Potential Ability debate

SuperJackLester · January 10, 2013

I beleive there should be two abilities running alongside each other. A talent based ability (current / potential) and a form based ability (a calculation of current form, season average rating, first team games etc). You could then divide the two and have a "present" ability rating. I think would help reproduce RL "yo-yo" players who play for Prem teams, have a drop in form, sign for a L1/L2 team then work their way back up to leagues. It would also help the amazing non-league players rise up the leagues that we are seeing more and more frequently these days.

Obviously the form based ability would have to be weighted with which leagues they're playing in.

Dave C · January 10, 2013

PA is only realistic if you can actually know the future, when you can't.
It's a prediction, and should be treated as such (i.e. it can be wrong in both directions).

Of course, such a vague number doesn't really serve a purpose if it is vague (it is a number that the player "may or may not reach", depending on what he does in-game - how useful!), which is why you can in theory just get rid of it.

The real solution is a dynamic training model, which will show up PA anyway, because what matters is the many deltas of development over time, not some pie-in-the-sky prediction of a very vague and unfalsifiable limit.

But every single piece of attribute data the researchers put in is 'just a prediction'. Does that mean we should throw it all out?

Of course not, it's a game/simulation. At some points you have to use the appropriate method to make it work, and PA is the most appropriate method.

x42bn6 · January 10, 2013

But every single piece of attribute data the researchers put in is 'just a prediction'. Does that mean we should throw it all out?
Of course not, it's a game/simulation. At some points you have to use the appropriate method to make it work, and PA is the most appropriate method.

"Today's" attributes are required, because we know about them "today", and can possibly even measure them objectively. There is, of course, a degree of uncertainty about the measurement of these attributes, but the uncertainty about the attributes "today" are much, much less than the uncertainty about the overall player 10 years in the future. So you cannot just get rid of today's attributes. A more accurate comparison with PA would be something like "potential passing", similar to PA but for passing: The game could get rid of that if it existed, because it is not "today", it's "the future".

In addition, because the attributes work on a 1-20 scale in terms of researcher data input, it is a lot less difficult to get it wrong compared with a 1-200 scale.

"Potential" has the second issue that it is unfalsifiable: We don't even know if, say, Zidane ever reached his potential. There are two schools of thought: Zidane could do no better, or Zidane could have been so much better than he was. We don't know which is true, and may never know. So what would Zidane's potential be if we went back in time to when he was 16? Would be be 185 (or whatever his peak was in FM), or even higher? You simply don't know.

The most appropriate method would be to not use such a vague and unfalsifiable measure in the first place. The game needs to rely on objective measures, things we know "today" (with uncertainty).

Now, of course, there may be merit in some form of prediction and using this prediction in-game to a certain extent. However, this prediction could be wrong (in both directions), and if the game progress can influence this prediction, then it should. So, if SI decide they want to use a researcher's prediction of the player's future potential in some way, then that could be fine: However, the game should allow for in-game progress to break that prediction if it is underestimated (i.e. Luca Toni, late-bloomer, 170/130), or fall short of that prediction if it is overestimated (i.e. Nii Lamptey, flop, 80/130).

One easy way to shift that is to let researchers pick "average peak" rather than "absolute maximum peak". "Average peak" will naturally be lower than "absolute maximum peak", is acceptable in terms of semantics to be overestimated and underestimated, and indeed, could be easier to think about because averages are easier to think about (for one thing, the distribution of all averages will be similar to the distribution of the recent historic peaks of players, which is a function of the club's youth and training facilities).

So this way, the Manchester United researcher could mark Tom Cleverley to have a future "solid squad player at top-level side" level for "average peak", and depending on how Cleverley does in-game, he could either turn out to be absolutely key for Manchester United, or crash down the tiers like van der Meyde.

It is just that my suspicion is that this is really not necessary because Cleverley should be able to have both these outcomes, regardless of whatever the opinion of the Manchester United researcher is, because what really matters is how Cleverley develops, not the thoughts of one person. But maybe that can be a more comfortable interim step if removing PA is too drastic.

lonestar190 · January 10, 2013

PA is only realistic if you can actually know the future, when you can't.
It's a prediction, and should be treated as such (i.e. it can be wrong in both directions).

Of course, such a vague number doesn't really serve a purpose if it is vague (it is a number that the player "may or may not reach", depending on what he does in-game - how useful!), which is why you can in theory just get rid of it.

The real solution is a dynamic training model, which will show up PA anyway, because what matters is the many deltas of development over time, not some pie-in-the-sky prediction of a very vague and unfalsifiable limit.

No, PA is realistic because it exists in real life.

However, it cannot be known for sure so researchers have to guess at it.

You want a game that mirrors the future, which will never happen no matter what is done with the training model. Even if PA was gone, most players would still not mirror their real life growth, and you would still be annoyed.

sirgiorgio · January 10, 2013

I think PA is neither realistic nor unrealistic, it's just useless. I copy what I wrote in another thread:

1) 16 years old, CA 10, Professionalism + Ambition + Pressure 20 + 20 + 20. He plays in Conference. He knows ManU scouts came to see his match (maybe they want to scout another presumed raw gem). He gives everything he got, ManU scouts are impressed and sign him. So here he is, in one of the best clubs in the world. He's willing to learn and wants to make it, but unfortunately by the time he's 18 his CA is "just" 85. So they sent him in League 2, but his will is strong, he trains hard, finds a good tutor and plays competitive matches. He'll never be a star, but the time he's 24-25 he'll be a good championship player.

2) 16 years old, CA 110. Every club in the world is interested in signing him. Professionalism + Ambition + Pressure 20 + 20 + 20. He makes the right choices (which might as well be to stay where he is and see if he can make it) and becomes the new Messi.

3) 16 years old, CA 110. Every club in the world is interested in signing him. Professionalism + Ambition + Pressure 1 + 1 + 1. He makes the wrong choices and will be the biggest waste of talent in the history of football.

Of course is too simplistic, but you got the idea: it's CA that "defines" PA, while the "right" attributes (and training facilities, and games played, and injuries...) determine a player's development. There's no need in guessing a player's PA, everything is already there. You see Balotelli and think: "This guy could be better than Messi, but he's wasting his talent", you see another player and think: "This guy will never be better than Messi, but he works hard, is willing to learn and has a good attitude". And so on.

x42bn6 · January 10, 2013

No, PA is realistic because it exists in real life.
However, it cannot be known for sure so researchers have to guess at it.

You want a game that mirrors the future, which will never happen no matter what is done with the training model. Even if PA was gone, most players would still not mirror their real life growth, and you would still be annoyed.

I don't want the game to mirror real-life growth exactly. The game world and in-game world diverge the moment the game begins. It would be boring if your actions could not influence the game, after all.

However, the full range of possibilities available in real-life should be possible in-game. This is not the case because the only fair estimation of probability of any player late-blooming is clearly greater than zero (because you cannot tell the future), yet a player who is 120/120 cannot late-bloom in-game. Because of PA.

PA is a very vague and unfalsifiable metric so there are a whole host of reasons why it should not be in-game. But even if it were, there are some characteristics it should have, notably the ability to be underestimated and the game compensating for it - it is a prediction, and it should be treated in that way. However, the game doesn't let PA be underestimated.

PA is not "the limit" - it's a guess of "the limit". So PA should be weakened. Weakened slightly, or to the extent where it should be gotten rid off? It doesn't really matter - it's a not black and white decision. I just think that it can be removed safely with balance.

Spart · January 10, 2013

One easy way to shift that is to let researchers pick "average peak" rather than "absolute maximum peak". "Average peak" will naturally be lower than "absolute maximum peak", is acceptable in terms of semantics to be overestimated and underestimated, and indeed, could be easier to think about because averages are easier to think about (for one thing, the distribution of all averages will be similar to the distribution of the recent historic peaks of players, which is a function of the club's youth and training facilities).
So this way, the Manchester United researcher could mark Tom Cleverley to have a future "solid squad player at top-level side" level for "average peak", and depending on how Cleverley does in-game, he could either turn out to be absolutely key for Manchester United, or crash down the tiers like van der Meyde.

