Potential Ability

[obinna]

Sounds to me that you two don't really want a dynamic PA but player attributes that are generated by what the player does in matches you observe. Players real attributes would be hidden from gamers the way CA is now, instead we would see a player profile that reflects the players recent observed performances. For example if you scouted a guy for one match and he missed two open goals you would see a Finishing attribute of 3 if you managed him and say that he scored one chance in two all season you would see he had a finishing attribute of eighteen.

That would certainly make scouting more interesting and long term scouting more important but It would add a hell of a lot of extra processing to keep separate observed player profiles for every AI entity that looks at a player.

Something along these lines is what I'm talking about. TBH I just want something that makes player development less predictable.

[tylerazevedo]

Of course a player can still develop but physically the way the player has grown as a child will already be set and also a persons intelligence is pretty much set by then also. Of course they can still learn things but if your stupid at 16 you will be at 26.

Sooo not true. People are constantly getting smarter, even if they aren't trying to. I would expect a 26 year old to be smarter than his/hers 16 year old self.

[Chris_l]

Untill today I didn't even know potential abillity was capped.

What's the point of buying and nurturing youngsters I wonder and investing in coaches and facilities when they can't get past a certain determined hidden limit? I feel like I've wasted an entire season with a weaker team. It's better to just start over and play with Barca/real/Chelsea and just buy a great team every year, all the time and energy and money spent on scouting for promissing youngsters, staff, fascilities and trying to become a good team by nurturing could be wasted on low capped players and you wouldn't know unless you have great scouts, and even still you won't know which attributes will be good and wich not.

[hursty2]

The dream is that there shouldn't be any attributes at all, the game should be completely stat based. For example a 16yr old has scored 20 goals in League 1, bigger clubs are sniffing round and buy him based on those goals not his PA.

[sinugie]

The dream is that there shouldn't be any attributes at all, the game should be completely stat based. For example an 16yr old has scored 20 goals in League 1, bigger clubs are sniffing round and buy him based on those goals not his PA.

this +1 the more i play the game the more i get disturb how easy to analyze player when the game giving the exact value of skill. rather than showing it, leave it under the hood and just give the player quick summary of few stat perhap throw some dynamic attribute displayed base on form, or perhaps report from the coach regarding player attitude in the training.

[Guest]

The dream is that there shouldn't be any attributes at all, the game should be completely stat based. For example a 16yr old has scored 20 goals in League 1, bigger clubs are sniffing round and buy him based on those goals not his PA.

+2

They could use the existing profile page but remove the attributes panel & replace it with a more general overview of the players core ability, maybe using the coach report as a base & then adding in feedback on recent performances that have shown up particular strengths & weaknesses in his game.

[x42bn6]

I showed some promise as a footballer in my youth but there was not a single set of circumstances in existence that could have aligned in a way that I would ever have reached the ability of Messi.

You're talking utter nonsense & it's to such an extent that I truly question your knowledge of elite sport & more specifically football.

It's not nonsense. You may have heard the 10,000 hours idea, where 10,000 hours are required to become "elite" at something. Had you devoted 10,000 hours to the sport, it is entirely possible that you could have become a very good footballer, and with a bit of luck and other things, it could well have been possible.

It is possible - it is just very unlikely (after all, Messi is 1 in around 7 billion). Realistically, you wouldn't bet your house on it. You're probably more likely to win the lottery. However, you cannot say it was impossible at 16 - especially since there was a large degree of uncertainty around predicting your future as a teenager.

Ronaldo was not playing anywhere near the top level of football when he was 16, yet look how he turned out.

Ah the old "everyone has the potential to be as good as messi" line. To be honest, that alone sums up why that particular suggestion is terrible, if that is in mind when devising this system then it will fail. I showed no promise as a kid, none at all, there isnt anything in the world that could have been done to make me a player of that caliber, anyone who says different knows very little about football.

Science disagrees. Studies show that if you work at something, you could become better. Maybe you could have worked harder?

Some players have risen from semi-professional levels to Premier League status. Bébé rose from homeless child to relative stardom - making his first Turkish league appearance the other day. That's unlikely enough!

Let me ask you - at what point does the probability of a player being unable to reach Messi's level become zero? Can you even measure it all, given the uncertainty around measuring the future and the fact that there is no definitive scale at measuring a player's ability? If you can't answer those questions, I question whether this is a sensible idea at all to definitely rule out the possibility of player X reaching Messi's level, as long as he is alive.

If the probability ends when he dies then it starts when he is born. After being born that player has 15/16 years before entering the game where their development will be different from anyone else and therefore, by the time we get them, they have already lost some of their initial potential.

Hold on, I thought you said potential was fixed? Why is it changing?

This means that at the stage of entering the game it is impossible already for them to be the best player of all time(200).

The implication being that the development time cannot be recovered at all? I'm not sure I agree with that. Drogba's career was going in a relatively mediocre direction prior to Chelsea.

Sure, it's more difficult - but not impossible. If it is - I'd like to know at what point it reaches 0%.

Of course a player can still develop but physically the way the player has grown as a child will already be set and also a persons intelligence is pretty much set by then also. Of course they can still learn things but if your stupid at 16 you will be at 26.

This isn't always true. Genetics only tell part of the story. Football is a sport where it doesn't require full geniuses (i.e. music/arts) or physical monsters (i.e. pituary gland issues, sprinters). It's a blend of the two. Things like hand-eye coordination can be developed as a youngster that help. Strength such as playing with weighted blocks, or crawling exercises can also help.

A baby is not born with a football brain. A baby doesn't even know how to control its own bladder, let alone play football!

Going down the pure genetics route is a pretty dangerous step, since we know that it is a blend of nature and nurture that produces talent. It is an open question whether there is such a thing as "superior genetics" when it comes to football, as it is such a complex sport.

As an example, the best footballer in the world in the future could well be a physical monster who has Elokobi's strength, Messi's agility and Walcott's pace, but limited technical ability. It is just that you cannot stop him when he has the ball, so he always brings the ball to a halt and picks out the correct pass/goes for goal. Would this be superior genetics? Nope - it's just different genetics that just so happen to produce a better footballer at that point in time.

