PA CA Again, 1 more solution

[robertcornell68]

"I'm not meant to see North Korea, therefore there is no problem with North Korea."

And that's just plain surreal.

[CrazyEastboy1]

Havent had the time to check the forums will do sometime tomorrow

[Riz Remes]

FM07's Pedro could not follow real-life, therefore PA is flawed, and therefore needs to be fixed.

You are taking a human evaluation that was wrong at the time (the PA in the DB, based on human evaluations in real life) and making it the fault of the system within the game. The system (the player development module in FM) would have allowed Pedro from FM07 to follow real-life, had he been rated 100% correctly in the DB with a higher potential. Which in most cases is simply impossible due to humans occasionally not being able to accurately estimate the exact PA nevermind individual attributes. Maybe he should have had a higher PA but some low mental attributes to counter the high PA or simply a low CA, to make it possible for the manager in the game to tutor him and turn him into a star ? Looking back at players modelled in the past versions of the game is a bit odd way to approach the subject of CA/PA

However, a newgen created by the game, in the game, looked at in the gameworld only with the scouting reports and coach reports available to the user would easily be able to replicate the development of Pedro. All it takes is the suitable PA, some non-exact scouting or missing mental attributes and possibly some good tutoring.

[milnerpoint]

i think thats a very good point, even if there was an unlimited PA or a variable PA, instances like Pedro would have still happened because of the inability of researchers to see into the future.

As i stated before there would have to be a restriction somewhere that would prevent every player becoming world class, be it personality or anything else possible, so if this was not acurate in the first place the unlimited PA would be worthless, and you still would not have a Pedro situation.

[Suge]

Thing is with the current system, the potential ability of players that exist in the game at the start are based entirely on human interpretation. This allows massive amounts of undervaluation to occur like we saw with Pedro in 2007.

The revamped system outlined in the OP would at least go some way to give every player a level playing field and totally eradicate the need to draw a crystal ball and guess the future.

I'm all for change.

[milnerpoint]

I think the big thing the system works very well for newgens, it is very possible to take an average looking newgen and turn them into something unexpected because you have no frame of reference to compare the player too, wheras with real players we see their actual futures pan out in front of us whilst then comparing their current ability to what was rated in earlier versions, which then leads to people questioning the system.

As Suge says all players that start off in the database are rated on human assumptions, and some researchers will be better than others, hence some players may get attributes not quite fitting their ability, this is no critisism of the spanish/barca researcher, its impossible to predict exactly how the future of a player in real life will turn out and as such no matter what system is in place, there will always be mistakes and players who do not pan out in the game as they have done in real life, because each game runs on a players guessed attributes at one given point in time. With newgens there is no preconception of what the media have said about him to influence your opinion on his future so you can develop them happily without watching them on TV and thinking "why isnt he as good on FM as this".

[Martyr1777]

You are taking a human evaluation that was wrong at the time (the PA in the DB, based on human evaluations in real life) and making it the fault of the system within the game. The system (the player development module in FM) would have allowed Pedro from FM07 to follow real-life, had he been rated 100% correctly in the DB with a higher potential. Which in most cases is simply impossible due to humans occasionally not being able to accurately estimate the exact PA nevermind individual attributes. Maybe he should have had a higher PA but some low mental attributes to counter the high PA or simply a low CA, to make it possible for the manager in the game to tutor him and turn him into a star ? Looking back at players modelled in the past versions of the game is a bit odd way to approach the subject of CA/PA

However, a newgen created by the game, in the game, looked at in the gameworld only with the scouting reports and coach reports available to the user would easily be able to replicate the development of Pedro. All it takes is the suitable PA, some non-exact scouting or missing mental attributes and possibly some good tutoring.

This is the heart of how the argument has changed since yesterday... People are focusing on real players and how they were limited in game and therefore didn't turn out as good as they did IRL.

That's a 100% completely flawed way of judging the system however because that's based on research and the database. it's a judgment error in translating a real player into the game and has nothing to do with the PA mechanic itself. I mean if you want to talk about real players not matching up with what we see in game there is a striker in the MLS that's I've heard referred to as the best striker in the league, yet in game aside from his pace he's barely even average. That goes to the same thing as claiming real players aren't reaching their real potential. It's not because of the PA system, it's just when they set that players PA they didn't think it would be that high.

If you really honestly want to judge the PA system you can -NOT- use comparisons for real players. You have to consider how newgens are created and how the system works in those cases. If you are point at real players it's basically just whining 'WAA!! so and so didn't have a high enough limit like in irl!'. The limit in those cases was -set- by the database team. But again, if you don't set a limit there is no balancing system and I have yet to see someone explain a balancing system without limitation.

Easier yes, but why give SI easy tasks? It's their job to give us customers what we want.

Uhh, because there a more important things that are actually broken, so why waste time trying to over complicate a system that works when you know the end result of that over complication is going to cause bugs that are far worse then any perceived issues with the current system. Also remember when you consider this statement just how critical I've been of SI and the game in the past. If I honestly thought there was any merit at all to changing the CA/PA system I'd be agreeing. But it's one of the areas that doesn't have a problem imo.

[x42bn6]

i dont really care what player you will never be as good as or worse than, my point was you and me will never ever in our wildest dreams get even close to the level of Messi regardless or anything, it just will not happen, forget about the probability of it, the reality is it will never happen.

That's fine - if you can't pick a name out, then I'm guessing you "can't" know. That's really the point - it is impossible to put your finger on it, so the fairest way is to assume it is possible. "Never" is too definitive.

Ok so name the conditions, and outline what will stop the big clubs all constantly developing world class players in every position, because this is the big point i am struggling with, where are you drawing the line, what stops every player at good clubs becoming top players, except player number restrictions.

Is that fair?

I'll humour you. In order to develop quickly, a player must train and play well, especially with the first-team. We all know what could potentially happen if you chucked a bunch of 16-year-olds into your first-team without blooding them properly - they aren't ready and pressure gets to them, and they don't perform. This breaks one of the requirements (a player must play well) - hence they don't develop. In other words, they have to be ready.

So how do you get them ready? They need to develop well in your youth teams. Again, sometimes players fall here - some don't adapt, some have pathetic attitudes, and some are simply stuck behind other players.

Also, some players simply aren't talented, and one idea is that less-talented players are more likely to develop slowly, in a similar way to how some people gain knowledge easily, some need to work to gain knowledge, some need to work very hard to gain knowledge, and some don't bother at all. A less-talented player is, if you like, handicapped by the fact he's not that talented - but can of course overcome this if he performs well for the youth teams and first-team, and/or trains well.

Of course, things like injuries play a part - suffer a serious injury and your development will likely slip away. Recurring injuries will also play a part.

It certainly doesn't guarantee things but then again there is no point in considering the most unlikely outcome and assuming this will be the thing to judge it against. An extreme scenario for the current system would be that no player develops at all - every single player suffers recurring cruciate injuries and all games are played with grey players. Worth considering? Probably not. Worth using to bash the current system? Of course not.

But they have limitations... Huge ones in Pepe's case... No amount of training or playing time will ever turn them into something they aren't and they can't ever be.

They're the typical "better than nothing" second-choices

You never know, you never know...

Of course, I'm not going to bet on Pepe turning out to be the next Ronaldo any time soon, but you can never really rule it out - it's just overwhelmingly unlikely.

Nobody expected Luca Toni to go on a late spurt, especially after he had arguably "peaked".

So no, you never know... Pepe might go on to do something similar.

But an "average" EPL guy is still better than an "average" League One player, who's better than an average BSN player, who'll probably look like Messi if comapred to us amateur players...

We aren't "all created equals"! And the game should still reflect that one way or another.