It is just that my suspicion is that this is really not necessary because Cleverley should be able to have both these outcomes, regardless of whatever the opinion of the Manchester United researcher is, because what really matters is how Cleverley develops, not the thoughts of one person. But maybe that can be a more comfortable interim step if removing PA is too drastic.

...so create more powerful and varied development models and add an extra 20 points or something to all players PAs?

The player's development path, rather than PA, would then be shaped by form, club facilities, coaches, tutoring, attitude, injuries, game time etc. etc.

Then, with higher PAs and a more complex array of development paths, a "Late Developer" status could be implemented to kick in for some players after triggering specific criteria, boosting their development post-24 in a similar way to how "Wonderkid" accelerates development for youngsters.

x42bn6 · January 10, 2013

...so create more powerful and varied development models and add an extra 20 points or something to all players PAs?
The player's development path, rather than PA, would then be shaped by form, club facilities, coaches, tutoring, attitude, injuries, game time etc. etc.

Then, with higher PAs and a more complex array of development paths, a "Late Developer" status could be implemented to kick in for some players after triggering specific criteria, boosting their development post-24 in a similar way to how "Wonderkid" accelerates development for youngsters.

20 PA is arbitrary. There is no 100% confidence interval, so in some extreme, extreme cases, 20 PA might still not be enough.

A "late developer" status is very FIFA Manager-like, like an arcade. Players aren't late-bloomers or not - it's not black and white - you can have "small" late-bloomers. Perhaps like Charlie Adam at Blackpool, to Liverpool's detriment. Or other one-season wonders.

There is no number that is really "enough", aside from 200 PA, really.

Another way of putting it is that every player has 200 PA, but every player "fails" (at different degrees) at reaching it. Some will fail harder than others, and that will be down to (the lack of) training, match experience, fitness, etc.

YKW · January 11, 2013

I don't really agree with that. If we break the attributes of a footballer into three categories, Physical, Technical, and Mental, then the one that sets the absolute best players apart is the mental side. That's the hardest to improve. But the easiest to improve, the easiest to be good at, is the physical side. I've watched lower league football the best part of thirty years. I've seen some absolute beasts, real man mountains. But they have no technique, or great technique but they are bottlers. I've seen players as quick as any Premier League star, but without the technique, decision-making, and composure to make it truly valuable.
So I would say it's mentality that provides the greatest limitation, then technique, then physique.

You can feel otherwise, but things like hand-eye (or foot-eye) coordination have been shown to have a genetic component, therefore provide a real limit on technical potential.

I'm with you most of the way here, I won't deny genetic limitations at all. But to me, it seems as though if any of the three categories you named was going to have clear genetic limits, it would be the physical side. So i don't disagree with you about which is the easiest to improve or be good at, but if we were talking about which would have the clearest limits then I would say it would have to be the physical side. If you were to go around the world and look at all the people who could do kick up tricks with a football (a technical skill) and compare them to all the people who could run the 100m above a certain time (a physical skill), I would speculate that there would be a much greater variance in the people who could do the kick ups compared to running the 100m. For example you could find a 13 year old girl or 60 year old man who had some awesome tricks, but neither of these people is likely to be in the 100m group.

My point is that anyone in the world has a much greater chance of mastering any particular technical skill than any particular physical skill. If you were to take a random person off the street and train them at/into something, you would be much more likely to be successful if what you are trying to teach them is a skill like juggling, as opposed to the long jump, for example. Not to say you won't meet some people who show no signs of ever being able to juggle, or that you won't uncover some natural long-jumpers, but your chances will be better as learning to juggle has less (but not no) genetic limiting factors.

I entirely agree about mentality providing the greatest limitation. For me this is because technique and physicals (and their improvement or degradation) depend heavily on the mentals, but not the other way around.

We seem to agree on dividing player's ability into technical, mental and physical attributes. We also seem to agree that each of these areas is harder/easier to improve depending on the individual in question. I know you have already dismissed it, but I would love it if you would reconsider the concept of split PA for each of the areas, I think we could come up with something good along these lines.

Meh we have this every year, the usual multi-quoting and going round in circles.

You forgot one: Annoying members who complain about a thread they don't like instead of just ignoring it, and show no intention whatsoever of adding to the discussion. You should think about changing your username, missthepoint would be much more appropriate.

Dave C · January 11, 2013

I'm with you most of the way here, I won't deny genetic limitations at all. But to me, it seems as though if any of the three categories you named was going to have clear genetic limits, it would be the physical side. So i don't disagree with you about which is the easiest to improve or be good at, but if we were talking about which would have the clearest limits then I would say it would have to be the physical side. If you were to go around the world and look at all the people who could do kick up tricks with a football (a technical skill) and compare them to all the people who could run the 100m above a certain time (a physical skill), I would speculate that there would be a much greater variance in the people who could do the kick ups compared to running the 100m. For example you could find a 13 year old girl or 60 year old man who had some awesome tricks, but neither of these people is likely to be in the 100m group.

I'm not saying there's no physical limitation, just that of those three groupings, the physical side is the 'easiest' to improve, relatively speaking. And this is easily demonstrable by looking at real-world football (as opposed to using lots of non-football illustrations). Go to a non-league match. You'll see plenty of guys who are big, strong, quick. You'll see fewer guys with great technique. And even fewer with great reading of the game.

My point is that anyone in the world has a much greater chance of mastering any particular technical skill than any particular physical skill. If you were to take a random person off the street and train them at/into something, you would be much more likely to be successful if what you are trying to teach them is a skill like juggling, as opposed to the long jump, for example. Not to say you won't meet some people who show no signs of ever being able to juggle, or that you won't uncover some natural long-jumpers, but your chances will be better as learning to juggle has less (but not no) genetic limiting factors.

I don't agree. I think, for example, anyone (outside of extreme cases) could be trained to improve their stamina to an extent they could run a marathon (I did it, anyone can). But I don't agree that anyone could learn to juggle. You mention long jump, but the real problem there is the technique as opposed to the physique. Plenty of pwerful guys and girls still wouldn't do well there, as the coordination/technique will let them down.

Dave C · January 11, 2013

20 PA is arbitrary. There is no 100% confidence interval, so in some extreme, extreme cases, 20 PA might still not be enough.
A "late developer" status is very FIFA Manager-like, like an arcade. Players aren't late-bloomers or not - it's not black and white - you can have "small" late-bloomers. Perhaps like Charlie Adam at Blackpool, to Liverpool's detriment. Or other one-season wonders.

There is no number that is really "enough", aside from 200 PA, really.

Another way of putting it is that every player has 200 PA, but every player "fails" (at different degrees) at reaching it. Some will fail harder than others, and that will be down to (the lack of) training, match experience, fitness, etc.

But every player *doesn't* have 200PA, not in real terms. If you can't accept that basic fact, there's no progress to be had in the discussion. The idea it's only luck/location/attitude that holds some players back is utterly wrong.

milnerpoint · January 11, 2013

You forgot one: Annoying members who complain about a thread they don't like instead of just ignoring it, and show no intention whatsoever of adding to the discussion. You should think about changing your username, missthepoint would be much more appropriate.

I was heavily involved in the last 3 threads on the subject, when all the exact same points, by mostly the same posters were made. This hasnt moved forward in three years, its the same tired arguments over and over involving Luca Toni and Pedro and no one has yet come up with anything new in a long time.

Good word play there at the end tho, im sure you got a wee chuckle to yourself over it.

x42bn6 · January 11, 2013

But every player *doesn't* have 200PA, not in real terms. If you can't accept that basic fact, there's no progress to be had in the discussion. The idea it's only luck/location/attitude that holds some players back is utterly wrong.

They might not have 200 PA, but effectively, there is no other value for PA that can be given with absolute certainty, which the hard-cap nature of PA implies.

Put another way: There will definitely be some form of weather tomorrow, but that does not mean SI need to hardcode the game such that the 12th of January, 2013, has a particular type of weather. The most sensible way of dealing with this is to let "everything" be possible (to a particular probability distribution) and let the game decide when a game hits the 12th of January, 2013.