[x42bn6]

I suspect I will need to head back to University to read some online journals about this, but here is a quote:

The prior example of intelligence shows that both nature and nurture work together in producing intelligence. Having great genes alone is worth little if intelligence is not nurtured in healthy and stimulating environments. Sometimes people assume that they have great "native" intelligence ("good genes") and they think that they do not have to work hard at being smart. They may be hurt by their own assumption that nature alone is important in developing their potential. Nature and nurture always work together: Both are important.

A study in Hawaii showed that children who had serious setbacks to intellectual development in their early years were sometimes able to excel if the early damage was counteracted with wonderful environmental inputs as the years passed. The negative effects of "damaging factors" can be offset by the positive powers of "protective factors," which include loving mentors, exciting opportunities, challenging tasks or jobs, and rewards for using what brain capacity one has. John will discuss these protective factors in the part of Soc 142 called "American Socialization."

http://www.soc.ucsb.edu/faculty/baldwin/classes/soc142/n&n.html

The implication of things akin to "destiny", to me, is a flawed idea and an easy way out that doesn't actually explore the idea properly. It's easy to state that the average person won't become Lionel Messi - I'm more interested in seeing if there is a point at which this happens (i.e. P(person becoming Lionel Messi) = 0), and even if it is not, how this is calculated. I think saying "never" is a very vague and dangerous way of abusing probability.

I think the chances of me becoming Lionel Messi are miniscule, but it would be a fruitless exercise to determine what that probability is, so it is easiest in layman terms to say it is zero. But I don't think it is, because "you never know".

[Guest]

You're over thinking a problem that in my opinion does not exist, I'd go back to one of my previous posts & ask how many players actually realise their starting PA in FM.

[x42bn6]

You're over thinking a problem that in my opinion does not exist, I'd go back to one of my previous posts & ask how many players actually realise their starting PA in FM.

You mean reach?

If so, it still misses the point. This thread is not about "reaching" your PA - it's about the notion of "PA" itself.

A player can fail to reach their PA. That's perfectly fine and I don't have too many issues with that (except the possible shape of the probability distribution, but that's a minor thing). In other words, I do not have too much of an issue with players failing to live up to the expectations of the researcher who decides it (peak CA < PA).

What I do have a problem with is a player who outperforms the researcher's opinion and ends up hitting their PA, implying they've matched the researcher's opinion. They haven't! They've exceeded it!

In other words: Players can match a researcher's opinion; they can fail to reach that opinion but they can never exceed it.

You might say that it's a limit, so it doesn't make sense to exceed it. Well, how about not calling it a limit and actually calling it what it is - an assumption. A guess. A number plucked out of not-so-thin air based on some statistics and opinions. An attempt to see into the future.

A limit is nice, concrete and not vague. A guess is dirty, dubious and vague. They are not the same. So we should stop calling it a limit. It's not.

If we don't call it a limit, then a player can be allowed to fail to meet expectations, meet expectations and exceed those expectations based on how he does within a saved game.

It is like saying that the game assumes "2 + 2 = 5" to justify something in the game. The fact that the game assumes "2 + 2 = 5" can and should be scrutinised as well.

[Guest]

A player can fail to reach their PA. That's perfectly fine and I don't have too many issues with that (except the possible shape of the probability distribution, but that's a minor thing). In other words, I do not have too much of an issue with players failing to live up to the expectations of the researcher who decides it (peak CA < PA).

What I do have a problem with is a player who outperforms the researcher's opinion and ends up hitting their PA, implying they've matched the researcher's opinion. They haven't! They've exceeded it!

Is your problem that a player can't exceed his FM potential in the game whereas he can exceed his FM potential in the real world? That's just the way it is, once the game starts you have to remove many links to the real world.

IMHO in the real world a player will never exceed their potential, what they will do is exceed their perceived potential based on the opinion of decision makers at any given club & this is where FM moves away from realistic behaviour because even without the use of FMTRE or Genie the potential of any given player can be fairly accurately defined because of how the scout/coach report works & the added aspect of years of game play.

If SI were t remove the unswerving accuracy of player scouting we would be heading in the right direction from a users perspective but this would need to be linked with a change in how the AI defines the value of a player, removing the absolute PA figure from all equations & replacing it with an algorithm based on CA/past form/present form.

[Smurf]

Potential ability can't be a "number". Especially for young players.

There really should be some elasticity on "some" players in regards to their PA. It should be able to go up or down.

Your scouts are not going to be able to give a definitive potential rating for a player, that's impossible. Currently in the game, that's what it gives you, you can list all your players by their potential rating and you'll have a good idea who are the good and bad players.

This makes it very easy to pick the "highest potential" and send them on loan. Which is perfectly fine.

But it also makes it far to easy to dismiss the other players, don't renew contracts or sell them on. The game makes it easy to see that these players will never amount to anything.

But in reality - they could get a growth spurt, or increase strength rapidly, which would make them perhaps a better defender. But the parameters of the game doesn't allow for this posibility, does it?

How about retraining a player to a new position. Surely the "potential" for him as a Winger as say 130. Retrain to a striker, who knows what his ability would be up there.

The game doesn't compensate for changes to environment. I could have top notch facilities. All this does is increase the speed in which they reach their PA.

But look at Ian Wright, he was playing for Dulwich at the age of 21 before he went to Crystal Palace. I'm sure their training and facilities greatly enhanced whatever potential he had, and the role of a striker.

Sure people will say "well his PA was always high, just never got a break" - but that's not the point really.

What about "partnerships" on the pitch, the young player with an experienced defender, or a young striker with an experienced striker.

Surely there would be some wiggle room in their Perceived potential and their actual potential?

[x42bn6]

Is your problem that a player can't exceed his FM potential in the game whereas he can exceed his FM potential in the real world? That's just the way it is, once the game starts you have to remove many links to the real world.

Yes, because "FM potential" is really just "researcher's opinion".

Many links have to be removed from the real world, true, but I think the aim should be to remove as few "mechanics" from the game as possible. The ideal game doesn't have to throw anything away.

IMHO in the real world a player will never exceed their potential, what they will do is exceed their perceived potential based on the opinion of decision makers at any given club & this is where FM moves away from realistic behaviour because even without the use of FMTRE or Genie the potential of any given player can be fairly accurately defined because of how the scout/coach report works & the added aspect of years of game play.

But they cannot exceed the opinion of potential expressed as an input by a researcher. Yet in the real world, a real player can indeed exceed the opinion of potential expressed as an input by a researcher.