We aren't all created equal, but no PA doesn't imply that this will happen.

It doesn't give us all the same level of opportunity - a BSN-level player will almost never play a part in a Premier League side (but there are notable exceptions - Smalling, Kightly, Bullard...), but under the assumption that they do, they should not be restricted to only their initial beginnings.

Still... plenty of La Masia graduates play in Spain Level 3, some even had a solid chance at Barça but never made it...

How do you explain Giovani Dos Santos then? Or, moving to Madrid, the likes of Portillo and Pavón?

A player can come though a prestigious academy, get adequate first-team football and STILL not turn out as good...

A system with no PA does not imply this!

They still have many hoops to jump through to succeed. It is just that one of the hoops - the initial starting opinion of PA - is removed - for reasons I have mentioned again and again as being unrealistic.

And there's the point you're missing!

Flexible PA is not a PREDICTION of the future, it's a reassessment of a previous prediction, and it's based on in-game events and "facts".

The only "prediction" involved is the one at the start of the game, but thanks to flexible PA you can still have all the Drogbas and the Pedros you want, while not having to deal with nonsense like "anyone can become good".

A 19yo BSP goalie can get "better" if he plays and trains consistently well, but he'll NOT be better than Hart, or even than Green.

We are arguably talking about the same thing.

Researchers "predict" the future by assigning a PA. Over time, this PA is looked at and likely adjusted. As more and more information comes in, it allows them to make more-informed adjustments about a player's PA. PA is an opinion.

A mechanism that predicts future (flexible PA) renews PA over time as information comes in. Over time, this PA is looked at and likely adjusted. As more and more information comes in, it allows the engine to make more-informed adjustments about a player's PA.

The difference is that a computer doesn't "think" the same way as us in the sense that it lacks human opinion.

However, PA is the translation of our opinions into the game. A player that does well could be bumped up from a -8 to a -9, for example - this is based on how he trains, the chances he has got in the first-team and how he has done, amongst other things. These are objective quantities that will surely go into your objective flexible PA engine. The game will take into account these (virtual) things and reevaluate the PA; a researcher will take into account these (real-world) things and reevaluate the PA. They are essentially the same thing.

Have you heard of Brownian motion? A gas particle, if suspended in some medium, will move in a random process yet its mean will be concentrated about its starting position. If the medium is an infinitely-large tank, it will still vibrate in this manner. An infinite set of possible outcomes, no limitations, yet a predictable outcome. This is why "nonsense like "anyone can become good"" is not nonsense - no limits, predictable outcome.

That could happen as well with flexible PA!!!

The unknown Luis Nazario da Lima would have started with a rather low PA, then after his breakthrough at Cruzeiro it would have risen to a decent "Brazilian youngster" level, then after the first season at PSV it would have gone up even more, and the stellar season at Barça would have just sealed the deal...

1990: PA 100

1993: PA 140

1995: PA 160

1997: PA 190

It doesn't look that complicated or outlandish IMO...

Surely easier than a messy "free for all" system where any player can turn into a Gary Cahill just by piling up appearences at decent clubs.

The main difference between my idea and yours is that yours requires an explicit calculation of PA. Your system implies that anyone can become Messi too, after all - just do a Ronaldo.

The reason why you accept PA being reevaluated is because PA may be "wrong" and needs to be reevaluated over time, based on newer information. In other words:

- If a player does well, then PA goes up;

- If a player does badly, then PA goes down.

Mine is:

- If a player does well, then he does well;

- If a player does badly, then he does badly.

Having PA moving around is essentially no different to no PA - there does not exceed a definite number which I can say, "THIS is DEFINITELY the ABSOLUTE ABSOLUTE DEFINITE maximum PA value for this player" (unless he has peaked, of course, bearing in mind the Luca Tonis of this world). If at some time a player's PA is 130, then a player can exceed PA 130 in some time in the future - it is no different.

All flexible PA says is that "This is his expected PA at his peak. Subject to change." It's not a limit - it's an expected PA. An educated guess. A bit like predicting the number of runs in a one-day cricket match by the run rate - if India are averaging 6 runs per over, then in 40 overs they will have 240 more runs. On the 50th and final over, the run rate is quite a good guess at the final score.

So "no PA" is basically saying "India have a run-rate of 6 runs per over. The final score can be computed, but it doesn't actually need to be stored anywhere - it can be calculated if you want. Multiply 6 by 40."

That is it! Essentially no difference at all.

This is a bit silly really. CA/PA is fine. Why is it not?

And no, I don't want a pseudo-philosophical answer about the nature of potential.

So I'm sorry, you want to know why it is wrong, but don't want to know why it is wrong?

No game is perfect. The most perfect system in the world cannot exist, since it would be equivalent to life itself, which is of course not possible. Therefore CA/PA is an imperfection, and it so-happens some people have ideas on how to improve on it.

Potential is set by it's defenition. It's the maximum a individual can ever become. Whether a person reaches that potenial is determined by a multitude of reasons, many beyond our control. This is portayed in the game by injuries, personality, training, tactics and the people around the player.

I know what PA is, but it is wrong to assume that the initial PA is correct, correct and correct.

How good a player becomes is also dependent on how he does throughout his career - not just his initial starting guess.

Since development determines how good a player will become, it makes no sense to say, "the starting guess is correct throughout is career" - it is wrong.

You are taking a human evaluation that was wrong at the time (the PA in the DB, based on human evaluations in real life) and making it the fault of the system within the game.

The game can compensate for mistakes made by users in the same way that if a researcher in the Championship accidentally added an extra zero for the stadium capacity, the club wouldn't be able to afford the maintenence and would shrink the stadium. Maybe not to a perfect level of that of the stadium's original size, but it can compensate.

The system (the player development module in FM) would have allowed Pedro from FM07 to follow real-life, had he been rated 100% correctly in the DB with a higher potential. Which in most cases is simply impossible due to humans occasionally not being able to accurately estimate the exact PA nevermind individual attributes. Maybe he should have had a higher PA but some low mental attributes to counter the high PA or simply a low CA, to make it possible for the manager in the game to tutor him and turn him into a star ? Looking back at players modelled in the past versions of the game is a bit odd way to approach the subject of CA/PA

I don't think so. It illustrates my point perfectly - over research phases, attributes, CA and PA change. Pedro's PA was wrong in FM07 - it was corrected over time (it could still be arguably wrong, but it is less wrong than before).

Therefore for regens, over time, attributes, CA and PA should change.

However, a newgen created by the game, in the game, looked at in the gameworld only with the scouting reports and coach reports available to the user would easily be able to replicate the development of Pedro. All it takes is the suitable PA, some non-exact scouting or missing mental attributes and possibly some good tutoring.

Er, no. If a regen was equal to Pedro in FM07, he would not be anything like Pedro in real-life 4-5 years on. Pedro's PA was not suitable to begin with.

This is what I mean by "the system is perfect IF PA is high enough to begin with." Sometimes, it simply isn't - because researchers can (wildly) underestimate a player.

[x42bn6]

This is the heart of how the argument has changed since yesterday... People are focusing on real players and how they were limited in game and therefore didn't turn out as good as they did IRL.

That's a 100% completely flawed way of judging the system however because that's based on research and the database. it's a judgment error in translating a real player into the game and has nothing to do with the PA mechanic itself.

No, it is not.

PA changes over research phases because researchers can make better decisions about their players' PA. Do well, PA goes up. Do badly, PA goes down.

This is the heart of RBKalle's idea around flexible PA - essentially, the game is replicating the researcher every season.

