality its a very minor issue as even though the game progresses into the future it is still currently 2012 in real life. The players we consider stars now are stars within FM, the players we think will be stars of the future will be stars in the future as we progress through a save and those players that are "Late Bloomers" who will get noticed in the next few years we don't really care about as we don't know anything about them at present.

[bigdunk]

It's actually easier for a researcher since they only need to consider "realistic" scenarios and "take an average". They don't care about the case where a player hits 10.00 every season. They just look at Tom Cleverley and say, "He's probably going to become a solid Premier League player in the future. 'EPA' 155 sounds sensible." Next player. It can be designed to be wholly-independent of things like training facilities. Entirely possible.

The bold bit is nearly what I'm after! I want something where a researcher gives some form of input (an "average"), and the game simply takes this input into consideration (along with many other factors, like coaching facilities and personality) when determining the player's development.

It's okay plucking values out of the air for Cleverly, he's already at a big club with good facilities etc. It's a hell of a lot harder doing it for some random young player in league 2 as there are so many variables that could potentially affect him. You are making a huge guess at his EPA because you are also guessing at his likely career trajectory (i.e. whether a bigger club buys him, plays him etc). In fact it's a much bigger guess that just guessing at a players actual PA which the researchers do now - actual PA factors as much out as possible by assuming that a player gets perfect training etc to give an absolute maximum and then determines whether a player actually reaches that based on his other factors.

So let's say you have a 20 year old playing fairly regularly in league 2, and doing fairly well - what EPA should he have? He has a chance of making the prem, but he probably won't. So do you give him the EPA of a league 2 journeyman and allow the other factors to massively affect his actual PA (so he could potentially become a prem star), which is nonsense, since anybody could potentially become a prem star if you allow this. Or you give him a prem level PA and make sure he doesn't reach it often, which is exactly what we do now so nothings changed.

And I doubt that it's easier for researchers. If you give Cleverly an EPA of 155, but give him a very high determination, he's almost always going to exceed that EPA, so you have the stupid situation of a researcher saying "I think Cleverly will be EPA 155, but I'm going to have to give him EPA 140 since he's always going to exceed what I give him" or vice versa, you have a researcher giving a player an EPA which he knows will be almost always ignored.

I say "nearly" because I believe that you don't actually care what the limit is. If I need to perform some complex calculation in order to determine the limit, what does that give me? Not much. What really matters is that the average is used in player development - not the complex calculation of the limit.

All I'm saying is, you either have a development model that is based on an average + factors that wil let the player exceed the average, or you a model that is based on a maximum with factors that determine whether you reach that maximum or not. Either way the outcome is exactly the same. In either model the player will ultimately reach a point where he can develop no further (either because he has reached his max PA, or because he has reached the limit at which the other factors will allow him to exceed his average PA). Of both systems, the Max PA one is much easier to use for researchers and for database balancing and it makes much more sense to stick to it.

[Cougar2010]

@ x42

Also how do you propose your system would deal with the millions of players who have a reputation of being a good/star player in the mid-lower leagues but struggle when promoted/transferred into higher leagues?

As far as I can see with your system the player would stand out in say League 1, get an improved PA as a result which would lead to a better performance in the Championship whereas in real life the player struggles in the Championship but plays well again when dropping back to League 1.

I also want to point out in general that far too much emphasis is placed on CA/PA rather than looking at attributes.

This last season I've had three DLs:

A) 31yo rated at 3*

B) 21yo rated at 2.5* with 3.5* potential

C) 34yo rated at 1.5*

Despite the difference in stars the best performer of the three has been the 34yo this season and if you looked at the relevant attributes for the position there actually isn't that much difference between any of the three.

[Mittens2317]

I honestly don't ind the PA system, it's a very realistic system that is pretty much based on the opinion of the Assistant Manager or whatever coach you're asking for a report from. At the end of the day, I want to know the opinion of my backroom staff. I want them to give me a little nudge every now and again, and I want to know what they think of the squad I'm building.