There is a real world-FM world mismatch that needs to be fixed.

If SI were t remove the unswerving accuracy of player scouting we would be heading in the right direction from a users perspective but this would need to be linked with a change in how the AI defines the value of a player, removing the absolute PA figure from all equations & replacing it with an algorithm based on CA/past form/present form.

If it produces a better game, why is this a problem?

[Smurf]

But they cannot exceed the opinion of potential expressed as an input by a researcher. Yet in the real world, a real player can indeed exceed the opinion of potential expressed as an input by a researcher.

Or an opinion expressed by a Scout.

There are many real world examples of players who were scouted by various top clubs and were deemed not to have the potential to play at that level.

David Platt for example. Joined a 4th Division from Man Utd. But then went on to have a very successful career in England and Italy.

Who's to say that the youth players you get at intake are correctly interpreted by your current scouts or coaches?

[marty78]

Sooo not true. People are constantly getting smarter, even if they aren't trying to. I would expect a 26 year old to be smarter than his/hers 16 year old self.

People gain further knowledge based on their intelligence they do not become considerably more intelligent. We actually learn faster at a young age than when we are older but with age comes experience and therefore kowledge.

@x42bn6

You can nurture all you want but a player who runs 100m in 17seconds at 16 will never be quick without the help of some sort of biotics. I didn't mean they have actually lost potential just the ability to reach that potential. They have a potential ceiling at 16 because full potential would require perfect physical, perfect mental and perfect technically ability. Since it would take perfect progression up to 16, and humans are incapable of being perfect at anything over such a period of time, then the ability they can reach must be less than the ability they could of possibly achieved. If achieveable PA is less than actual potential it therefore must have a limit.

I think you are overthinking things in an attempt to persuade everyone about the merits of your arguements. Even if you raised all PA's to 200 it would still be a limit so surely that would still be a problem?

[x42bn6]

People gain further knowledge based on their intelligence they do not become considerably more intelligent. We actually learn faster at a young age than when we are older but with age comes experience and therefore kowledge.

@x42bn6

You can nurture all you want but a player who runs 100m in 17seconds at 16 will never be quick without the help of some sort of biotics.

I'm quite reluctant to compare football to things like sprinting and music because these require fairly "raw" genetic capabilities. Football is a combination of both, as well as mental skills that are more "nurture" (things like decision-making and positioning, for example, although they have minor genetic traits such as spatial awareness). I accept that genetics plays a huge part in sprinting - I'm just less convinced it matters as much in football. Like I've said before, Messi wins few genetic lotteries thanks to his size and medication requirements. Ronaldo wins tons more genetic lotteries but is an inferior player.

I didn't mean they have actually lost potential just the ability to reach that potential. They have a potential ceiling at 16 because full potential would require perfect physical, perfect mental and perfect technically ability. Since it would take perfect progression up to 16, and humans are incapable of being perfect at anything over such a period of time, then the ability they can reach must be less than the ability they could of possibly achieved. If achieveable PA is less than actual potential it therefore must have a limit.

Where did this arbitrary 16-year-old value come from when you mention "potential ceiling at 16"? What science declares that 16-years-old is the important part when it comes to development ceilings? Why not 15? Or 14? Or 15 years 250 days?

Also, even if this reasoning was fine (it's not, see late bloomers, but I digress): Say the limit was X. How would you calculate X without the benefit of hindsight? Can you state a value for X when a player is 16 that there is absolutely no chance of a player reaching a value X + e where e > 0? Or is there always going to be some uncertainty involved, including you underestimating X?

I think you are overthinking things in an attempt to persuade everyone about the merits of your arguements. Even if you raised all PA's to 200 it would still be a limit so surely that would still be a problem?

200 is a modelling limitation. I'd argue it should be infinity, but there are computational limits involved, of course. Realistically, all the attributes should go from 1 to infinity. It is just easier to scale the players to 1-200 (CA/PA) or 1-20/100 (attributes). But that's the point - it's easier - it's an assumption - and that can be relaxed in the future. It might take a while before that happens, of course, given the fact that humans today can't really pass better than, say, Xavi. Evolution will let us in the future, but that is a slow-paced thing we don't need to worry about in the short-term.

I don't think I'm overthinking things. I don't think a lot of people are thinking enough.

[marty78]

16 represents the point newgens come into the game. 16 was just easier than writing "when new players enter the game".

[x42bn6]

16 represents the point newgens come into the game. 16 was just easier than writing "when new players enter the game".

How do you determine the potential for a real-life 16-year-old? What equation or equations do you use?

[marty78]

Also, even if this reasoning was fine (it's not, see late bloomers, but I digress): Say the limit was X. How would you calculate X without the benefit of hindsight? Can you state a value for X when a player is 16 that there is absolutely no chance of a player reaching a value X + e where e > 0? Or is there always going to be some uncertainty involved, including you underestimating X?

Late bloomers potential doesn't increase, ever, they just take longer to get there than usual.

No-one knows the value of X as it unknowable, that is the whole point. Even with hindsight we don't know someones potential. All we know is what level they reached not what they were capable of. What we can say though is that X must have a limit regardless of what the limit is.

200 is a modelling limitation. I'd argue it should be infinity, but there are computational limits involved, of course. Realistically, all the attributes should go from 1 to infinity. It is just easier to scale the players to 1-200 (CA/PA) or 1-20/100 (attributes). But that's the point - it's easier - it's an assumption - and that can be relaxed in the future. It might take a while before that happens, of course, given the fact that humans today can't really pass better than, say, Xavi. Evolution will let us in the future, but that is a slow-paced thing we don't need to worry about in the short-term.

I don't think I'm overthinking things. I don't think a lot of people are thinking enough.

All attributes to infinity? Infinite pace? Faster than the speed of light? I reckon talking about infinities and the possibilities human evolution could bring is overthinking.

As long as PA is hidden properly (which it is not ATM) then that is in line with real life more than unlimited potential. I can understand the questioning of PA's of real players as they are guessed by someone, but I have fallen in line with the "FM is god" view with regards regens. As the creator the game has the divine right to set the parameters of the created. Even with most real players that are young enough to get a - rating there is an element of randomness. There is also an official editor should you wish to change the PA's of real players.

I could also understand raising PA's accross the board but making it pretty much impossible to achieve full PA. I can't believe in no limit though.