I mean if you want to talk about real players not matching up with what we see in game there is a striker in the MLS that's I've heard referred to as the best striker in the league, yet in game aside from his pace he's barely even average. That goes to the same thing as claiming real players aren't reaching their real potential. It's not because of the PA system, it's just when they set that players PA they didn't think it would be that high.

If you really honestly want to judge the PA system you can -NOT- use comparisons for real players. You have to consider how newgens are created and how the system works in those cases. If you are point at real players it's basically just whining 'WAA!! so and so didn't have a high enough limit like in irl!'. The limit in those cases was -set- by the database team. But again, if you don't set a limit there is no balancing system and I have yet to see someone explain a balancing system without limitation.

See my post.

Uhh, because there a more important things that are actually broken, so why waste time trying to over complicate a system that works when you know the end result of that over complication is going to cause bugs that are far worse then any perceived issues with the current system. Also remember when you consider this statement just how critical I've been of SI and the game in the past. If I honestly thought there was any merit at all to changing the CA/PA system I'd be agreeing. But it's one of the areas that doesn't have a problem imo.

This is called issue prioritisation - but issues cannot be prioritised if issues are not brought up.

I think everyone wants SI to improve the development model - just not at the expense of silly bugs like stats not showing up properly on profiles or too many corners converted. That's fine. But I would still like the issues to be put on the board for SI to look at.

I've mentioned why it's broken - please critique.

[milnerpoint]

thats not what he means, what he means is it is easy to recreate a late bloomer newgen in FM, just like Pedro in real life, a player that out of the blue gets a chance and takes it in a big team, he didnt say a carbon copy of pedro's profile.

The more and more i read your posts on this it seems like the issue you have is more to do with the research rather than the way the CA/PA system works, if every single player was acturately rated, with people having the ability to see into the future then there would be no issues really, everything would happen as we expect it too, however this is impossible, with any system in place its impossible for every player to develop or not develop as in real life, but with game generated players the system is perfect, we should have no idea of what a PA number of a player is and as such we can try and develop these players without the need to compare them to real life.

[Martyr1777]

Is that fair?

I'll humour you. In order to develop quickly, a player must train and play well, especially with the first-team. We all know what could potentially happen if you chucked a bunch of 16-year-olds into your first-team without blooding them properly - they aren't ready and pressure gets to them, and they don't perform. This breaks one of the requirements (a player must play well) - hence they don't develop. In other words, they have to be ready.

So how do you get them ready? They need to develop well in your youth teams. Again, sometimes players fall here - some don't adapt, some have pathetic attitudes, and some are simply stuck behind other players.

Also, some players simply aren't talented, and one idea is that less-talented players are more likely to develop slowly, in a similar way to how some people gain knowledge easily, some need to work to gain knowledge, some need to work very hard to gain knowledge, and some don't bother at all. A less-talented player is, if you like, handicapped by the fact he's not that talented - but can of course overcome this if he performs well for the youth teams and first-team, and/or trains well.

Of course, things like injuries play a part - suffer a serious injury and your development will likely slip away. Recurring injuries will also play a part.

It certainly doesn't guarantee things but then again there is no point in considering the most unlikely outcome and assuming this will be the thing to judge it against. An extreme scenario for the current system would be that no player develops at all - every single player suffers recurring cruciate injuries and all games are played with grey players. Worth considering? Probably not. Worth using to bash the current system? Of course not.

This is no way explains how an unlimited system will have any balance at all. It just says there are lots of chances for bad luck to prevent a world class player. As for why the top clubs should be balanced as you've asked before. Well how many of the kids from top club youth teams become top world stars? Also how many world class players come from smaller clubs as a kid?

You are asking for a system with no checks and balances which leads to a complete stupidity in a game. There -HAVE- to be limitations, no if's and's or but's. Games that don't balance things fail because they are too easily exploited or are simply boring either because of being too easy or too hard. You are asking for the easy road, like the youth players that have low PA are an annoyance to you, but it's realistic.

Edit:

I've mentioned why it's broken - please critique.

You've talked predominately about how PA in-game doesn't match real life. Well CA doesn't match real life in many cases either, simply because it's a judgment made by the researchers and DB team. it has NOTHING to do with the development mechanic at all. Again, if you completely cut all thought of real players out of consideration and only consider how newgens work... where exactly does the system fail? Oh yeah, your other failure point is every player can't be a super star... That's no argument.

[Biscotti]

Er, no. If a regen was equal to Pedro in FM07, he would not be anything like Pedro in real-life 4-5 years on. Pedro's PA was not suitable to begin with.

This is what I mean by "the system is perfect IF PA is high enough to begin with." Sometimes, it simply isn't - because researchers can (wildly) underestimate a player.

You're not talking about the same thing as Riz in this case.

Riz is talking about a newgen now, you are talking about a newgen in a 4-year old version of the game based on a player that has already been acknowledged as inaccurately researched at the time.

Also, you seem to be taking serious issue with the PA of already existing players, which is not an issue with the PA system, but most definitely an issue with the research involved.

[milnerpoint]

.

Is that fair?

I'll humour you. In order to develop quickly, a player must train and play well, especially with the first-team. We all know what could potentially happen if you chucked a bunch of 16-year-olds into your first-team without blooding them properly - they aren't ready and pressure gets to them, and they don't perform. This breaks one of the requirements (a player must play well) - hence they don't develop. In other words, they have to be ready.

So how do you get them ready? They need to develop well in your youth teams. Again, sometimes players fall here - some don't adapt, some have pathetic attitudes, and some are simply stuck behind other players.

Also, some players simply aren't talented, and one idea is that less-talented players are more likely to develop slowly, in a similar way to how some people gain knowledge easily, some need to work to gain knowledge, some need to work very hard to gain knowledge, and some don't bother at all. A less-talented player is, if you like, handicapped by the fact he's not that talented - but can of course overcome this if he performs well for the youth teams and first-team, and/or trains well.

Of course, things like injuries play a part - suffer a serious injury and your development will likely slip away. Recurring injuries will also play a part.

It certainly doesn't guarantee things but then again there is no point in considering the most unlikely outcome and assuming this will be the thing to judge it against. An extreme scenario for the current system would be that no player develops at all - every single player suffers recurring cruciate injuries and all games are played with grey players. Worth considering? Probably not. Worth using to bash the current system? Of course not.

i never ment it as a dig or a go, just actually want to know what you would consider as restrictions. The thing is yet again it would back to the point that even with that all in place, there would still be human error because people cannot get everyones personalities exactly spot on, people also can change who they are and change their attitude quickly in life, so you may still end up with a pedro situation, where even with an unlimited PA his mental stats were not deemed good enough at the time the data was entered, so he will never reach the level we expect him too.

Until the human element of research is completely removed there is no way for everything to be perfect in terms of real players, there really just isnt, no way to account for all possible senario's, each system that has been mentioned has its upsides and down sides but frankly none are a perfect solution, at least the one in place now has been worked on for years and years and SI have a good grasp on how it works and what its potentials are.

[Cougar2010]

What it all boils down to is a human researchers inability to see the future.

Lets take Pedro as he is being used as an example, you only have an issue with his PA now in 2011 back in 2007 when he had a much lower PA it was never a problem because in 2007 people didn't expect him being as good as he is four years later. There wasn't an outcry about his PA on the forum, there weren't 1,000,001 threads created saying why is his PA not higher because its only an issue now. But since 2007 his PA has been reviewed upwards on an annual basis by the researchers increasing with each FM released.

This is only a problem with RL players as the game functions perfectly well with newgens.

TBH I would quite happily be open to all players say U25 having a random PA assigned when starting a save providing it was always =/> than their current CA.

[x42bn6]

thats not what he means, what he means is it is easy to recreate a late bloomer newgen in FM, just like Pedro in real life, a player that out of the blue gets a chance and takes it in a big team, he didnt say a carbon copy of pedro's profile.