If I have a youth player eligible for a professional contract, I'm going to want my Youth Coach to tell me if they think that player's good enough to offer a contract, because I certainly don't want to be paying wages to one of those players who retires the season after their contract's expired because he simply wasn't good enough to be picked up by another team.

This system saves money, time, effort, and makes your backroom staff an invaluable asset. It also makes choosing your backroom staff much more tactical. Do you want an Assistant Manager who can manage the team if you're on holiday/got a touchline ban, or would you prefer one that can analyse your squad and give you the ol' nudge nudge wink wink when a player's past his peak?

Also, the PA system doesn't always work out the way you think. I splashed out £5m on a youngster with a four star PA rating after almost a month's worth of being scouted, and now after being at the club for a few years, he's two star. On the other hand, I had a youth player who had a one and a half star rating (the minimum, imo, for offering a professional contract), and now he's two star CA and three star PA.

At the end of the day, it's all down to the stats of your backroom staff. Yes, I know that a PA system out 200 is auto-generated in the database, but you don't know what that rating is for newly generated players. That's why attribute masking is one of the most simple, but innovative features that was added to the game, because you simply don't know. And yes, it is a fact that some players (irl) will be severely limited in potential ability. This could be down to getting into football late, or not being naturally fit, or just flat out not good enough.

To say that the PA system is unreallistic is absolute bollocks, and if you don't like it, play an earlier version of the game.

[obinna]

If he is scoring 50 goals with finishing 13, why does it need to increase? How many more goals does the player need to score? What would be the benefit of a player with finishing 13, and that kind of scoring record having his attributes increase further?

The increase should be gradual not sudden, just like any other progressing player. There are no players with Demba Ba's or Clint Dempsey's career path in FM. This is the problem and I really don't see why its so hard for people to understand.

[whilewolf]

Because a 120/120 player averaging 8.50 won't develop, while a 80/150 player will surpass this player as long as his personality is sensible.

Because a player who averages 8.50 consistently over many seasons would realistically be compared to Messi and Ronaldo, and would quite rightly rival them in ability. Yet he will be stuck at CA 120.

If a top flight player is bossing games and scoring more then a goal a game but FM doesn't treat him as a equal to Messi and Ronaldo that is a problem with the way FM uses PA\CA in determining value and reputation.

If a player scores more then a goal a game and creates half as many again then he does rival Messi and Ronaldo if you can't reconcile his attributes with his performances that is a problem with the match engine.

Under the present system when Kevin Phillips improbably performed beyond expectations and scored 30 goals in 99/00 the researcher would have amended his attributes and CA to a number that would make that possible. Under your proposed system the AI would increase his attributes and CA to a level where he could score 35 goals the next season.

Your proposing a system where good performance leads to improved ability but in any sporting activity I have been involved in its always been the case that improvements in fitness, skill, mentality or understanding happen before or contemporaneously with improvements in competitive performance.

[obinna]

This only happens with real players, its a total non-issue with newgens.

and in reality its a very minor issue as even though the game progresses into the future it is still currently 2012 in real life. The players we consider stars now are stars within FM, the players we think will be stars of the future will be stars in the future as we progress through a save and those players that are "Late Bloomers" who will get noticed in the next few years we don't really care about as we don't know anything about them at present.

This is false. It happens with newgens as well. They have a PA just like the rest of the players in the game. I once sold a 22 year old, in-form, French national team striker b/c it was obvious he reached his peak. How did I know this? His ratings dropped across the board (he wasn't injured) which told me he hit is PA limit. No matter how well he played (he was top scorer in European championship and won the CL) I knew he wouldn't improve. This is pretty silly and wouldn't happen in the real world.

It has very little to do with real players and newgens.