[Smurf]

The potential could have a +/- percentage depending on individual training and mental abilities.

With the right coaching they could exceed their potential. And with the wrong coaching their potential decreases.

It's kinda a silly thing to say, either you're good or your not. And they start with a definitive potential and it is up to the player/club to reach that potential.

The main flaw within the game is that it shows the potential too accurately.

That's where the individual training and mental attributes would come into play.

If you have the coaches (depending on their stats) giving an AVERAGE of what they think the potential of a player is, perhaps the players mental attributes are low. The coaches would then rate them as potentially reaching a 3 * rating.

But if you were to assign a good mentor - then this gives the coach an evaluation to say well actually he could be 4 star.

If the mentoring goes bad, then the coach might say - "well he's still got potentially 3 stars".

Similarly - if you train a player on individual things, then perhaps the extra training could cause a higher evaluation from the coaches.

That's what the problem really is - the potential is too accurately displayed in the game.

[Kriss]

That's what the problem really is - the potential is too accurately displayed in the game.

It isn't displayed in the game though;)

How scouts etc perceive potential could be a problem though.

[marty78]

That's what the problem really is - the potential is too accurately displayed in the game.

Yeah thats the problem for me too. The things is even if you have a 150PA player and give him a chance to increase that by 10% all things going in his favour he still has a PA of 165 max, thats is still a limit. You don't need a minus swing to PA as that is just equivilant to not reaching it as it stands.

@Kriss

The thing is if you have two players of similar standing and age with slight differences between potential you can't tell in game. Increase that gap and it is glaringly obvious. For me this problem is biggest with scouting and youth intake and it is easy to dismiss players at a glance without too many fears of misjudging.

[Guest]

But they cannot exceed the opinion of potential expressed as an input by a researcher. Yet in the real world, a real player can indeed exceed the opinion of potential expressed as an input by a researcher.

There is a real world-FM world mismatch that needs to be fixed.

This is where you're going wrong, at some point all games must leave behind real world logic, going back to your racing game analogies why are you still able to compete in the next race at a 300kph smash that in the real world could have resulted in broken bones, concussion or even death? Playability.

Taking the uncapped PA to its finite conclusion all we will end up with are FM'ers complaining that it is too easy to create a dream XI through focused & in the extreme exploitative training, who cares if Theo Walcott was given a 175PA in FM09 but in FM12 has a 135PA or Chris Smalling was rated at 115 but is now 160, once you load up the game you need to leave behind any real life development & the acceptance that potential is a subjective concept that can never be given an absolute value outside of a simulated environment.

Where FM falls down is on the mental aspect of the game & how players learn as they progress through their career, looking at pure technical ability players generally do not improve after their mid 20's but what does happen is they learn how to use the skills they have to provide more variation to their game, the true greats appear to do this with ease, good players work hard to achieve it & the majority fail to adapt. This is also the case as players start to decline physically, some will change the way they play to remain effective but in a different way whereas others will continue to decline to the point that they are no longer effective.

The problem with FM is that players do not change how they approach a game, in essence they do not learn & this is what needs to be addressed but it has nothing to do with a PA glass ceiling.

[x42bn6]

All attributes to infinity? Infinite pace? Faster than the speed of light? I reckon talking about infinities and the possibilities human evolution could bring is overthinking.

The question was around the 200 CA limit. In theory, yes, the limit should be infinity. It is just that football evolves very slowly - by the time Xavi's passing, say, would be considered 19 at his peak relative to the best player in the world, he would probably be retired. 1-200 is model scaling, not player development modelling.

As long as PA is hidden properly (which it is not ATM) then that is in line with real life more than unlimited potential.

No, it is not, since a researcher can get it wrong by underestimating a player, denying a player his correct real-life performances modelled in-game.

Ronaldo at São Cristóvão would never have got a PA in the 190s. In fact, if he did, he would have been smacked down by the Brazilian researcher at the time. Yet he became the world's best player.

I can understand the questioning of PA's of real players as they are guessed by someone, but I have fallen in line with the "FM is god" view with regards regens. As the creator the game has the divine right to set the parameters of the created.

Except that this isn't true at all. Regen PAs are generated via a random number generator based on things like training facilities and coaching quality. However, if any of these teams move to a league that is many levels above, their PAs do not get re-rated as they would have in reality.

An amateur side would never realistically wonder how their players would develop should they move to Barcelona, say. In fact, given it's so rare, it's probably not worth considering due to the lack of data. But if an amateur player does move to Barcelona, there would be some re-evaluation.

A regen player who is 120/120 and hitting 8.50 as his average rating has the exact same issue as a real player at 120/120 hitting 8.50. It is just that real players have the additional uncertainty of bad data input.

Even with most real players that are young enough to get a - rating there is an element of randomness.

That's funny because this implies that a researcher's rating cannot be trusted, but an unfalliable deity is willing to claim his knowledge is perfect based on imperfect mortal opinion.

But let's say a researcher assigns a player a PA rating of -9, implying a PA between 150-180. Why doesn't the researcher just give him a PA of 180? After all, that is his effective limit and implies perfect facilities, coaching, tutoring, personality and game performances. Why -9?

There is also an official editor should you wish to change the PA's of real players.

It doesn't help games that have already started.

I could also understand raising PA's accross the board but making it pretty much impossible to achieve full PA. I can't believe in no limit though.

It's more like "everyone has a limit, but it makes no sense to decide what it should be at 16".

Or "just because there is a limit does not mean we need to define one in-game."

Or "even if there was a limit, you can only guess what it is at 16. So why doesn't the game just treat it as a guess?"

[x42bn6]

This is where you're going wrong, at some point all games must leave behind real world logic, going back to your racing game analogies why are you still able to compete in the next race at a 300kph smash that in the real world could have resulted in broken bones, concussion or even death? Playability.

So it is a gameplay-enhancing experience for a potential wonderkid to be underestimated by a researcher and therefore you are unable to turn him into a world-class player in-game?

The more accurate analogy is putting the top speed of a Gold GTI as 150mph, then the top speed is stuck at 150mph when you go downhill. Because the circumstances in determining the top speed no longer apply.