It's the same thing here: Even if we did have a PA, there is no need to hardcode it, but let the game (and user, via user influences) "decide" it when the time comes. What then matters is the sum of all development up to that point being accurate and reasonable, and the engine being accurate and reasonable - which will produce a reasonable result, given the sum of all development up to that point.

What is clear is that pure genetics is not enough as an argument, so why would it be a hard limit? Opportunities play a huge, huge part in most things - which month you are born plays a role through when sports seasons begin; where you are when a scout visits your youth club; who you know in terms of footballing contacts; injuries... In a lot of ways, hand-eye coordination and sprinting speed are pretty small issues compared with the vast majority of things that are out of your control to a large extent. The book Outliers goes into way more detail, more of less debunking the idea of "destiny" or "fate" and some notion of a limit that is down to "natural talent".

x42bn6 · January 11, 2013

I was heavily involved in the last 3 threads on the subject, when all the exact same points, by mostly the same posters were made. This hasnt moved forward in three years, its the same tired arguments over and over involving Luca Toni and Pedro and no one has yet come up with anything new in a long time.
Good word play there at the end tho, im sure you got a wee chuckle to yourself over it.

The reason why nothing new has come up is because the Luca Toni and Pedro examples all end up being defended with the same old silly reason: The game was wrong at that point ("The researchers should have given him a higher PA!").

But more importantly, the game was wrong, and today's game will be wrong in a few years time, so the game is... always going to be wrong. Now imagine a game where the engine could compensate for human error - it's probably still going to be wrong, but it's obvious that it will be less wrong than before. Wouldn't that be better?

It just so happens that to compensate for human error, you may need to relax some of the assumptions, or perhaps even just get rid of them. There's nothing wrong with exploring those possibilities, especially when that assumption itself is one of the key problems (PA being vague, unfalsifiable, unknown and unmeasurable). Maybe PA does exist. But it's a bit pointless when nobody can accurately define or measure it, and is always going to be wrong without hindsight. It sounds like a pretty bad thing to put in a simulation, quite frankly.

kingpug · January 11, 2013

If you get an 18 year old playing in your first team and has 3 great seasons and reaches his PA he never improves even if his PA is something low like 130, why would someone so young and playing so well suddenly hit a brick wall?

Dave C · January 11, 2013

They might not have 200 PA, but effectively, there is no other value for PA that can be given with absolute certainty, which the hard-cap nature of PA implies.
Put another way: There will definitely be some form of weather tomorrow, but that does not mean SI need to hardcode the game such that the 12th of January, 2013, has a particular type of weather. The most sensible way of dealing with this is to let "everything" be possible (to a particular probability distribution) and let the game decide when a game hits the 12th of January, 2013.

If I could apply one rule to these forums, it would be "don't try draw comparisons from something other than football, to football"

It's the same thing here: Even if we did have a PA, there is no need to hardcode it, but let the game (and user, via user influences) "decide" it when the time comes. What then matters is the sum of all development up to that point being accurate and reasonable, and the engine being accurate and reasonable - which will produce a reasonable result, given the sum of all development up to that point.

It's not the same thing here. There's an entire branch of the football world dedicated to assessing and estimating potential. it's called 'scouting'. it exists because everyone working in football believes, knows, that there is a hard cap on a players future development. The game replicates this by having real people estimate that.

What is clear is that pure genetics is not enough as an argument, so why would it be a hard limit? Opportunities play a huge, huge part in most things - which month you are born plays a role through when sports seasons begin; where you are when a scout visits your youth club; who you know in terms of footballing contacts; injuries... In a lot of ways, hand-eye coordination and sprinting speed are pretty small issues compared with the vast majority of things that are out of your control to a large extent. The book Outliers goes into way more detail, more of less debunking the idea of "destiny" or "fate" and some notion of a limit that is down to "natural talent".

It's a hard limit because everything in the history of football suggests there is a hard limit. As I have said, that's part of why scouting exists. Yes, lots of other stuff comes into play, but all that stuff is about how likely you are to make your potential. it doesn't actually change it.

Destiny/fate is nonsense, obviously. But 'natural talent' isn't, especially not in sport.

Dave C · January 11, 2013

If you get an 18 year old playing in your first team and has 3 great seasons and reaches his PA he never improves even if his PA is something low like 130, why would someone so young and playing so well suddenly hit a brick wall?

But they do. In real life too. In fact, the ones that hit a brick wall are far more common than the ones that don't.

YKW · January 11, 2013

I was heavily involved in the last 3 threads on the subject, when all the exact same points, by mostly the same posters were made. This hasnt moved forward in three years, its the same tired arguments over and over involving Luca Toni and Pedro and no one has yet come up with anything new in a long time.
Good word play there at the end tho, im sure you got a wee chuckle to yourself over it.

I did :lol:how cool am i?

Re "anything new," surely my BA idea falls under this category? I'm sure I have never read anything like this in any FM discussion all over the interwebs.

x42bn6 · January 11, 2013

If I could apply one rule to these forums, it would be "don't try draw comparisons from something other than football, to football"
It's not the same thing here. There's an entire branch of the football world dedicated to assessing and estimating potential. it's called 'scouting'. it exists because everyone working in football believes, knows, that there is a hard cap on a players future development. The game replicates this by having real people estimate that.

You say that "real people estimate that [limit]". So why not treat that number as an estimation, that can be underestimated? If you like, this number a researcher estimates has an uncertainty, and the realistic thing is to model that uncertainty too. But in doing so, the uncertainty must allow for underestimation (as in reality, the researcher may underestimate a player).

That said, scouts don't look for "players with high ceilings". They look for talent: Players who have good technique, players who can spot good passes, players who have good decision making, players with good attitude, players who have good scoring records... Every single scout knows that some players never star at youth level but develop rapidly later-on in life (i.e. Tom Cleverley: A ho-hum right-back at 17, to one of Manchester United's best midfielders this season), so you cannot judge a limit at a tender age.

Take Raheem Sterling: He's not seen as a possible future great because people think he has a high ceiling: They think he's a possible future great because he's getting lots of first-team football and he's delivering the goods in a top league; and his development has been "sterling" for a number of years, with many good games at youth, reserves and now first-team level. If he continues his rate of development, there is no reason why he can't become a great player in the future. His "ceiling" never comes into the picture - nobody cares about his ceiling.

Let's face it: If you are a League Two scout, you know a player's ceiling is only going to be known if they move to Barcelona early, play for their first-team and deliver the goods there - which is very unlikely to happen to a League Two player anyway. So the limit has absolutely no point in being considered at that level. In fact, it's not worth considering at any level, because Barcelona may still train players sub-optimally (and importantly, we don't know if this is true). You scout a player based on what you know now and extrapolate to the future.

It's a hard limit because everything in the history of football suggests there is a hard limit. As I have said, that's part of why scouting exists. Yes, lots of other stuff comes into play, but all that stuff is about how likely you are to make your potential. it doesn't actually change it.
Destiny/fate is nonsense, obviously. But 'natural talent' isn't, especially not in sport.

Even if we did have a hard limit, how would you measure it? How would you even know if a player had hit his limit?

Did Zidane hit his limit? You could argue Zidane hit his limit, or you could argue he didn't and could have been so much better. It is impossible to tell: It is unfalsifiable.

Natural talent is very different to a realistic limit within the confines of football, which is very different to the ultimate limit of a human being. I have no issues modelling natural talent (although I think that is CA, really), but I see that as different to deriving a realistic limit from it, and deriving the ultimate limit is an unknown science.

Destiny and fate are nonsense, but PA is nearly there. PA influences fate by limiting certain possibilities.

Spart · January 12, 2013

You say that "real people estimate that [limit]". So why not treat that number as an estimation, that can be underestimated? If you like, this number a researcher estimates has an uncertainty, and the realistic thing is to model that uncertainty too. But in doing so, the uncertainty must allow for underestimation (as in reality, the researcher may underestimate a player).
That said, scouts don't look for "players with high ceilings". They look for talent: Players who have good technique, players who can spot good passes, players who have good decision making, players with good attitude, players who have good scoring records... Every single scout knows that some players never star at youth level but develop rapidly later-on in life (i.e. Tom Cleverley: A ho-hum right-back at 17, to one of Manchester United's best midfielders this season), so you cannot judge a limit at a tender age.