So Riz has just stated the obvious?

There are a specific set of circumstances that allow late-bloomers and one of those is that PA must be high enough. That's right - the initial PA, generated from a random variable that ONLY takes into account the circumstances where the player was generated from initially, determines whether you are a late-bloomer.

Has nobody wondered why Luca Toni was a late bloomer? It's not because he started his career in Modena! It's because he found a specific set of circumstances that allowed him to consistently outperform what Modena expected of him when he was a little lad at 16. The fact he was a late-bloomer has nothing to do with his perceived potential a year before - it has everything to do with his current circumstances!

The more and more i read your posts on this it seems like the issue you have is more to do with the research rather than the way the CA/PA system works,if every single player was acturately rated, with people having the ability to see into the future then there would be no issues really, everything would happen as we expect it too, however this is impossible, with any system in place its impossible for every player to develop or not develop as in real life, but with game generated players the system is perfect, we should have no idea of what a PA number of a player is and as such we can try and develop these players without the need to compare them to real life.

I hold nothing against the researchers (except a certain person who rated Babacar) - let's get that straight first.

We shouldn't have an idea of what a player's PA is. This doesn't mean that there is no issue with PA.

It is almost surely impossible for a game to mirror real-life exactly, and it would be boring if it did - but the fact is that with PA, errors are introduced. PA basically says, "this player will never exceed X" - which, by the way, might actually happen in real life. Look at PAs getting increased!

This is no way explains how an unlimited system will have any balance at all. It just says there are lots of chances for bad luck to prevent a world class player. As for why the top clubs should be balanced as you've asked before. Well how many of the kids from top club youth teams become top world stars? Also how many world class players come from smaller clubs as a kid?

You are asking for a system with no checks and balances which leads to a complete stupidity in a game. There -HAVE- to be limitations, no if's and's or but's. Games that don't balance things fail because they are too easily exploited or are simply boring either because of being too easy or too hard. You are asking for the easy road, like the youth players that have low PA are an annoyance to you, but it's realistic.

Let's play an arcade game. The game is as follows: A machine flips a 1p coin - if it is heads, you win the coin, and the machine flips another one. This repeats until the coin flip is tails, in which case the machine stops. You pay 5p to play the game. Also say that the machine is directly hooked-up to the arcade's cash vault and can draw an infinite overdraft if necessary.

You run this arcade.

Say one day your boss asks, "Can you guarantee this machine won't bankrupt me?"

You can't guarantee him, no. There could exist some possibility that the machine flips several million heads, resulting in a massive loss for the company.

So why is this machine making him money? In fact, on average, each play earns a profit of 3p!

An infinite process - no limits (no maximum number of coins) - predictable outcome. The worst-case scenario is terrible, but not worth considering. The average-case scenario is a profit.

Would you put this machine in an arcade? Hell yes, of course you would..

Edit:

You've talked predominately about how PA in-game doesn't match real life. Well CA doesn't match real life in many cases either, simply because it's a judgment made by the researchers and DB team.

Two wrongs don't make a right. CA should match in real-life and in virtual-land.

it has NOTHING to do with the development mechanic at all.

Wrong, because PA has flaws, and researchers input flawed logic - everything to do with the game engine.

Again, if you completely cut all thought of real players out of consideration and only consider how newgens work... where exactly does the system fail?

It fails because a player with a low PA can never late-bloom, and that PA fails to take into account recent developments.

Oh yeah, your other failure point is every player can't be a super star... That's no argument.

I never implied otherwise.

You're not talking about the same thing as Riz in this case.

Riz is talking about a newgen now, you are talking about a newgen in a 4-year old version of the game based on a player that has already been acknowledged as inaccurately researched at the time.

Also, you seem to be taking serious issue with the PA of already existing players, which is not an issue with the PA system, but most definitely an issue with the research involved.

See above - you are basically saying Riz is stating the obvious (late-bloomers are possible under certain conditions) - I say one of those conditions is erroneous (see Luca Toni and Modena).

I am not going to bash the Barcelona researcher for getting Pedro wrong in FM07 - nobody knew any better.

What I do think is that the game can compensate for it anyway. Low PA, player still performs brilliantly in the first-team and has a good attitude => development that exceeds expectations - i.e. PA.

[x42bn6]

i never ment it as a dig or a go, just actually want to know what you would consider as restrictions. The thing is yet again it would back to the point that even with that all in place, there would still be human error because people cannot get everyones personalities exactly spot on, people also can change who they are and change their attitude quickly in life, so you may still end up with a pedro situation, where even with an unlimited PA his mental stats were not deemed good enough at the time the data was entered, so he will never reach the level we expect him too.

No, but thankfully, we don't have "potential passing", for example, so even if the attributes are wrong, it can be made up somewhat by the engine. This isn't the case for CA and PA - get PA wrong by underestimating? Oops...

Until the human element of research is completely removed there is no way for everything to be perfect in terms of real players, there really just isnt, no way to account for all possible senario's, each system that has been mentioned has its upsides and down sides but frankly none are a perfect solution, at least the one in place now has been worked on for years and years and SI have a good grasp on how it works and what its potentials are.

The human bit will never be removed.

Humans make mistakes. Computers can compensate for it. Look at data sanitation, error analysis and convergence. Look at the normal distribution, and outlier detection.

This is only a problem with RL players as the game functions perfectly well with newgens.

No, it doesn't - you still need high PA to late-bloom.

- If a real-life player has a low PA, he cannot be a late-bloomer.

- If a regen player has a low PA, he cannot be a late-bloomer.

The problem isn't fixed for regens.

TBH I would quite happily be open to all players say U25 having a random PA assigned when starting a save providing it was always =/> than their current CA.

More arbitrary limits... Where will it all end?

[Paoloboy]

I would love that SI would implement a possibility to choose random potential for all the players. Eg. player with fixed potential of 170 would have -7 and player with fixed potential of 160 to have -6 etc. (maybe also tune the range where negative potentials can hit). This would create a lot of variation for us who like to play several saves. AS said this could be an option like "attribute masking" and would be very easy to implement programming wise.

[Cougar2010]

No, it doesn't - you still need high PA to late-bloom.

- If a real-life player has a low PA, he cannot be a late-bloomer.

- If a regen player has a low PA, he cannot be a late-bloomer.

The problem isn't fixed for regens.

Do you get newgens with high PA - Yes

Can he become a late bloomer - Yes if his CA doesn't rise too quickly in the early stages of his career.

Conclusion: The problem is with CA development, not PA.

I really am lost for words with you on this issue x42. Last time you claimed the game shouldn't be acting as god when creating newgens. The fact is life has limits, you just don't want to accept that.

[x42bn6]

Do you get newgens with high PA - Yes

Can he become a late bloomer - Yes if his CA doesn't rise too quickly in the early stages of his career.

Conclusion: The problem is with CA development, not PA.

What if a player has a low PA, but his CA rises quickly anyway, he stagnates and then late-blooms?

Not possible with PA.

I really am lost for words with you on this issue x42. Last time you claimed the game shouldn't be acting as god when creating newgens. The fact is life has limits, you just don't want to accept that.

Oh, life has limits, but considering Usain Bolt isn't a model for 20 pace and acceleration, or the world's strongest man isn't a model for 20 strength, I'm not sure that really matters, to be honest.

Life has limits, but nobody knows what they are today - you'll only know your maximum when you, well, hit your maximum. Until then, you should leave all doors open - although perhaps some doors more open than others.

[Biscotti]

No, it doesn't - you still need high PA to late-bloom.

- If a real-life player has a low PA, he cannot be a late-bloomer.