Taking the uncapped PA to its finite conclusion all we will end up with are FM'ers complaining that it is too easy to create a dream XI through focused & in the extreme exploitative training,

Then the issue surely lies with exploitative training, no?

who cares if Theo Walcott was given a 175PA in FM09 but in FM12 has a 135PA or Chris Smalling was rated at 115 but is now 160, once you load up the game you need to leave behind any real life development

Why? It would be a better gameplay experience that improves immersion if we could mimic reality to a greater extent (at least within our parameters). Or even have "larger-than-life" experiences where Heskey leads England to World Cup glory.

& the acceptance that potential is a subjective concept that can never be given an absolute value.

So why is PA, in-game, an absolute concept with an absolute value?

Where FM falls down is on the mental aspect of the game & how players learn as they progress through their career, looking at pure technical ability players generally do not improve after their mid 20's but what does happen is they learn how to use what skills they have to provide more variation to their game, the true greats appear to do this with ease, good players work hard to achieve it & the majority fail to adapt. This is also the case as players start to decline physically, some will change the way they play to remain effective but in a different way whereas others will continue to decline to the point that they are no longer effective.

I don't see why we can't have this and a better PA system.

The problem with FM is that players do not change how they approach a game, in essence they do not learn & this is what needs to be addressed but it has nothing to do with a PA glass ceiling.

Again, we can have both.

[Guest]

So it is a gameplay-enhancing experience for a potential wonderkid to be underestimated by a researcher and therefore you are unable to turn him into a world-class player in-game?

It doesn't matter what rating the player was given by the researcher, at the precise moment in time that the data is entered it is correct & that is all that counts.

If a player develops beyond his expected ability so what, it's of zero value to FM because the game world needs fixed data points to work in a balanced manner.

So why is PA, in-game, an absolute concept with an absolute value?

The key phrase that you chose to delete from the quote was ..outside the simulated environment, this is crucial to my point.

I cannot reiterate this enough but you need to separate real world progression from game world progression, until you do I am done because it is quite clear that you are not open to any opinion other than your own.

[tylerazevedo]

I think we can all agree that it may be impossible to accurately represent potential in a game. Hell it's impossible to accurately predict potential consistently in real life.

I do think the game should be more stat based. Look at how OOTP gives star ratings to their players. It is like 40% attribute based & 60% stat based. It has to be a hybrid of those two. Right now, it's almost purely attribute based in FM.

As far as not showing any attributes, that's just bonkers. You have to be able to see attributes, but maybe not the exact attribute. Maybe a range for each attribute would be more appropriate.

[x42bn6]

It doesn't matter what rating the player was given by the researcher, at the precise moment in time that the data is entered it is correct

But it isn't! Stop calling a guess "correct" when it is just... A guess.

If the game assumes it's correct... The game is making an incorrect assumption. The last time I checked, I think that's called "a bug".

If a player develops beyond his expected ability so what, it's of zero value to FM because the game world needs fixed data points to work in a balanced manner.

Nope. Just as we don't need maximum scorelines to prevent 40-40 results. Just as a racing game doesn't need to store maximum speeds of its cars.

The key phrase that you chose not to quote was ..outside the simulated environment, this is crucial to my point.

If it can't be given outside a simulated environment, what are we doing putting it inside the simulated environment?

I cannot reiterate this enough but you need to separate real world progression from game world progression.

I don't see why we need to. The ideal game should be able to mimic reality. If it doesn't, then that's a problem with the game.

The game might not be able to predict the future but it should be able to make it happen anyway.

Imagine playing "Football Manager 1990" and being able to mimic Ronaldo's career, despite the inevitable low PA that would have been assigned by São Cristóvão's researcher.

A game that can do that would be a better game than not being able to mimic Ronaldo's career. If it is a better game, why shouldn't we push for it?

[tylerazevedo]

A game that can do that would be a better game than not being able to mimic Ronaldo's career. If it is a better game, why shouldn't we push for it?

The holy grail.....

[marty78]

A regen player who is 120/120 and hitting 8.50 as his average rating has the exact same issue as a real player at 120/120 hitting 8.50. It is just that real players have the additional uncertainty of bad data input.

I still don't understand why this is a problem for you. If his PA stars were hidden, as I've suggested, how would you know he has no potential left to fulfill?

That's funny because this implies that a researcher's rating cannot be trusted, but an unfalliable deity is willing to claim his knowledge is perfect based on imperfect mortal opinion.

But let's say a researcher assigns a player a PA rating of -9, implying a PA between 150-180. Why doesn't the researcher just give him a PA of 180? After all, that is his effective limit and implies perfect facilities, coaching, tutoring, personality and game performances. Why -9?

Has to be by definition. Omiscience would be necessary for an unfallible deity so it would know all that is knowable. Lets not go further down that road though and there is no reason to use racing simulators in the discussion either.

If they gave every player a set PA and the same starting attributes there would be little variation and the game would lose replayability to a point.

If a number shouldn't be set for PA should one be set for CA as that is also a parameter that can not be accurately defined? What about individual attributes should they be defined?

If you have a alternative system that could work and can explain exactly how that can be modelled in game I'm all ears.

[achilles-the-victorious]

What the hell is this kid still on about?

"Then the issue surely lies with exploitative training, no?"

If you design a game, and it has an exploit in it, is it the fault of the designers, or the exploiters? The designers! Thats why it is important for games not to be ruined by exploits. If you design a new PA system that is variable, then that means you have to make sure its not exploitable as well. Problem is, thats a LOT easier said than done. We talk about soak tests and gameplay balancing, etc. but the truth is, that takes time and money. A lot more of both than we commonly imagine, IMO.

People would try to figure out how the variable PA system works, what training, match experience, and personality is the ideal one, and there you go.

I've already given examples of scouts being wrong for YEARS in my game, and a player who exceeded his potential (for all intents and purposes, because I never checked to see what his potential was, I am saying he exceeded the potential my scouts consistently told me he has). Doesn't that count for anything?

[x42bn6]

I still don't understand why this is a problem for you. If his PA stars were hidden, as I've suggested, how would you know he has no potential left to fulfill?

Because if, say, he is 120/120 and getting an average rating of 8.50, he wouldn't be developing much, if at all. At least not as much as he should be in reality.

Has to be by definition. Omiscience would be necessary for an unfallible deity so it would know all that is knowable.

So this sounds like the game isn't perfect after all in this sense. Even more reason to improve the game.