Take Raheem Sterling: He's not seen as a possible future great because people think he has a high ceiling: They think he's a possible future great because he's getting lots of first-team football and he's delivering the goods in a top league; and his development has been "sterling" for a number of years, with many good games at youth, reserves and now first-team level. If he continues his rate of development, there is no reason why he can't become a great player in the future. His "ceiling" never comes into the picture - nobody cares about his ceiling.

I disagree. That video of Redknapp defending a young Lampard at a West Ham fan conference (watch here:

) shows that managers, coaches and scouts do indeed seek players who are not only good for their level but who also have a high development ceiling, sometimes doing away with other talents who lack the potential and future of others. At least one West Ham fan didn't think Frank was up to it so it wasn't just a case of a high level at a young age. Redknapp and his staff, if you like, though Lampard had a high PA and invested their time and resources developing that rather than progressing with players who didn't have the ceiling for greatness.

A more recent example: why have United sold off Brady to Hull this week? If what you say is true he should have instead stayed, worked hard and developed exponentially into the next Ryan Giggs or beyond - he certainly isn't a limited player technically or a troublemaker.

Also, why did United sign Nick Powell? If that isn't a transfer based around a player's development ceiling, I don't know what is. He was good for Crewe but he didn't show Premier League quality, but Premier League potential.

Let's face it: If you are a League Two scout, you know a player's ceiling is only going to be known if they move to Barcelona early, play for their first-team and deliver the goods there - which is very unlikely to happen to a League Two player anyway. So the limit has absolutely no point in being considered at that level. In fact, it's not worth considering at any level, because Barcelona may still train players sub-optimally (and importantly, we don't know if this is true). You scout a player based on what you know now and extrapolate to the future.

Why don't Barca's scouts hoover up every single player who isn't terrible on their worldwide searches then and pump out 11 Messi's every year? Could it be that different players do have different potentials with their own development limits?

This extrapolation you're suggested scouts partake in, isn't that just their judgement on a player's potential? How is that different from the decisions researchers make when judging a player?

If a club has a flawless development centre would every single youth player become great? And if they did, would that be because all players have the same potential or that the club finds the right type of high-class talent to feed their academy, like I dare say Barca do? Could that be why the likes of Xavi graduated through their academy while many others weren't good enough to be footballers?

Gunblade · January 12, 2013

I'm gonna repeat this again for the 3rd time, cause no one seems to understand.

I'm not saying i'm right but by god some of you guys are thick headed. It's like you can't discern between black and white and only take extreme cases.

"Oh so why aren't there Messi's, everywhere???"

"Everyone has infinite potential!!!"

NO NO NO!!!

There is a middle ground argument here. I've discussed it, others have discussed it.

Yes it'd be feasible to train anyone to be a good player BUT yes there IS also a LIMIT.

MIDDLE GROUND. Again I'm not saying I'm right but some of you are so damned stuck on some forsaken belief that the other is arguing for the extreme end of the spectrum.

We ALL know that there is a physical limit, not everyone is built the same. No one, or certainly most people would never believe that everyone can run as fast as Ronaldo, or as strong as that brick of a defender who plays for who ever (Can't find of someone >_>).

We also know that there MAY be a mental limit, this however is difficult to prove.

Now I believe (My opinion, not fact) that there is no technical limit.

"BUT WAIT", I hear you cry. "SO where's all the footballing gods!?"

Potential unfulfilled, not through some technical cap. You ALWAYS get better at someone the longer you do it. The reason people start getting worse, in later careers then?

Well obviously, I answered my own question. "Late in their career", it's not their techincal skill (Opinion). It's, and this is fact, their physical side. We all hear how players get tired more often, start struggling to cope with the game with age.

It's not that technicality goes down it's that age catches up with them and they can't run as fast they use to, can't muscle the other like they use to, aren't as agile. It slows them down.

Sure they can still make that 30 yard pass, but if they can't last even 45 minutes you start becoming a liability.

And that's just one part.

MENTALITY.

Barca aren't pumping out 100 Messi's every year not cause the player can't learn. Given enough time they could become technical wizards. But again there we go with that one word.

TIME.

Not everyone has the brain to learn as fast as another. Our brains our built differently and some might be amazing at academics but not sports, the other, a brain who's able to adapt physical lessons like Einstein to physics.

They learn faster, develop faster.

That kid, Barca let go, the one's no one ever heard of? Wasn't developing fast enough, maybe his physical presense was weakm maybe his mentality just wasn't strong enough to match the ability of the other players.

Could still be a good player though, just not good enough for Barca (No one in particular, just a random example).

Determination, work rate, decision making, pressure handling, concentration, to take some stats from FM.

These in real life may also determine how well you can handle the game.

As well as the mentality to determine if you learn fast enough.

The average pro footballer has 14 years in him if minor to no injuries?

Not a lot of time.

I've said this before again previously (And this is new since it was only reported last year I believe) but it's been scientifically proven that anybody can reach a pro level in their field by putting in the hours. That number was 10000 hours. An average, nothing more.

It could be possible that anybody, even me, could even be as good as Messi but where Messi took 5000 hours playing football (for illustration purposes, don't know how many hours he's actually put in). I'd probably have to take 50000 hours (even that might be a stretch).

Yup, as long as I live forever everything is possible but no, we are limited by age and time.

Summary. All players aren't magic cause of body build and time. So to finish up with my opinion. 2 seperate PA's.

Physical, mental (Hard to gauge though), but no technical cap. The cap for technical is time and mental attributes.

Crap mental attributes, you'll never train yoursdelf up to be amazing.

Crap physical attributes, even with good skill and mentality, you'll struggle to last a top level game, hell you even struggle to keep up with training.

Great at everything? Well that's just rare then isn't it?

Jibby123 · January 12, 2013

I'm gonna repeat this again for the 3rd time, cause no one seems to understand.
I'm not saying i'm right but by god some of you guys are thick headed....

It's a novel idea online; being the expert talking to idiots (everybody else)....but I'd stick with it mate. :thup:

Spart · January 12, 2013

Summary. All players aren't magic cause of body build and time. So to finish up with my opinion. 2 seperate PA's.
Physical, mental (Hard to gauge though), but no technical cap. The cap for technical is time and mental attributes.

Crap mental attributes, you'll never train yoursdelf up to be amazing.

Crap physical attributes, even with good skill and mentality, you'll struggle to last a top level game, hell you even struggle to keep up with training.

Great at everything? Well that's just rare then isn't it?

Pssh, thickheaded my eye. What does your supposed middle ground that isn't already in the game and doesn't already convey real life football?

Ryan Giggs, Frank Lampard and to a lesser extent Michael Owen have all adapted their game after losing physical abilities with age and injury. This is reflected in game by CA shifts through different attributes.

What your suggestions doesn't help with is making player development more complex and less linear - the real crux for making PA more interesting/realistic.

pret · January 12, 2013

I agree to an extent. PA is too limited. All regens (and to an extent, young players in the game) should have huge potential ability, but also, a more realistic figure set.

Potential states how good someone can become, and when Chris Smalling was in non-league, did anyone say "This lad could be one of the best defenders in the world one day"? No. Now, who knows, he might not progress much further than he is now, but at the same time, he could. It happens to a lot of players who are late developers.

Now correct me if I'm wrong, but a player's current ability can't exceed what his potential ability was can it? I think it should be able to. People understate potential regularly, and if a player has the right attitude, training, development etc. there's no reason IMO as to why his 170 potential prevents him from being a 185 player.

Spart · January 12, 2013

I agree to an extent. PA is too limited. All regens (and to an extent, young players in the game) should have huge potential ability, but also, a more realistic figure set.
Potential states how good someone can become, and when Chris Smalling was in non-league, did anyone say "This lad could be one of the best defenders in the world one day"? No. Now, who knows, he might not progress much further than he is now, but at the same time, he could. It happens to a lot of players who are late developers.