[Äktsjon Männ]

Has any player ever reached their maximum potential? Who do we see as the ultimate 200 PA player? Let's say it's Diego Maradona at his peak for arguments sake. What if a 16 year old Diego Maradona got picked up by Barcelona's youth academy today? Don't you think there is a chance he could have been an even better player if he got the opportunities that players get in this day and age as opposed to the 1970's and 80's?

That's one problem I have with the fixed PA system. Players reach their PA rather often in FM which stops their development process right there and then. The system also says there will pretty much never be a better player than Messi/Ronaldo. If there needs to be an artificially determined upper limit on ability then only about 0.00000000001% of the players should ever reach it.

[obinna]

If a top flight player is bossing games and scoring more then a goal a game but FM doesn't treat him as a equal to Messi and Ronaldo that is a problem with the way FM uses PA\CA in determining value and reputation.

Agreed.

If a player scores more then a goal a game and creates half as many again then he does rival Messi and Ronaldo if you can't reconcile his attributes with his performances that is a problem with the match engine.

Disagree. Players exceeded and fall short of expectations all the time. Attributes should reflect this.

Under the present system when Kevin Phillips improbably performed beyond expectations and scored 30 goals in 99/00 the researcher would have amended his attributes and CA to a number that would make that possible. Under your proposed system the AI would increase his attributes and CA to a level where he could score 35 goals the next season.

His attributes don't need to correlate precisely but they should increase somewhat. For example, if his finishing and composure was 13 and 12 respectively I think a jump to 15 and 13 (or something along those lines) would be expected, not 17 and 16. Under the current system, if he's reached his PA this has 0 chance of happening.

[milnerpoint]

The increase should be gradual not sudden, just like any other progressing player. There are no players with Demba Ba's or Clint Dempsey's career path in FM. This is the problem and I really don't see why its so hard for people to understand.

You didnt answer my questions, Why, if he has that kind of scoring record already, does he need to improve? What would be the benefit of the worlds top goal scorer improving?

[milnerpoint]

This is false. It happens with newgens as well. They have a PA just like the rest of the players in the game. I once sold a 22 year old, in-form, French national team striker b/c it was obvious he reached his peak. How did I know this? His ratings dropped across the board (he wasn't injured) which told me he hit is PA limit. No matter how well he played (he was top scorer in European championship and won the CL) I knew he wouldn't improve. This is pretty silly and wouldn't happen in the real world.

It has very little to do with real players and newgens.

Again how much better can you get as a striker, than one of the best in Europe? What is the big obsession with a player constantly improving? At what point would said player stop improving? After a 200 goal season? If he is a french national striker he is going to be very good, what are you expecting from players in FM, that eventually they can all become like Messi?

[obinna]

You didnt answer my questions, Why, if he has that kind of scoring record already, does he need to improve? What would be the benefit of the worlds top goal scorer improving?

So that his attributes reflect that he exceeded expectations. The same way Torres' ratings dropped b/c he didn't meet expectations. Unless you think ratings should stay the same? No offense but that question is little silly.

During the course of the season, the worlds top scorer should definitely improve. It happens in real life and every time a new version of FM is released, so why is it so crazy that it can't happen in the game world? Once again, if it's gradual there shouldn't be a problem. He wouldn't just turn into a world class striker overnight, nor would the attribute increase be unrealistic.

[obinna]

Again how much better can you get as a striker, than one of the best in Europe? What is the big obsession with a player constantly improving? At what point would said player stop improving? After a 200 goal season? If he is a french national striker he is going to be very good, what are you expecting from players in FM, that eventually they can all become like Messi?

Please stop with the straw man arguments. NO, WE ARE NOT SAYING EVERY PLAYER SHOULD HAVE THE ABILITY TO BE MESSI. The main point is the relationship between CA and PA shouldn't be so rigid. You guys keep dumbing down this debate with the same baseless accusation.

This striker's ratings were good but not great. His production exceeded his ratings for a while but his ratings either dropped or hovered around the same level b/c of his PA. The same way his rating would drop if he has a bad season, they should increase if he plays well.