Lets not go further down that road though and there is no reason to use racing simulators in the discussion either.

They're analogies and useful. This would also mean people would have to stop using the argument that the game is God.

If they gave every player a set PA and the same starting attributes there would be little variation and the game would lose replayability to a point.

I thought that there was some randomness involved where a player didn't reach their PA?

What's wrong with a little more randomness to allow players to exceed this expectation?

Maybe we could "transfer" some randomness on the negative end to the positive end by skewing the distribution a little.

If a number shouldn't be set for PA should one be set for CA as that is also a parameter that can not be accurately defined? What about individual attributes should they defined?

I don't think CA is required, as it is a weighted average.

I think attribute numbers are required because it just so happens that the best way to compare an attribute across two players is a number. It's also a sensible method if scaled properly.

If you have a alternative system that could work and can explain exactly how that can be modelled in game I'm all ears.

A few ways:

1) Re-rate PA at certain times in the season (Every season? Halfway through the season? Every month? Every day?)

2) Remove PA in favour of some "EPA" value that is a guideline (that can be exceeded or missed, possibly drastically)

And of course both solutions (and indeed, any solution not mentioned) will be appropriately balanced, and any exploits that might break this system will also be fixed.

I'm not sure if I should release another spreadsheet model. The last time I did, nobody downloaded it for days - I know this because I uploaded it on Rapidshare and I can track the download counts. Yet people were still debating me. People also giggled because it wasn't as sophisticated as SI's decades-long and worked-on-by-many-people Football Manager game (well, I'm sorry, but there's only so much I can do with Excel by myself).

[x42bn6]

What the hell is this kid still on about?

"Then the issue surely lies with exploitative training, no?"

If you design a game, and it has an exploit in it, is it the fault of the designers, or the exploiters? The designers! Thats why it is important for games not to be ruined by exploits. If you design a new PA system that is variable, then that means you have to make sure its not exploitable as well. Problem is, thats a LOT easier said than done. We talk about soak tests and gameplay balancing, etc. but the truth is, that takes time and money. A lot more of both than we commonly imagine, IMO.

Well then, maybe SI should get a move on and work a little harder?

Who cares how hard a feature is to implement, or a bug takes to fix?

And yes, the issue lies with the exploitative training (by definition of the word "exploit"). If the match-engine produces rubbish results due to some corner bug, it's the corner bug that is the issue, not the match-engine. Similarly, if a player develops really quickly because of some training exploit, then it's the training exploit that is the issue, not the match-engine.

People would try to figure out how the variable PA system works, what training, match experience, and personality is the ideal one, and there you go.

And they will quickly find out that the best way to develop talent is to pick the players with the most raw talent, have the best personalities, and have the best coaching facilities and training facilities as possible, and perhaps tutor the player appropriately.

... Oh! That sounds very much like real-life! Stupid reality, exploiting things.

I've already given examples of scouts being wrong for YEARS in my game, and a player who exceeded his potential (for all intents and purposes, because I never checked to see what his potential was, I am saying he exceeded the potential my scouts consistently told me he has). Doesn't that count for anything?

Say this player was 120/160 and your scouts said he was a 120/130 player. I'm not surprised he exceeded your scouts' opinions.

Now imagine if he was a 120/120 player. He wouldn't be able to exceed your scouts' opinions.

But a 120/120 player who is playing very well is still learning, and therefore should develop (as long as he is playing sufficiently well and he isn't too old, where the "decreasing pressure" will outweigh it).

[Kriss]

(well, I'm sorry, but there's only so much I can do with Excel by myself).

Should be using Access, no?

[marty78]

So if a player plays well because the user is tactically spot on in the use of him his PA would essentially continue to rise? Until what age and at what rate?

If the increases are varied and based on parameters other than ratings, how many parameters, what are they and how much processing is going to be required to re-rate periodically?

If the increases are very small then does a small increase really matter when a player is consistantly brilliant? If they can be large with consistent high ratings, do you think you can stop exploits without re-modelling training? Are any other areas going to have to be changed much to avoid exploits and balance the system?

Why would negative adjustments be required when it already exists that a player might not reach their potential?

How high should this 120/120 be able to go with superb ratings and no injuries in your opinion?

[achilles-the-victorious]

You are telling me you aren't surprised he surpassed my scouts expectations. My scouts earning $1m a year, who have scouted every player I signed for the last 10 years, who are right time and time again, but rate this kid as 2.5 stars potential, and you aren't surprised? I was very very pleasantly surprised to see him continue to match my expectations, while simultaneously exceeding the expectations of all my staff.

A player who is 120/120 will still develop. Some area of their game will have to be sacrificed, but they can still become more well-rounded, or more specialized. At 26 he will be a better player than at 22, despite the same CA/PA. He won't be 4 years of development better, though, he will be 4 years of experience better (meaning to me, better mentals, slightly lower physicals).

Who cares how hard a feature is to implement? Businessmen. Like the ones who manage SI, they care about things like expenses, revenues, and rates of return. When the rate of return is negative, believe me, you can desire a change as much as you want at that point, but that doesn't mean its going to happen.

Well then, maybe SI should get a move on and work a little harder?

Who cares how hard a feature is to implement, or a bug takes to fix?

Say this player was 120/160 and your scouts said he was a 120/130 player. I'm not surprised he exceeded your scouts' opinions.

Now imagine if he was a 120/120 player. He wouldn't be able to exceed your scouts' opinions.

But a 120/120 player who is playing very well is still learning, and therefore should develop (as long as he is playing sufficiently well and he isn't too old, where the "decreasing pressure" will outweigh it).

[hursty2]

I wonder what would happen if you have every player in game a PA of 200? Would the game become a superworld or still be balanced because of the effects that training, coaches, mentality, game time, reputation and club facilities have on a players CA?

[x42bn6]

Should be using Access, no?

It's not a database people are asking for - it's a mathematical model. Excel works there.

So if a player plays well because the user is tactically spot on in the use of him his PA would essentially continue to rise? Until what age and at what rate?

Likely indefinitely, but eventually the rate of growth will slow down as adults of course have less leeway to develop than a youngster or newborn child.

If the increases are varied and based on parameters other than ratings, how many parameters, what are they and how much processing is going to be required to re-rate periodically?

Premature optimisation is evil. It will likely require more processing power but then again why does that matter now? It's the ideas that matter, not the implementation.