Now correct me if I'm wrong, but a player's current ability can't exceed what his potential ability was can it? I think it should be able to. People understate potential regularly, and if a player has the right attitude, training, development etc. there's no reason IMO as to why his 170 potential prevents him from being a 185 player.

If you're not checking a player's PA with the editor, Genie Scout or the like I doubt you'd be able to tell a 170 player with such an excellent attitude from a 185 player, especially if they have good stats for their role, consistency and big game stats, especially when in form.

lonestar190 · January 12, 2013

I'm gonna repeat this again for the 3rd time, cause no one seems to understand.
I've said this before again previously (And this is new since it was only reported last year I believe) but it's been scientifically proven that anybody can reach a pro level in their field by putting in the hours. That number was 10000 hours. An average, nothing more.

This is a theory based on research of some already successful people, it is far far from scientifically proven. Don't call other people thick and then spout stuff like this.

Kain · January 13, 2013

But more importantly, the game was wrong, and today's game will be wrong in a few years time, so the game is... always going to be wrong. Now imagine a game where the engine could compensate for human error - it's probably still going to be wrong, but it's obvious that it will be less wrong than before. Wouldn't that be better?

Surely that's only true if you think researchers are wrong more often than they're right? Because the engine compensating for human error will also be changing things that are pretty much correct too.

Gunblade · January 13, 2013

Maybe i used the wrong word, I'm not calling people stupid just stubborn. It annoys when I see people give the same argument, and to use the same example again, about "So why aren't there more Messi's in the world" if everyone had no limits?

We've given you an answer, we've given you a theory. There are multitudes of posts given you their opinions about this question yet have you referenced any of them? No you haven't.

At the least disagree with it.

And about the 10000 hours theory. I apologise yes it's Theory, but it's a solid and sound theory in my opinion and is certainly not old news if we're comparing this a 7 year argument on FM PA.

i'd like to ask all of you.

Do you believe, yourself, could only hit a peak with a technical ability? Do you think that if you had trained from the age of, let's say 6, for hours on end on passing that your passing skill will still suck?

Really?

I honestly don't see how. train long enough on a repetive regime and your passing "Skill" would be amazing no?

Again this doesn't mean you'd make that great pass in a real game, where other players are moving, closing you down, putting you on pressure. Then there's you creativity, decisions and concentration.

That what makes a good passer, technical skill could 20/20 and you could still suck. But as for this example, your technical skill would still be amazing. Train hard and long enough and you'll be good at it.

Isn't this what were taught every day, in general life?

And to finish on a different note, I'd like SI to actually test some ideas, can't be that hard, give it a shot, even if it's an inhouse only test. If everyone's arguing about it prove them wrong or right. Does it work or not?

At the very least take a look into the current system and look into changes, not just stick with the status quo for the next 10 years.

dragonheat · January 13, 2013

http://www.myfacewhen.net/uploads/239-oh-look-its-this-thread-again.jpg

Nikopol · January 13, 2013

I think a lot of people make it more complicated than it should be. I agree that fixed values are unrealistic and unauthentic, they lead to a lot of problems. I also agree that the Ability that a player can reach or not reach should be determined by training, first-team football, hidden attributes, form and manager.

To make the game more open-ended the players should get a ''development-value'' from -0 to -10 (as an example) and that should determine how easy a player can progress and get better. So a -7 player can still become a world class player if the manager works hard on improving him, tutors him by good veterans etc etc. (Cristiano Ronaldo for example) A -10 player would see those increases easily, it isn't a problem for him to get on a good football lever, it comes naturally (Ronaldinho,Messi)

To balance this a bit more and not turn this into Godlike-system. Mentals should play a much much higher role. Even with a bad personality every player can reach his maximum PA easily. If you know through your scouts that a player has 4 or more PA-stars you know that if you give him playtime, that he will become a world-class player. That should be different. Determination should be a hidden attribute too, to make it harder to distinguish how good a player can become.

So example a player like Balotelli who got incredible skills and potential (let's say a -9 by my system) will maybe never reach a level where he belongs to the great players because his personality make it hard for to progress and develop.

Edit: When i say that it is no problem to reach a good level (ronaldinho,messi) i don't mean that it happens automatically. Every player at the top levels worked hard to get at such a high level, but some already have godlike talent they are born with and some have to work their ass of every day to get at such a level.

Dave C · January 13, 2013

ITo make the game more open-ended the players should get a ''development-value'' from -0 to -10 (as an example) and that should determine how easy a player can progress and get better. So a -7 player can still become a world class player if the manager works hard on improving him, tutors him by good veterans etc etc. (Cristiano Ronaldo for example) A -10 player would see those increases easily, it isn't a problem for him to get on a good football lever, it comes naturally (Ronaldinho,Messi)

Hang on, how is that system any more realistic? You are still predicting what level that player will reach, just in a slightly different way.

I find it ironic that you say others make it too complex, then propose something even more complicated than we have now!

Dave C · January 13, 2013

You say that "real people estimate that [limit]". So why not treat that number as an estimation, that can be underestimated? If you like, this number a researcher estimates has an uncertainty, and the realistic thing is to model that uncertainty too. But in doing so, the uncertainty must allow for underestimation (as in reality, the researcher may underestimate a player).

Which is why you should have a better development model, and estimation that tends to optimism.

That said, scouts don't look for "players with high ceilings". They look for talent: Players who have good technique, players who can spot good passes, players who have good decision making, players with good attitude, players who have good scoring records

I'm going to pull you up there. Have you spoken to many real-life scouts? I have. When I started doing research for SI I actively spoke to real professional scouts to find out how they judge players. The stuff you are talking about is the evidence they draw on to predict where a player can go. Things like exceptional technique or vision, or strong mentality, are the things they see as indicators of high potential.

Take Raheem Sterling: He's not seen as a possible future great because people think he has a high ceiling: They think he's a possible future great because he's getting lots of first-team football and he's delivering the goods in a top league; and his development has been "sterling" for a number of years, with many good games at youth, reserves and now first-team level. If he continues his rate of development, there is no reason why he can't become a great player in the future. His "ceiling" never comes into the picture - nobody cares about his ceiling.

You are absolutely wrong about that. He gets the time because people think he can become great. His manager, coaches, scouts, all care a lot about his 'ceiling'.

Let's face it: If you are a League Two scout, you know a player's ceiling is only going to be known if they move to Barcelona early, play for their first-team and deliver the goods there - which is very unlikely to happen to a League Two player anyway. So the limit has absolutely no point in being considered at that level. In fact, it's not worth considering at any level, because Barcelona may still train players sub-optimally (and importantly, we don't know if this is true). You scout a player based on what you know now and extrapolate to the future.

Yes, you predict their potential based on what they show now. Just like...

Dave C · January 13, 2013

...Now I believe (My opinion, not fact) that there is no technical limit...?

I've cut everything else out to prevent massive embedded posts, because this is actually the key point. Your entire 'middle ground' system, which would make life far more difficult, is based on your belief that there is no technical limit. But there's a mountain of evidence that stacks against that. Yes, there's a good book that argues otherwise, but it's certainly not 'scientifically proven'. It's a good hypothesis, not even a theory really, but that's all.

Dave C · January 13, 2013

And to finish on a different note, I'd like SI to actually test some ideas, can't be that hard, give it a shot, even if it's an inhouse only test. If everyone's arguing about it prove them wrong or right. Does it work or not?

So they should use thousands of hours of development and testing time to look at an alternate system, based on a hypothetical principle, that massively over-complicates development, that may not even make it into the game? At the expense of just working on making the existing game better?

BiggusD · January 13, 2013

If it is true what people say (contrary to my experience), that all the players need is match experience and then everything is sorted, then the development module needs to be looked at. That's a major imbalance in the user's favour since the squad selection logic of the AI is still poor.

I think the game simulates well how exceptional talent develops, as well as how some kids can fail despite having the potential. It is the creation and development of decent players that needs the most work. It cannot be the case that all players of decent-to-good PL quality and below are failed worldwide star prospects (because of bad personalities). Neither should it be the case that all top international players are super-serious professionals. Maybe Professionalism (rate of CA gained), Ambition (amount of CA gained), match experience and events (causes of CA gained) are not enough factors? If so, what is missing?