- If a regen player has a low PA, he cannot be a late-bloomer.

The problem isn't fixed for regens.

The low PA newgen can't late bloom because he's not capable of that level of ability. A high PA newgen is still entirely capable of late-blooming. If a real-life player has low PA and late-blooms in real life, then the researcher was wrong, not the game.

You are still basing your assessment on a player that was judged by a human when it later turned out that human was wrong. With newgen players, that is a physical impossibility, because there is no human involvement in the process. Therefore, the issues you seem to have with the system are only relating to real-life players and the research surrounding them.

[Martyr1777]

Let's play an arcade game. The game is as follows: A machine flips a 1p coin - if it is heads, you win the coin, and the machine flips another one. This repeats until the coin flip is tails, in which case the machine stops. You pay 5p to play the game. Also say that the machine is directly hooked-up to the arcade's cash vault and can draw an infinite overdraft if necessary.

You run this arcade.

Say one day your boss asks, "Can you guarantee this machine won't bankrupt me?"

You can't guarantee him, no. There could exist some possibility that the machine flips several million heads, resulting in a massive loss for the company.

So why is this machine making him money? In fact, on average, each play earns a profit of 3p!

An infinite process - no limits (no maximum number of coins) - predictable outcome. The worst-case scenario is terrible, but not worth considering. The average-case scenario is a profit.

Would you put this machine in an arcade? Hell yes, of course you would..

That's basically a slot machine and only people that want the 'thrill' of gambling but don't know how to play real casino games, either by ignorance or lack of desire, play them. Hardly an example to use to ignore balance in a game as deep as FM. After all if everything in FM came down to luck I wouldn't touch the thing, nor would many other I bet. You need depth and balance to that depth. Just figuratively flipping a bunch of coins when it comes to developing youth in-game is a joke and not credible game design.

Two wrongs don't make a right. CA should match in real-life and in virtual-land.

Well if it's so easy to do perfectly without making mistakes then you do it. To be that critical of how real players are represented is just stupid because everyone has their own opinion of who's faster, who's more skilled, who's a better student of the game, etc. You can't fault the game over something that's basically 100% subjective. I'd be just as happy without real players in my game, you seem to not be able to separate yourself from them.

It fails because a player with a low PA can never late-bloom, and that PA fails to take into account recent developments.

You mean like a player with loads of potential but at a weaker club with poor facilities/coaches, then moving up to a larger club and coming into that potential when he has the tools at hand? Perfect example right there. Or a player with loads of potential and at a top club but has his youth career hampered by injury.

You are grasping at straws and your entire thought process is flawed. You completely discount the need for limitations and balance, you are basing everything on the fact that real players are different then in-game, and you are ignoring the fact that the system does indeed work you just want it to work differently.

[milnerpoint]

What if a player has a low PA, but his CA rises quickly anyway, he stagnates and then late-blooms?

Not possible with PA.

but that is a situation that happens in real life, Reo-Coker is a good example, was very good when he was younger and stagnated and never really reached anything near people expected, you cannot remove that senario from the game, the fact is you will have newgen players with a high PA who develop later in their career, and you will have players with a high PA never reach anything close to what that number says, you will have players who develop early and stagnate and you will have players who are never going to be good enough full stop, the game accounts for all of these senarios.

Your issue really is with the research done on human players, that wont change, and the system your suggesting wont change that, because there has to be a limitation built into the game to prevent an database mostly populated with world class players and this limitation will have to be researched, and will never be 100% for all human players in the game.

Judge the system on the players it generates, because that is the true level of if the CA/PA system works, real life research is just data entered into the game based on one persons opinion at the time he inputs the data. This is updated each year or in some cases twice or three times a year because no one can predict the future.

[ac13]

I would love that SI would implement a possibility to choose random potential for all the players. Eg. player with fixed potential of 170 would have -7 and player with fixed potential of 160 to have -6 etc. (maybe also tune the range where negative potentials can hit). This would create a lot of variation for us who like to play several saves. AS said this could be an option like "attribute masking" and would be very easy to implement programming wise.

This plus CA/PA stars masking and much better attribute masking system would be nice...but seems the game is moving more towards the casual player (CA/PA stars introduced...player role replacing sliders..tactics creator doing the same) so not sure it'll happen.

I suppose they kept sliders as an option so no reason they couldn't have more "advanced" options.

[lucatonix]

I think the CA and PA system in game work fine. It gives a nice distribution of players across all levels. I can see how it can be conceived as inaccurate or unrealistic because one doesn't actually know someones potential ability. However for the game to continue to function properly there must be a fixed PA limit. A reason for this is that no fixed PA limit would lead to an imbalance in player ability. It would be very hard for the game designers to actually get the coding spot on to work without faults for such a complicated system to get a PA-less system to work at this current time.

There are also several flaws with the PA-less system.

1. It is based upon correcting any mistakes carried out by researchers, therefore making late blooming possible, however once newgens are in the game, the need for a PA-less system is nullified. The game can create a PA boundary on these players because it can act as 'God' as it were to say this is the upper bound of the player. What should change in this respect is the scouting system so it isn't as easy to tell if a player has high PA. Although I have had times where my scouts have been wrong numerous times about a winger who had good base attributes. He later developed to become one of the best wingers in the world despite my scouts saying he would struggle to get into a top 2nd tier side at his maximum potential.

2. Late-blooming, and player development as such isn't dependant on how a player has played recently. For example, Jack Wilshere looks promising this season but last year he made about 6 appearances for Bolton in which he was pretty dire to be honest. In the PA-less system his development would be small or virtually none but he is a much improved player from last season.

Scott Parker has played at a high level for the past 3 or so years, and I would argue he has reached his peak. In the PA-less system what would have stopped him from getting a break-through season where he would have a late-bloom, despite him not actually having one last year or the year before. If you can find a way to stop this from happening:

What makes a player undergo a late-bloom. Randomness. This seems rather pointless, since it could mean anyone could get a late bloom despite playing poorly or not at all.

3. How would you stop not only the AI but the human player getting an immensely strong squad. Say I'm managing at Arsenal, it would be pretty easy with all the players with good base attributes, to focus on 3 or 4 to turn world class by giving them lots of games, at a high reputation club, with success and good training facilities. By employing a rotation system of loan, tutoring and game time, it would be then possible to make 6 or 7 world class thus the next generation of players would all stay at one team, making them a super club. This would unbalance the game and make it less enjoyable since it would be easy being the big team or super difficult say being an International manager for 15 years then deciding to go into club management. Your starting midtable club would be very far from the likes of Man Utd and Arsenal in player ability due to the PA-less system and good squad management.

4. How would you take into account PA increase in different countries. Say for example Brazil. There are many promising youngsters there at the moment such as Neymar, Paulo Henrique etc... Would their PA max out at say 165 because they stayed in Brazil or would it be possible to get to 190ish say if they became amazing but noone picked them up.

5, How much effect will attitude have on player and PA development? For example at my school, my football skills have improved much more than others but I'm still not in the first team. I am one of the best in the second XI though. Surely those in the first XI with more regular training, more matches and slightly worse attitude should have improved more than me according to your PA-less system.

[CrazyEastboy1]

In the system that I have mentioned it will be easier to do late boomers as well.

Most of you guys are talking about just match ratings, but what I am saying is that the PA (or you call call it by a different name) should be referenced or tied down to CA. So, when the CA of the player increases the PA or the glass ceiling for their development the PA should also be increased.

The OP is just an example, if u want late boomers then just change the percentage change in corresponding ages. Which will be unknown by human players.

That means a Drogba could have a percentage change of PA as 30%, 50%, 10%, and so on with his 100% change occuring around the 26/27 age mark.