If the increases are very small then does a small increase really matter when a player is consistantly brilliant?

They might not be small increases, but even if they are, there are lots of small increases if a player outperforms slightly, so they will add up.

Does it matter? Yes, since better players are treated more harshly by the opposition (tighter marking, more closing-down, etc.). Eventually, most players will fall flat as the opposition finally deals with them. However, the odd player like Lionel Messi defies the odds.

Of course, Messi is rare.

If they can be large with consistent high ratings, do you think you can stop exploits without re-modelling training?

Yes, if the system is well-designed and balanced.

Are any other areas going to have to be changed much to avoid exploits and balance the system?

Probably, but these all improve the game, no?

Why would negative adjustments be required when it already exists that a player might not reach their potential?

Because "EPA" or whatever it is called isn't defined as the limit. It's a moving target. Which is why I do question why we would need PA at all. But PA-re-rating sounds a lot easier to understand.

How high should this 120/120 be able to go with superb ratings and no injuries in your opinion?

As high as it deserves, depending on things like balancing and real analysis.

You are telling me you aren't surprised he surpassed my scouts expectations. My scouts earning $1m a year, who have scouted every player I signed for the last 10 years, who are right time and time again, but rate this kid as 2.5 stars potential, and you aren't surprised? I was very very pleasantly surprised to see him continue to match my expectations, while simultaneously exceeding the expectations of all my staff.

Good for you. But you've missed my point again, that there exists some player that simply cannot exceed your best scouts' opinions.

A player who is 120/120 will still develop. Some area of their game will have to be sacrificed

That's not development. That's rebalancing.

Real development doesn't necessarily sacrifice other areas you are good at.

, but they can still become more well-rounded, or more specialized.

How about just plain "better"?

At 26 he will be a better player than at 22, despite the same CA/PA.

Not that much better than if he had the spare CA to move.

He won't be 4 years of development better, though, he will be 4 years of experience better (meaning to me, better mentals, slightly lower physicals).

But inevitably hindered by his low CA-PA difference.

Who cares how hard a feature is to implement? Businessmen. Like the ones who manage SI, they care about things like expenses, revenues, and rates of return. When the rate of return is negative, believe me, you can desire a change as much as you want at that point, but that doesn't mean its going to happen.

And this matters to you, because...

You are a customer. You shouldn't care about how much feature X costs. If you want a superb game, demand it. That's not wrong. That's you trying to get value for money.

Do SI care? Probably. Boo hoo, poor SI, asking them to work harder.

I wonder what would happen if you have every player in game a PA of 200? Would the game become a superworld or still be balanced because of the effects that training, coaches, mentality, game time, reputation and club facilities have on a players CA?

Regarding "PA of 200": I'm assuming we do have some criterion (say "EPA") that distinguishes between two players' levels of "potential" or "talent". A bit like personality - two players have different personalities; two players have different levels of talent - it isn't just the case that we let PAs all become 200 and off we go.

It will be balanced because not every player would have access to the same level of facilities/coaches/game time/etc. For example, there are only so many first-team goalkeepers available to go round, so it makes no sense to say that every single goalkeeper will become a Casillas. That's the game time criterion covered. There are also only so many clubs with world-class training facilities in the world, and no club has an infinite wage budget. That's the facilities criterion covered. And so on.

[achilles-the-victorious]

I don't understand why you don't review and adjust your club and the opposition players PA at the end of a season. The tools exist. You know what you want, you know how to achieve it, what is stopping you?

[x42bn6]

I don't understand why you don't review and adjust your club and the opposition players PA at the end of a season. The tools exist. You know what you want, you know how to achieve it, what is stopping you?

1) You have to pay for FMRTE to fix your game

2) FMRTE is an external tool that risks ruining your game

3) It is tedious to fix players in your game. Do you really want to re-rate and edit every single player in the game?

4) Bias on your behalf. If I re-rated my game, it wouldn't be fair as I am not impartial nor consistent

[achilles-the-victorious]

Improve the PA of the top 2% of performers, or make a similar rule.

One time I edited a player who was injured on my team, a star. When my top opposition had their star get the same injury (broken foot) I healed him as well.

[x42bn6]

Improve the PA of the top 2% of performers, or make a similar rule.

One time I edited a player who was injured on my team, a star. When my top opposition had their star get the same injury (broken foot) I healed him as well.

2% of around 30,000 players is 600 edits. Do you really think this is a fun thing to do?

And it doesn't solve the paying issue, the external tool issue and bias.

[Guest]

Someone mentioned this earlier, it's worth considering the sort of processing/memory requirements such a fluid system that has a significant weighting based on form would require as full detail processing of all competitions is a must to provide balance to all player development, what spec PC/MAC would the consumer require? Obviously x42nb6 doesn't care about this as he's already made it quite clear elsewhere that based on SI's current method of distribution he's not going to be buying any future releases.

[x42bn6]

Someone mentioned this earlier, it's worth considering the sort of processing/memory requirements such a fluid system that has a significant weighting based on form would require as full detail processing of all competitions

Nope - something that is more lightweight can be employed for non-playable leagues in the same way that teams in non-playable leagues still simulate matches, likely in a more inaccurate way.

is a must to provide balance to all player development, what spec PC/MAC would the consumer require?

It would increase the processor requirements a tad. I don't think it's a massive deal, though, as Football Manager 2012 has a minimum system requirement of a Pentium 4/Intel Core/AMD Athlon.

It might increase the RAM requirements depending on the level of caching but that would be inversely-proportional to the processor speed increase and could well be a non-factor if Football Manager moves to a 64-bit game.

Also, I don't think FM really makes use of cores (based on various task manager screenshots on this forum) nor RAM (as it's still stuck at 32-bit) properly. I think there's a lot more juice that can be squeezed out of our processors yet.

But until a concrete model comes out, it's really not worth considering the system requirements. There may even be savings based on various optimisations that SI do in the background.

What's important is that people realise that there is a problem and that there are ways to fix it.

It then boils down to things like better 3D graphics, which are always, well, better (as the phrase suggests), where this is just evaluated against typical hardware requirements at the time. But most importantly - it's being planned to address a real problem.