What about Favourite Training Category? Technical, Mental or Physical; 1-20 in each with the same logic as Positions (one is always 20), with templates for football culture as well as primary position (Defender, Midfielder, Forward). Hidden, but with hints in the personality screen? That way, even though a player has everything else necessary to develop into a class player, if he as a youngster has poor physique but remain focused on technique throughout his career, he may max out his position's technical attributes but still not be world class thus -not- reaching his potential... in a logical way.

Ackter · January 13, 2013

Yup, it's all about the development system for me.

Get that working well and potential will never be a problem.

milnerpoint · January 13, 2013

Yeah thats been my standpoint for the last 2 years on this, they need to work on how players develop, not worry about PA as a number, you should never know it, so you should never worry about it, what should happen is it should be next to impossible to know when a player has hit this level, which is not the case at the moment. It also should be much harder to reach this level, and should not happen as young as it does for the majority of players.

OldSwanner · January 14, 2013

Have a 1/2/3 attribute that affects 'growth' rate. Simple.
1 = Grows normally as per your training facilities et al (as we have now)

2 = Ability grows too damn fast - more likely to reach potential before 20 (Early-bloomers)

3 = Ability stagnates but grows x4 speed once the player hits 26+ years of age. (late-bloomers)

It would have to be hidden, random and scouts should NOT be able to identify it at all. An early bloomer with LOW PA will likely sucker an unsuspecting manager into thinking they got a diamond in the rough and so on.

To be honest, this discussion crops up every installment and yeah it does need changing but it won't be.

Yes! Vote for this!!

x42bn6 · January 14, 2013

I disagree. That video of Redknapp defending a young Lampard at a West Ham fan conference (watch here:
) shows that managers, coaches and scouts do indeed seek players who are not only good for their level but who also have a high development ceiling, sometimes doing away with other talents who lack the potential and future of others. At least one West Ham fan didn't think Frank was up to it so it wasn't just a case of a high level at a young age. Redknapp and his staff, if you like, though Lampard had a high PA and invested their time and resources developing that rather than progressing with players who didn't have the ceiling for greatness.

'Arry in defending his player shocker...

'Arry didn't see Lampard as having a high potential. He saw Lampard as having great technique and awareness at a young level, with a great attitude. As we see today. His game has never been based on pace and strength, too, which is an even bigger bonus. If Lampard had continued his development and first-team performances throughout his career, he would have turned into a great player, as we see too. That's really all there is to it.

A more recent example: why have United sold off Brady to Hull this week? If what you say is true he should have instead stayed, worked hard and developed exponentially into the next Ryan Giggs or beyond - he certainly isn't a limited player technically or a troublemaker.

We have 8 attacking players competing for 4 attacking spots and you wonder why Brady left? Not to mention he now has 3 players ahead of him at left-back, which is a weird position he's played in for friendlies in the past, because of the 8 in front of him up front.

I've always said that for a player to develop well, he needs to to train well and avoid injuries, get first-team football at a high level and perform when he gets those chances. Brady has sacrificed the first, in the form of world-class training facilities, in exchange for much more first-team football. He'll look at the likes of Shawcross (constantly linked with Manchester United, although I personally think it's nonsense) and realise that he might still have a chance to make it at a top-tier club, maybe even at Manchester United again.

Also, why did United sign Nick Powell? If that isn't a transfer based around a player's development ceiling, I don't know what is. He was good for Crewe but he didn't show Premier League quality, but Premier League potential.

Again, like Lampard, it was nothing to do with "potential". Sir Alex did not look at Powell and imagine him as a future superstar: Sir Alex has had plenty of successes and failures in young players and he knows better than to extrapolate.

So why was Powell, like Morrison and Pogba, perhaps, considered so "promising"? The answer in Powell's case is simple: technique. Powell's passing and shooting ability are fantastic for his age, and he is physically quite strong and nimble already - so a lot of the hard-to-train attributes are already there to a large extent, which increases his chance of success. But more importantly, he's not only a good player "in theory": He's a good player in practice, too, delivering the goods for Crewe at a young age, which justifies his good technique and possibly suggests more things (i.e. maturity, performances under pressure, etc.), rather than him just being a beach football player with great "tekkers".

Why don't Barca's scouts hoover up every single player who isn't terrible on their worldwide searches then and pump out 11 Messi's every year? Could it be that different players do have different potentials with their own development limits?

If a club has a flawless development centre would every single youth player become great? And if they did, would that be because all players have the same potential or that the club finds the right type of high-class talent to feed their academy, like I dare say Barca do? Could that be why the likes of Xavi graduated through their academy while many others weren't good enough to be footballers?

Again: I've always said that for a player to develop well, he needs to to train well and avoid injuries, get first-team football at a high level and perform when he gets those chances.

Firstly, in order to develop well for Barcelona, you need to be technically excellent, and have great awareness. This rules out tons of players from the start. That's strike 1.

Then, the player needs to train well when at Barcelona, which involves mastering tiki-taka, one of the hardest tactics to play. The player must also sustain his development well. A lot of players are technically brilliant but there's more to tiki-taka than technique. The players must also develop well, which involves a professional attitude and avoiding injuries. A lot of players will fail at this level too. That's strike 2.

Then, the player needs to get into the first-team, which happens to be the best team in the world today. So that would mean being capable of playing for a La Liga side at a young age. Which rules out even more players. Let's not forget that a lot of Barcelona graduates don't even start (or even play) in La Liga - we only hear about the good ones. So that's strike 3.

Then, the player actually has to deliver for Barcelona, or he will get shipped out like Bojan. If the player gets this far, he could end up like many other Barcelona players, a Keita or Mascherano rather than a Messi or Xavi. This is not the worst career in the world, of course, but it's probably the biggest hurdle of them all.

Messi ticked all the boxes: Great technique and mentality from a tender age, developed fantastically well as a youngster, broke into the first-team as a result and most importantly, delivered. And boy, did he deliver - his performances for the first-team were exceptional for a world class player's perspective, let alone his age. To me, this is why Messi turned out brilliant, as opposed to the likes of, say, Busquets (good player, but was not ripping up trees and flinging them around).

In addition, signing hundreds of extra young players has seriously diminishing returns and introduces noise that can result in Barcelona missing out gems they already have.

This extrapolation you're suggested scouts partake in, isn't that just their judgement on a player's potential? How is that different from the decisions researchers make when judging a player?

There isn't much difference but one key one is that a scout can underestimate a player and watch him prove him wrong. A researcher can, but that "outperforming" can only happen in real-life, not in-game.

One other difference is that a scout will practice risk-reward because there's a different goal for a scout than a researcher (a scout may see a promising goalkeeper but that's a bit useless if the club has even more promising goalkeepers already). In addition, a scout's mentality may be to not operate on the same "level" as a researcher in finding the maximum, but rather, some sort of "average".

Surely that's only true if you think researchers are wrong more often than they're right? Because the engine compensating for human error will also be changing things that are pretty much correct too.

Well, there are 200 data points to pick from, so the odds are always against them. I think most researchers are correct in terms of getting it in the same ballpark range, but there are a few niche, corner cases where it doesn't quite work (Toni, Pedro, etc.).

Strictly-speaking, there is no "correct" value for PA because the moment you start a game, the game world and real world diverge: What is correct in reality (Neymar is awesome!) may not be correct in your game (Neymar suffers a damaged cruciate ligament!). So as long as the game compensates for human error "correctly" and produces "sensible" results, it is fine.

Which is why you should have a better development model, and estimation that tends to optimism.

That "optimism" is partly-derived from how a player develops - so that "optimism" must also be partly-derived from how a player develops in-game, too.

I'm going to pull you up there. Have you spoken to many real-life scouts? I have. When I started doing research for SI I actively spoke to real professional scouts to find out how they judge players. The stuff you are talking about is the evidence they draw on to predict where a player can go. Things like exceptional technique or vision, or strong mentality, are the things they see as indicators of high potential.

Key bits in bold.