For example:

[ac13]

In the system that I have mentioned it will be easier to do late boomers as well.

Most of you guys are talking about just match ratings, but what I am saying is that the PA (or you call call it by a different name) should be referenced or tied down to CA. So, when the CA of the player increases the PA or the glass ceiling for their development the PA should also be increased.

The OP is just an example, if u want late boomers then just change the percentage change in corresponding ages. Which will be unknown by human players.

That means a Drogba could have a percentage change of PA as 30%, 50%, 10%, and so on with his 100% change occuring around the 26/27 age mark.

For example:

This in effect is the same as having no PA. You may as well just have a potential rate of increase per year which varies per player....which is what LMA had. PES also has something to this effect.

Not a good idea.

[CrazyEastboy1]

This in effect is the same as having no PA. You may as well just have a potential rate of increase per year which varies per player....which is what LMA had. PES also has something to this effect.

Not a good idea.

Rather than just saying not a good idea can u elaborate why is it so?

[ac13]

Because if performance is all that matters then you just going to get all the best players (attribute wise or stats wise) who are young and your team will progress much faster than someone who has slightly better players who are older.

It will lead to the game becoming even easier. Anyway who is going to decide which players are to peak early or develop late...the researchers or will it be random.

The way it is currently it's still possible for players to do badly in game despite their pa's (if they not played regularly). I would suggest simply that CA's of players drop significantly as well when players have a run of bad form (or a bad season). This would mean that players reaching their PA becomes more random.

I think there are easier options...eg making the starting PA more variable...having an option so players are grouped into 5 groups eg 0-100, 60-120, 100-150, 120-170, 140-200. THen you wouldn't be able to just buy the same players as each game diff players would become really good. This along with making it more difficult for players to reach their potential and more difficult to see what the potential is (removing the PA star showing) would make playing more games less repetitive.

[CrazyEastboy1]

I havent said that performance should increase the PA, the increase in CA should be reflected in the PA.

I agree that you could have groups of PA an so on, my example is just a little effort to change things. It is not concrete.

As for deciding who peaks when the AI can do that for regens and the researcher for real players.

[milnerpoint]

it still all boils down to the fact that PA is always there, its your ability to fufil that potential that counts. This really is a non issue outwith human researched players, as the system works perfectly fine with newgen players, but the human element will always be there to skew things beacause we cannot get it exactly right, even with these variable PA's, or unlimited PA's there will still be a human error element and none of these systems will change that, the human restriction that would have to be in place in order to prevent an unbalanced database will always be questioned.

[Biscotti]

This plus CA/PA stars masking and much better attribute masking system would be nice...but seems the game is moving more towards the casual player (CA/PA stars introduced...player role replacing sliders..tactics creator doing the same) so not sure it'll happen.

I suppose they kept sliders as an option so no reason they couldn't have more "advanced" options.

I know it's off-topic, but I just wanted to reply to this bit.

Making a game more accessible does not equal catering to a casual player. 90% of "casual players" would still be ridiculously lost with the complexity of FM. I don't know if you've played the series for long or not, but I remember a time when the game did feel a lot like an interactive spreadsheet and it would take me weeks after a new release just to feel comfortable with where I could find all the information I needed.

The casual gamers are the ones you see posting a few times each week with "make this game more like FIFA" or "It's so unrealistic, I can't use one tactic all the time and dominate the league" type threads. If it were catered to those people, they wouldn't be the ones complaining.

[RBKalle]

There's a bit of a faulty logic in action in the whole "scrap PA altogether because we can't have a Pedro/Drogba/Toni" situation...

The starting database reflects the players' potential as perceived AT THAT POINT IN TIME... No researcher could or would have submitted a PA of 180 for Drogba or Toni for the 2002-03 edition of CM/FM... even if he had a crystal ball, he would have been asked to tone it down because there was NO WAY to think either man could have turned out so good.

If Barça's staff were struggling to "see" Pedro three years ago, which researcher in his right mind would have assigned him a PA of 180?

Let's be honest here... it's unfair using a few unexpected "late bloomers" to dismiss the whole concept of "Joe Hart has better potential than Gunnar Nielsen".

If we can provide a handful of players who have been given a lower PA, why not going for a list of the "early floppers"?! Those who get ludicrously high PA based on a good display as teenagers (or on hearsay) but then failed spectacularly?

The likes of Hadzimehmedovic and Cherno Samba are much better poster boys for the "we need PA" than Drogba or Pedro can be for the "let's get rid of PA" campaign

Wouldn't it be silly saying "we need LOW PA for most players because Knutsen and Sigþórsson had high PA and could turn into international stars while in real life they never went close to that level"??

PA is reasonably accurate for the specific time the game is released, but once you click on "Continue" in your first day of game-time, Real Life goes one way and Football Manager 20xx goes another way.

Or should I moan to Codemasters because in F12010 the outcome of many GPs is different than in real life?

As simple as it is: a player with 170PA can become a world beater 99 times out of 100, a player with 130PA can still overperform if used well. So for the good of the game and for realism's sake PA should stay.

(better if flexible, but we can't ignore our own natural "glass ceiling")

[Pangaea]

I think this whole debates is a bit odd. We are talking about a computer programme, a gigantic Excel sheet with numbers. To achieve any kind of balance over time the players' (potential) ability has to be limited. A fixed number is the easiest way to do that. Any other system will practically ensure an unbalanced system. It would be a nightmare to test and make sure the starting database fits well with the 2050 database, et cetera. The various suggestions I've seen in this thread would ultimately mean "Player in good team will turn good, player elsewhere will not". By putting a lot of importance on facilities we could end up with the big teams turning any kind of youth into a good player, while players elsewhere would be crap. Depending on how this is implemented we could end up with a player pool with a huge quantity of Messis or Cahills or whatever (good to great players), or way too many poor players (poor facilities vastly outnumber good ones after all).

I maintain that the current system is pretty good. Perhaps it is too easy to spot and develop good talent, perhaps some attributes should develop differently over time, perhaps some attributes can be tied to current/potential ability differently. But the core system works well. The Luca Tonis, Giggs' and Sheringhams are rare, very rare. So to use them to beat the current system over the head may be somewhat unfair, although I agree it's difficult in to maintain old players without physical decline. I think what is needed is slight modifications to the current system, not throwing it all out and starting afresh with something new.

All players have a certain talent, and very few players can turn into professional footballers, let alone be Man Utd, Barca or Real Madrid quality. You simply cannot turn any player into a good Premier League or world class player. It can't be done. Big clubs have worldwide scouting networks for a reason. I think the current system reflects this quite well. I certainly don't want to go through 3-4 versions of FM with a whole new "PA" system before it gets reasonably stable and well-tested. And I can guarantee you that a change like this will take years to even approach some balance and stability.

[ac13]

Anyway you do get Pedro situations. In my save Ramsey hardly played at all for Arsenal until he was 23/4. In that season he played regularly and suddenly all his attributes went up. Pedro's break out year was when he was 22 btw.

[ac13]

I know it's off-topic, but I just wanted to reply to this bit.

Making a game more accessible does not equal catering to a casual player. 90% of "casual players" would still be ridiculously lost with the complexity of FM. I don't know if you've played the series for long or not, but I remember a time when the game did feel a lot like an interactive spreadsheet and it would take me weeks after a new release just to feel comfortable with where I could find all the information I needed.

The casual gamers are the ones you see posting a few times each week with "make this game more like FIFA" or "It's so unrealistic, I can't use one tactic all the time and dominate the league" type threads. If it were catered to those people, they wouldn't be the ones complaining.

Well let's see what you can get the assistant to do...team talks...press conferences...select your team for you. You get told at the beginning of the season your best formation. You get told who to sign, what ppm's to suggest to your players..who to use for set pieces.