[Guest]

The difference between sim'd & full detail processing is massive in terms of player performance & match rating, as an example when sim'd on 12.1.1 Messi (yes him again) would score around 40 goal a season whereas on full detail he was easily hitting 70 goals a season & in one test run he smashed in an average of 5 goals per game.

I have seen increases in individual & team match performances across my entire save (all leagues) when I changed from sim'd to full detail processing, if you're going to base your EPA around a principle that match performances drives improvements then you're going to have have all competitions processed in the same manner, if you leave some leagues to run on a sim'd basis the players based in those nations are going to be disadvantaged which will in all likelihood cause an imbalance in the game which in turn destroys the realism that you're striving for.

[x42bn6]

The difference between sim'd & full detail processing is massive in terms of player performance & match rating, as an example when sim'd on 12.1.1 Messi (yes him again) would score around 40 goal a season whereas on full detail he was easily hitting 70 goals a season & in one test run he smashed in an average of 5 goals per game.

I have seen increases in individual & team match performances across my entire save (all leagues) when I changed from sim'd to full detail processing, if you're going to base your EPA around a principle that match performances drives improvements then you're going to have have all competitions processed in the same manner, if you leave some leagues to run on a sim'd basis the players based in those nations are going to be disadvantaged which will in all likelihood cause an imbalance in the game which in turn destroys the realism that you're striving for.

Or SI could fix the root cause that is full detail processing being very different to view-only processing.

[Harper]

I think newgens should come in with a basic position (Goalie, Defense, Midfield, Forward) and leave it up to the manager to train their primary position.

[faraowww]

I´ve been reading this thread and I more or less share the opinion of x42bn6.

I would say PA is nonsense and not necessary. It is nonsense, because it is a numerical abstraction of some unknowable variable. Nobody knows a real life PA, so why define it? It is not necessary, because it can be replaced with a talent attribute including different learning curves.

It certainly doens't make sense that a player can't improve if he meets his PA. If I would play for 10 years in League 2, I probably won't improve anymore, so the game would say I have reached my PA. But if I would be transfered from league 2 to Man Utd at age 33, and play matches in the PL, I almost certainly would improve. Playing at a higher level, with better coaches and facilities makes you better.

So I would say we need talent attributes with learning curves. It should be possible to always improve as a player, only the probability of improving gets lower once you get older, but it should never be impossible to improve. Players with high talent should improve faster, so that they can become worldclass players. Other players lack talent to reach above league 2. And some players would have talent, but reach higher than expected, because they work hard, or have the luck to have good coaches and play at a high level.

The learning curves should be somewhat different, because some players improve very much the first few years, but stagnate, whereas other have a more gradual improvement rate.

The only 'problem' with talent and learning curves, is that is also depends on the judgment of SI researchers. But I would not say it is a real problem, compared to PA. PA is simply impossible to estimate, whereas SI could make mathematical models of player improvement and learning curves. Just look at how different players have developed through the years, or even look at the CA of previous FM games. It requires some thinking and maths, since there are no limits anymore, and it should be avoides that the game will have too much or too little world class players, but I think it is not too hard to create a balanced system with these principles. It is still possible to have an element of randomness in talent and learning curves, like we have negative PA's know. Another advantage, is that scouts in the game can't judge the PA anymore, which adds realism.

I don't agree with x42bn6 on the point that PA (or any other variable) should be adapted based by match ratings in the game, even if the match rating system would improve. First of all, with the system I just described, players can always improve if they play at a higher level or get good match ratings, so we don't need to adjust anything. Second, I think it mixed up real life and virtual life in a wrong way. FM-life (W') is a model of real life (W), so it makes sense to change the input of W' when players perform good in W. But it doesn't make sense to change the input W' based on W', since they are not models of each other, so it can't improve realism.

I also don't agree with Messi winning few genetic lotteries, because I think that is not true . Yes, he is small, but he has high agility and acceleration, which has a lot to do with genetics. But especially his brain is amazing, and that also has a lot to do with acceleration. His perception and translating spational awareness into action is very quick. If you stick out your right leg, he is already dribbling past your left leg before you can react. That is because his brain works quicker than ours. (Ofcourse he also had thousends of hours of practice, and that is also very important). Every physical movement based on visual input, relies very much on brain activity, and that has a lot to do with genetics. There are already neuro-scientist, who claim that we are going to use brain-scanners to determine football talent. This is a bit off-topic but interesting nonetheless.

[Jops14]

I think the best suggestion ive read came in the start of the thread. Which was my not give regens a -PA number until they hit 21/22.

So for example you have 2 -9 PA players come through in one season (i forget the range of -9 i think its 150-180??) The first one, A winger, starts with a CA of 84, whilst the other one a striker, comes in with a PA of 70. The striker due to coming in with a low CA isn't played much in the first team, he is tutored but left in the U18's only played 5 first team games before his 21st. Meanwhile the winger is used on the bench a few times and breaks into the first team by the age of 19. Due to these differing development rates lets say the striker at 21 has a CA of 118, whilst the winger has a CA of 160 (in the threshold). therefore the game calculates on their CA and mental attributes what their peak will be. The striker is given a peak of 155 whilst the winger comes in at 180. This allows for both players to develop further, whilst also giving them a peak limit of ability (like in real life). So as you would assume the winger goes on the become a world class players (what you would expect given his ability at 21) and continues to develop. Meanwhile the strikers get released and starts playing lower league football. But there is still the potential[\u] for him to develop further (say if he were picked up by a better team or started scoring prolifically) and be a late bloomer. But not able to become a star (as was never expected). I think the scope for development into 20's and the unknown quantity of youth is what everyone wants and is still a more flexible system than the current one.

I also partly feel CA should be allowed to fluctuate 5 points either way depending on form and morale. So a player in great form sees his finishing go up etc and plays beyond his capabilities for a short period of time. It would then perhaps reflect examples such as Ba and Torres this season. But to the people saying "my player only has 120 Pa but hes a leading premscorer and scores 30 goals a season" you could argue with "so?" I dont see the problem with having a player who may not be the most gifted individual, but is perfectly suited to their role.

[marty78]

If a players development is based on a learning curve then the highest point in that curve is still a limit of PA (or maximum achieveable CA if you like).

As you have already stated, that means researchers would have to guess at talent as they do with PA now and therefore the PA limit(highest point in the curve) is still being guessed at by a researcher.

At this point that just sounds like a more complicated version of the current system.