Predict: What they make is a "prediction". One that can be underestimated, as well as overestimated. One that could be refined over time as they get more information.

Overestimation exists, but the mechanics of underestimation and refinement of the prediction do not exist in-game.

The second bold bit: This derives the third bit. "High potential" is in the sense that since many of the "ingredients" are there already, that player is far more likely to turn out great than an average player in his peer group.

Because these derive a prediction, in a programming sense, it's not actually necessary logic that needs to be explicitly stored as a hard-coded value. If a player has all the ingredients that will really help him develop, then he should be seen as "promising" regardless of whatever the (initial, defying) prediction was at the start. If you insist on the prediction existing, it should be exactly that - a prediction.

Analogous to attributes, perhaps: If a researcher were to be extremely accurate, the game would let them input attributes with confidence intervals (so passing 12 +/- 2.1 or something), and the game would generate it as it needs it (which is immediately, on game start). PA would be similar except that the reason for uncertainty is down to initial measurement error, plus any uncertainty associated with how they develop in-game (which has an analogous bijection to reality). So in a lot of ways, this is a plus for a researcher since they can never be wrong. They can never be right, too, but they are no longer expected to be.

It's just a "sign", anyway - there are doubtless many players out there who are technically excellent, mentally astute and have great professionalism and physicality, but are never scouted nor used properly. "High potential" is not a ceiling in a literal sense, but a suggestion of risk-reward in that these promising players have lower risk for higher rewards, because they have many good things going for them already.

You are absolutely wrong about that. He gets the time because people think he can become great. His manager, coaches, scouts, all care a lot about his 'ceiling'.

And why do people think he can turn out great?

It's already answered above: A combination of his ability, attitude and ability to deliver for the first-team at a young age.

Nobody is going to extrapolate a Heskey-like curve for Sterling at the moment, because all signs suggest that high ability + consistent performances = great player. It might be fair to say that he might even turn it up several more notches and surpass the Liverpool training staffs' estimations, too. Nobody knows.

Yes, you predict their potential based on what they show now. Just like...

They predict using the possibilities likely to a League Two youngster. However, do you really want to define potential in that way? Because you are basically defining it as "they won't be better than they will ever be", which is circular reasoning! You would have thought that "potential" in that sense would have been defined as "if this player gets the world's best training facilties, the world's best tutoring, and plays for the world's best team and averages 10.00 per match, his potential would be the sum of his current ability now, plus the development 'bonuses' attained by all these points". This is a far cry from "League Two training facilties, hoof-it-up football and average performances".

It would also be fair to say Luca Toni had the same issue before he late-bloomed: His potential would have been defined as that of a fairly mediocre Serie A/good Serie B player, but as we see here, the researcher failed to account for the possibility that he could turn into a striking monster and one day play for Bayern Munich. I don't really blame the researcher for that, but they should not be put into that position in the first place: The game should decide that.

martinji · January 14, 2013

In my opinion PA as a concept is fine. In general PAs should be higher, it should be much more common to fall short so that development plays more influence on ultimate CA than currently, and it should be much less obvious who the future stars are - more late developers, less accurate scouts. It should be rare for a player to reach 100% of their PA - and this would then leave room to get around some of the other issues, like attributes inexplicable getting worse just to accommodate another attribute getting better.

Separate PAs for different sets of skills would be a decent approach.

I'd also be perfectly happy to simulate the unpredictiveness of PA by having the PAs of all players randomised somewhat, or eg the game picks say 1-2% of players and adds a uniformly distributed random number between 0-100 to their PA. There are your unpredictable late bloomers. The development & scouting model would have to cope with this though, in that they should need some trigger to set off development rather than suddenly just start improving, and it shouldnt be obvious who they are either.

mixmkz · January 14, 2013

Perhaps the players should have a PA of an randomly assigned categories (ex: -9, -6 etc) until a certain age. So if players never get above CA of 150 by age of 21, then newgen's PA can remain random until then and then receive a final number after to reflect if they are peaking to or still have a lot of development to go.

Dave C · January 15, 2013

That "optimism" is partly-derived from how a player develops - so that "optimism" must also be partly-derived from how a player develops in-game, too.

Yes, in terms of how a scout views and rates the player. But I've always argued scouting should be based on attributes and performance, not just CA/PA.

Key bits in bold.
Predict: What they make is a "prediction". One that can be underestimated, as well as overestimated. One that could be refined over time as they get more information.

Overestimation exists, but the mechanics of underestimation and refinement of the prediction do not exist in-game.

Any system will fail at times to meet what happens in real-life. That's not an argument for throwing out caps in favour of something overly mechanistic.

The second bold bit: This derives the third bit. "High potential" is in the sense that since many of the "ingredients" are there already, that player is far more likely to turn out great than an average player in his peer group.
Because these derive a prediction, in a programming sense, it's not actually necessary logic that needs to be explicitly stored as a hard-coded value. If a player has all the ingredients that will really help him develop, then he should be seen as "promising" regardless of whatever the (initial, defying) prediction was at the start. If you insist on the prediction existing, it should be exactly that - a prediction.

But there are other factors too, some of which researchers have no way of adding in-game. The attributes are only part of the predictive process. To use a slightly extreme example, if I happen to know a player is in a bad circle, bad influences, I cannot code that in. But I know it's likely to hinder, even prevent, his development (and I've been proven right before!). The cap helps there.

Analogous to attributes, perhaps: If a researcher were to be extremely accurate, the game would let them input attributes with confidence intervals (so passing 12 +/- 2.1 or something), and the game would generate it as it needs it (which is immediately, on game start). PA would be similar except that the reason for uncertainty is down to initial measurement error, plus any uncertainty associated with how they develop in-game (which has an analogous bijection to reality). So in a lot of ways, this is a plus for a researcher since they can never be wrong. They can never be right, too, but they are no longer expected to be.
It's just a "sign", anyway - there are doubtless many players out there who are technically excellent, mentally astute and have great professionalism and physicality, but are never scouted nor used properly. "High potential" is not a ceiling in a literal sense, but a suggestion of risk-reward in that these promising players have lower risk for higher rewards, because they have many good things going for them already.

Far fewer than you'd expect. Worldwide scouting these days mean few players truly 'slip the net'.

And why do people think he can turn out great?
It's already answered above: A combination of his ability, attitude and ability to deliver for the first-team at a young age.

Nobody is going to extrapolate a Heskey-like curve for Sterling at the moment, because all signs suggest that high ability + consistent performances = great player. It might be fair to say that he might even turn it up several more notches and surpass the Liverpool training staffs' estimations, too. Nobody knows.

No, but you can make a fair prediction. And we do.

They predict using the possibilities likely to a League Two youngster. However, do you really want to define potential in that way? Because you are basically defining it as "they won't be better than they will ever be", which is circular reasoning! You would have thought that "potential" in that sense would have been defined as "if this player gets the world's best training facilties, the world's best tutoring, and plays for the world's best team and averages 10.00 per match, his potential would be the sum of his current ability now, plus the development 'bonuses' attained by all these points". This is a far cry from "League Two training facilties, hoof-it-up football and average performances".

Well, that's just plain wrong. Not much I can do about that. If a PL scout is looking at a L2 youngster, he is thinking how that player will be if his club brings him in. Similarly, if a researcher is thinking about a youngster, he is thinking how a player could be if he gets the breaks. In the current system, because the development model is limited, we tend to pessimism. But it's easy to change that if the model is more sophisticated.

Anyway, feel free to reply, but I'm remembering why I don't usually wade into this pointless, circular debate. I'm wading back out. SI aren't removing the cap, so you are shouting into a tin can.

Matshit · January 15, 2013

I think one of the biggest proof of the system not working is the fact that IRL Jordi Alba didn't even make it through the youth intake (or as close to it as it gets since there is no real intake IRL). I am pretty sure no one has ever even let go of a player of "bottom of PL"-quality in this intake. The current system is fine for most players, but it doesn't give any real surprises of the Alba/Di Natale/"other player with a sudden big increase in quality" type. Or for that matter a Adriano or Bojan who looked great as youths but stopped developing early.