There's really not much you need to do on your own other than make subs and negotiate transfers/contracts.

[milnerpoint]

Well let's see what you can get the assistant to do...team talks...press conferences...select your team for you. You get told at the beginning of the season your best formation. You get told who to sign, what ppm's to suggest to your players..who to use for set pieces.

There's really not much you need to do on your own other than make subs and negotiate transfers/contracts.

you can blindly follow the assistant manager, but it does not mean you will be successful, its just advice, and for the most part, bad advice. People who only follow these instructions will badly struggle at the game.

[ac13]

you can blindly follow the assistant manager, but it does not mean you will be successful, its just advice, and for the most part, bad advice. People who only follow these instructions will badly struggle at the game.

Got any proof on this?

Calling it making it accessible or dumbing it down is just semantics. There's far more help or hand holding than ever before and it's far more difficult to get anything wrong these days (signings for example). We're now told the players CA and PA and even have a nice little octagon so we don't even have to look at proper attributes. Plus we don't have to think for ourselves which attributes are important for each player cos it's highlighted for us.

Now we don't even have to set up our own training as it's done for us.

[Biscotti]

@ac13 - As milner said, that's all advice, and generally sub-optimal. It's also entirely optional.

Edit:- all you need to do is look around the tactics and training forum to see that the training is not "done for us". What is there is merely a basic template so that if people don't like to tweak too much in the background they aren't punished by the game. Not everyone enjoys messing around with every tiny aspect of the game.

@Pangaea - Debating these kind of issues is indeed a very odd thing to do, but I also think it's extremely healthy. Whilst I don't think any one contributor to this thread agrees 100% with a different contributor, there have been a large number of interesting ideas and concepts thrown around that cause people to actually think hard about the way it all works.

Whilst constant threads on this same topic would be annoying and pointless, a good long debate on the subject every now and then ensures that the player development module will never grow stale, because there are so many possible ways you can approach it.

[milnerpoint]

Got any proof on this?

Calling it making it accessible or dumbing it down is just semantics. There's far more help or hand holding than ever before and it's far more difficult to get anything wrong these days (signings for example). We're now told the players CA and PA and even have a nice little octagon so we don't even have to look at proper attributes. Plus we don't have to think for ourselves which attributes are important for each player cos it's highlighted for us.

Now we don't even have to set up our own training as it's done for us.

and yet despite all this help time and time and time again there are threads from people who do not understand the mechanics of the game, SI give help in game but its basic and does little in the long run to make you successful at the game.

Try it for yourself, start a new game and ONLY follow what your assman and coaches tell you, sign exactly who they say, play the exact tactics they say, let assman do all team talks and press conferences and see how well you do.

[ac13]

and yet despite all this help time and time and time again there are threads from people who do not understand the mechanics of the game, SI give help in game but its basic and does little in the long run to make you successful at the game.

Try it for yourself, start a new game and ONLY follow what your assman and coaches tell you, sign exactly who they say, play the exact tactics they say, let assman do all team talks and press conferences and see how well you do.

I'm sure i'll do fine. You willing to give it a try as well? After all as soon as you scout a u20 tourney and three 4 star plus players are available for 100k-2mil you can't really go wrong.

Anyway it's not like there's any certainty that every person who plays Fifa manager or cm is going to be successful either.

[Biscotti]

I'm sure i'll do fine. You willing to give it a try as well? After all as soon as you scout a u20 tourney and three 4 star plus players are available for 100k-2mil you can't really go wrong.

I'm not sure that doing "fine" is really the point. The point is that you won't be achieving the best possible results that you could (hence my statement of "sub-optimal").

I would like to see a proper test performed on this though, but I don't think I've got the time (or PC specs) to really give it a good go.

[milnerpoint]

I'm sure i'll do fine. You willing to give it a try as well? After all as soon as you scout a u20 tourney and three 4 star plus players are available for 100k-2mil you can't really go wrong.

Anyway it's not like there's any certainty that every person who plays Fifa manager or cm is going to be successful either.

As biscotti says, you will do ok, but if you want to really succeed at this game you need to learn the mechanics, but id be willing to give it a go, only follow what the game says, no making up your own mind except in the situations when you have no choice, dont attend press conferences, dont set up training, dont touch match pre, or team talks, let the ass man select the team and only use tactics recommended pre season. There is a good chance this thread will die off by the time any real results are through so feel free to PM me if you want.

[x42bn6]

I'm going to detail another example to get my point across.

Say that one day, someone in SI Towers decides that it would be a good idea to pre-determine the results of a match before it is played in-game. For example, Manchester United play Arsenal, and the game decides that the "maximum score" is 0-2 - i.e. Manchester United cannot score a goal, and Arsenal can score no more than two goals.

This is a bit like determining PA - the game determines the maximum score beforehand, like "God".

Here are the mappings between CA/PA and the game:

Current state: CA right now/current score at any one time

Maximum future state: PA/maximum score

Expected future state: Some value < PA/some score < maximum score

Late-bloomer: CA-PA gap high, CA suddenly goes up dramatically/0-0 after 85 minutes, Arsenal score twice

Early-bloomer: CA-PA gap becomes low at an early age/Arsenal score twice within the first 10 minutes

In other words, no matter what happens between the first minute and the 90th minute, you cannot get a scoreline better than 0-2. No matter if you do something entirely unexpected that will have a clear difference on the scoreline, like put Rooney in goal, or Almunia in goal.

There is a key reason why this is not acceptable - both managers can do things to influence the score. 0-2 is an expectation; a generation of some random variable, but does not say why it thinks 0-2 is a sensible maximum scoreline. During the game, 0-2 could look quite silly as an expectated maximum - say Arsenal go down to 9 men by half-time, and Almunia is in goal, although the score is still goalless.

I challenge you to think this is acceptable, and map this back across to PA.

----------------------------------------

The low PA newgen can't late bloom because he's not capable of that level of ability. A high PA newgen is still entirely capable of late-blooming. If a real-life player has low PA and late-blooms in real life, then the researcher was wrong, not the game.

Step back one second. What made Luca Toni a "late-bloomer"? He found a consistent level of form that was higher than anyone's expectations and for a whole season, everyone waited for his form to drop - it simply didn't. Form is temporary, class is permanent (in hindsight, one year's form is probably temporary too... But I digress).

Toni didn't "late-bloom" because his "PA let him" - he was a late-bloomer because he performed well consistently brilliantly for one whole season.

PA is an expectation, and since Toni exceeded that expectation, PA was wrong.

Which is why I think you're missing the point here:

then the researcher was wrong, not the game.

Any player - regen or not - that consistently performs beyond expectations at a relatively-old age should be a late-bloomer. If two players are identical in every way except one has a low PA that prevents "late-blooming", and both perform equally well at a relatively old-age on a consistently-brilliant basis, then I would expect these two players to gain identical rises in ability and identical rises in reputation - they should be equally-good late-bloomers.

However, the one with the lower PA doesn't get the benefits because his PA is too low. Isn't this contradictory? The expected PA generated at the start of the entry of the player into the game, which was around a decade ago, hinders the player, despite the fact they perform equally-well, beyond expectations.

This is like saying, "Nope, you can't get any better, because I didn't expect you to be that good in the first place."

You are still basing your assessment on a player that was judged by a human when it later turned out that human was wrong. With newgen players, that is a physical impossibility, because there is no human involvement in the process. Therefore, the issues you seem to have with the system are only relating to real-life players and the research surrounding them.

The fictional players I've mentioned above may be real or regen - it doesn't matter. It just happens that it is so much easier to point to real-life examples rather than "look at my Andriy Biskup regen in the game. He once played for Italian powerhouses Catania."