In game, every player develops the same way. Big increase from 16-22. Increase slowing down from 22-28. Player peaking from 28-30. Player starting decreasing from 30 until end of career. As long as the player keeps out of injuries this is mainly the way it goes.

An honest question. If a player had CA 140 and PA 200 at age 25. Would anyone consider him a potential world beater?

I do think a less standardized development curve with a lot more factors playing a part+a higher average PA is the way to go. Players reaching their PA should only happen in rare cases where the player is a model profesional from birth, start out with a CA close to the PA or/and have perfect training for a long time.

Ackter · January 15, 2013

I'd take a CA140 PA200 player at age 25 because i can guarantee I could get him into the 170s at least.

iwantpie · January 15, 2013

For my money, there are only two real problems with the system the way it is:

1) It's far too easy to get a reasonably accurate idea of a player's PA. 'Get Scout Report' with an average scout will come close to the mark which is just them (as far as I can tell) watching some video of the player.

2) A player's growth stops too early. If a player hasn't reached his PA by 24 he's not going to grow much more.

I think the system could be improved by addressing these two problems. I'm not saying it would fix the system but it would be a bit deeper and more interesting:

1) 'Get Scout Report' gives an accurate indication of the player's CA and a very, very rough idea of PA based on reputation largely (are they a known, widely reported promising youth or an unknown?) e.g. Ronald McDonald would be able to make some contribution to our first team right away and indications are that he has a bright future. How much more can you glean from a day's research? This can be influenced by your scout's knowledge (not their JCA or JPA, it's what they know already, not how good they are at assessing players). If Ronald plays in Austria and your scout has maximum knowledge of Austria then he'll be able to put a better report together after a day than someone with no or minimal knowledge of Austria who can, effectively watch a game or two on videotape and make a couple of calls to someone (e.g. jouno or agent) who may know more.

If you want more accuracy you have to set a proper scouting assignment on the player. Have scouts watch video of the player over weeks and attend a match or two. You receive a more complete report after a month that gives you more detail (a star rating like now for CA and 'Ronald McDonald could perform at a good level in a top European league'). You can then keep scouting him and after another month (or whenever) you get a final report with a star rating for CA and for PA and an idea of the player's development type (see point 2).

2) There are a number of development models for a player. I'll use three as an example but you'd probably want more than this.

a) Wonderkid = quick advancement at an early age but if you miss the window (injury, bad attitude etc.) development slows early on and player is unlikely to reach his potential. This means that a player will develop very fast from 16-21 (e.g. Wayne Rooney, Eden Hazard) but if they have problems in that time or don't train well it turns into largely wasted potential by early 20s (e.g. Michael Johnson, Francis Jeffers)

b) Steady = development is a straight line and the player will require time but will continue to develop into his prime years (c. 25-27 yo). This results in players who will be good at 21/22 but not superstars but if they continue to get games and play well their development will continue into their mid-twenties. They may even continue beyond this though at a reduced rate. I guess players like James Milner or Radamel Falcao might come under this category.

c) Late Bloomer = development will be unremarkable in their early years and to reach their PA they will have to break through ON THE PITCH. What this effectively means is that this type of player is unlikely to get that close to their PA until 22-24 years old but any time from that point until c.30 they can take big leaps forward if tey have a big season or two. Examples include Drogba, Diego Forlan arguably Stewart Downing (breakout season at Villa). Note that they won't necessarily be bad just because they haven't reached their PA. If their PA is 180 then they could be around 120/130 at age 24, have a big season that kickstarts their development and bumps them up to around 140/145 and then they power on from there.

As I said, there would need to be more models than this.

I also quite like the idea that some players could have completely hidden chunks of PA. Only small amounts, 10 points or so, but the idea would be that a player later on in life could get a sudden extra spurt of ability after a good season that takes everybody by surprise. They have a mental breakthrough or are playing in a position or team they really love and a little light just comes on that pushes them on a little bit.

3) I know this doesn't relate to a point up there but I also don't like the way players lose PA between versions of the game at the moment. I read an article about him earlier so I'll keep using him as an example but take Michael Johnson. Just released by Man City aged 24, when he was 18 he was widely thought to be a future England captain. A combination of injuries and poor attitude stopped him from making strides forward. In FM13 his rating is as a 'good Championship player' and he is 'playing at close to his full potential'. Clearly that's not the case. Maybe he's not future England Captain any more but at aged 24 if he goes to a good club with a manager who can get the best out of him he could yet salvage something from his career. Should he lose some potential? Yes, definitely, but I would think he should still have at least the potential to be a good Premiership player. Is it likely? No, and that should be represented, but it should be possible to tap some of that potential which it just isn't at the moment.

Any thoughts on my ramblings?

x42bn6 · January 15, 2013

Any system will fail at times to meet what happens in real-life. That's not an argument for throwing out caps in favour of something overly mechanistic.

I'm perfectly fine with relaxing the "hardness" of PA, not necessarily removing it. I do think that eventually removing it is the answer and is naturally derived from relaxing it and considering reevaluation over extremely short timescales.

But there are other factors too, some of which researchers have no way of adding in-game. The attributes are only part of the predictive process. To use a slightly extreme example, if I happen to know a player is in a bad circle, bad influences, I cannot code that in. But I know it's likely to hinder, even prevent, his development (and I've been proven right before!). The cap helps there.

Why not just change his professionalism then? If he's not smart enough to not hang around with the wrong people, he doesn't exactly have the right attitude.

And even if not, you can do that without a cap: Just introduce a new attribute "environment" or something. Because you never know, he might grow up and leave this gang of nasty people shortly after you submit your data. Or he might simply move to a new club in a new part of the country (or even a new country) and he can't take his nasty friends along with him. Ravel Morrison had that issue in Manchester - Sir Alex hinted at that when he moved to West Ham.

Far fewer than you'd expect. Worldwide scouting these days mean few players truly 'slip the net'.

The number is immaterial - it's the mechanics. We already know scouts and managers make mistakes.

No, but you can make a fair prediction. And we do.

And that is all you can do. The game should therefore treat that number as a prediction, rather than a scientific absolute cap.

Similarly, if a researcher is thinking about a youngster, he is thinking how a player could be if he gets the breaks. In the current system, because the development model is limited, we tend to pessimism. But it's easy to change that if the model is more sophisticated.

How would a League Two researcher consider the possibility of that player moving to Barcelona, breaking into the first-team and outscoring Messi, which would derive a very good guess at a player's limit?

Hurball · January 15, 2013

My main problem is not with players having a maximum potential - it's that your scouts will very accurately know this potential after only one look.

I simply suggest that your scouts error in judgement is vastly increased, or made more vague.

Multiple scouting reports should increase his accuracy for current ability, but scouts should never be able to accurately tell if a guy will be a star.

I feel that the game is getting too easy with regards to knowing how good players are. I snapped up a young goalkeeper simply due to his 4.5 star potential, but this didn't feel as impressive as it should have been. I know he'll become good goalie, there's no uncertainty really. Unless I try to destroy his career, he'll be excellent. And since making PA changeable seems unworkable, the only logical change is to alter how the game displays this.

Flukey · January 15, 2013

I have been thinking about this.

I think the current system is the best we have at the moment, although it does have its flaws.

Saying that, how about this proposal:

We get rid of PA completely, and replace it with a hidden attribute, of the players ability to learn. That ability would decrease as the get older, and a high natural ability to learn may not lead to them becoming a superstar, as they would also need a club with good coaches etc, as they need the information to learn.

This means that you can have a person with high CA, but a low ability to learn, meaning he will never improve much.

It would also mean you can have a person with a low CA, but a high ability to learn, meaning that with the right coaching, he can turn into a superstar.

Thoughts?

MelbVictory · January 16, 2013

The only problems arise when you look at those numbers using external tools. If you use the in-game tools (scouts/coaches) then it works as intended.

This is absolutely correct. If you play the game as intended and don't look beneath the hood, there is no issue at all.

hackerz22 · January 16, 2013

I prefer the absolute mystery :confused:

Sign In

Potential Ability debate

Recommended Posts

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post

Share on other sites

Link to post