That's basically a slot machine and only people that want the 'thrill' of gambling but don't know how to play real casino games, either by ignorance or lack of desire, play them. Hardly an example to use to ignore balance in a game as deep as FM. After all if everything in FM came down to luck I wouldn't touch the thing, nor would many other I bet. You need depth and balance to that depth. Just figuratively flipping a bunch of coins when it comes to developing youth in-game is a joke and not credible game design.

Not true at all. A no-PA system is a process without limits that has a predictable outcome. Said arcade machine is a process without limits that has a predictable outcome.

Another example. When you step out on the road to cross the street, there are an infinite number of ways you could die. One of those is that you get abducted by aliens and are turned into alien food. You can't guarantee this won't happen - but you never worry about this. Why? On average, you cross the street perfectly fine.

Well if it's so easy to do perfectly without making mistakes then you do it.

Ah, goody, I was wondering when the "so you do it" argument would come in... Software development is a lot more than just throwing an extra body in there.

I can't, quite frankly, because I don't work for SI, and probably won't for a while, if ever.

To be that critical of how real players are represented is just stupid because everyone has their own opinion of who's faster, who's more skilled, who's a better student of the game, etc. You can't fault the game over something that's basically 100% subjective. I'd be just as happy without real players in my game, you seem to not be able to separate yourself from them.

There is no need to think I'm deliberately targetting real players here - I only talk about real players because they are examples everyone can relate to. Nobody knows anything about Dominik Helm, my regen reserve goalkeeper, or my dilemma at right-back where I have two regens fighting it out for the first-team spot. On the other hand, I can say, "Luca Toni" and everyone knows what I am talking about.

You mean like a player with loads of potential but at a weaker club with poor facilities/coaches, then moving up to a larger club and coming into that potential when he has the tools at hand? Perfect example right there. Or a player with loads of potential and at a top club but has his youth career hampered by injury.

All possible, but the PA of the player never moves, despite the fact the expectation has changed. Chris Smalling may never play for a semi-professional team ever again, for example.

You are grasping at straws and your entire thought process is flawed.

No, I just disagree with you.

You completely discount the need for limitations and balance,

You don't need limitations if the system is balanced. Like the arcade example.

you are basing everything on the fact that real players are different then in-game,

Under Model-View-Controller, there is no reason to treat them differently.

I've mentioned why I talk about real players above.

and you are ignoring the fact that the system does indeed work you just want it to work differently.

So your argument is "it works because it works"?

No system is perfect. No system works. Some systems work better. I have detailed an explanation of why it is better, but you are assuming I'm targetting real players for some reason. I'm just using them as examples.

but that is a situation that happens in real life, Reo-Coker is a good example, was very good when he was younger and stagnated and never really reached anything near people expected,

I said, "rises quickly anyway, he stagnates and then late-blooms". Reo-Coker hasn't late-bloomed yet (if at all).

you cannot remove that senario from the game, the fact is you will have newgen players with a high PA who develop later in their career, and you will have players with a high PA never reach anything close to what that number says, you will have players who develop early and stagnate and you will have players who are never going to be good enough full stop, the game accounts for all of these senarios.

I have never denied it is possible.

I have said that it happens under a set of conditions that is not realistic.

A player can only late-bloom if his PA is high enough. To me, this is contrary to real-life.

- In-game: Players can late-bloom only if their PA is high enough and they don't reach it quickly-enough to begin with.

- In reality: Players can late-bloom if they consistently outperform expectations at a relatively-old age.

These are not equivalent. PA is an opinion of an expected maximum, but in reality, this opinion is wrong when a player starts to late-bloom! If we said "I don't expect Toni to be better than a mid-table striker", and Toni proved us wrong, then our opinion - i.e. PA - is wrong.

The reason is simple - PA partially-determines how a player can late-bloom, but fails to take into account why it happens.

Your issue really is with the research done on human players, that wont change, and the system your suggesting wont change that, because there has to be a limitation built into the game to prevent an database mostly populated with world class players and this limitation will have to be researched, and will never be 100% for all human players in the game.

No, it is not. I just use real players because nobody knows tons about my regen-full game (heck, even I don't). Real players are easy to refer to.

Judge the system on the players it generates, because that is the true level of if the CA/PA system works, real life research is just data entered into the game based on one persons opinion at the time he inputs the data. This is updated each year or in some cases twice or three times a year because no one can predict the future.

I have no issue with how the players are generated - just what is generated.

There's a bit of a faulty logic in action in the whole "scrap PA altogether because we can't have a Pedro/Drogba/Toni" situation...

The starting database reflects the players' potential as perceived AT THAT POINT IN TIME... No researcher could or would have submitted a PA of 180 for Drogba or Toni for the 2002-03 edition of CM/FM... even if he had a crystal ball, he would have been asked to tone it down because there was NO WAY to think either man could have turned out so good.

If Barça's staff were struggling to "see" Pedro three years ago, which researcher in his right mind would have assigned him a PA of 180?

I'm not denying it's an issue, but we both agree here that today's opinion cannot be used as gospel for the rest of his career. Room must be made for human-error within the game, which doesn't exist right now.

As simple as it is: a player with 170PA can become a world beater 99 times out of 100, a player with 130PA can still overperform if used well. So for the good of the game and for realism's sake PA should stay.

(better if flexible, but we can't ignore our own natural "glass ceiling")

Flexible PA has the issue that sometimes a player might smack his head on the ceiling, wait for his PA to rise, then smack into it again, then wait for it to rise, then smack into it again... The player "loses" a year's worth of development as a result of the fact that the "regeneration" happens once every season.

Imagine, then, that your flexible PA value is generated on a monthly-basis. Every month, the player is re-evaluated, instead of every season. The individual jumps may be less, of course, but it is another idea. This way, the player gets a smoother development curve, and it is more realistic because he doesn't "suddenly" get to develop during the start of a new season, if he does smack his head on the ceiling.

Then go even further - imagine it is generated on a weekly-basis. And then imagine it is generated daily. This PA value is recomputed every day instead. "Boss, he won't be better than League Two relegation fodder," your assistant says every day for 15 days, before it gradually improves... It becomes like the OP's tables, where the CA-PA difference eventually becomes zero at the peak, but analysed on a daily basis.

Then it begs the question - why do you need to compute the PA value? Every day, the game generates a new PA value that determines how they will peak. Every day the same question is asked, and you get a slightly different answer every day. 100. 110. 113. 115. 112. What does this number achieve?

If it is to stop players from developing too much, essentially the daily PA value is like a ceiling. You can't gain more than 0.1 CA per day, for example - that is essentially what it is. It limits players' development rates. Which is fine, if you need absolute certainty, but I argue you don't need to, in the same way you don't need absolute certainty you will not be turned into alien food by crossing the road to avoid crossing it.

There may be a theoretical maximum within the game and in real-life - but we don't need to worry about it. For example, the maximum value for development would be a player that has a perfect personality, suffers no injuries ever, never burns out, has the best coaches and training facilities in the world, adapts perfectly-well, gets first-team football for 90 minutes for every single game and gets a rating of 10.00 every game. So unlikely it is simply not worth considering, like alien-food-world.

I think this whole debates is a bit odd. We are talking about a computer programme, a gigantic Excel sheet with numbers. To achieve any kind of balance over time the players' (potential) ability has to be limited. A fixed number is the easiest way to do that.

No, it doesn't - I've listed unbounded processes that have predictable results (search for "arcade"), and the alien food example.

It is easier to limit things, but no, it is not necessary by any means.

Any other system will practically ensure an unbalanced system. It would be a nightmare to test and make sure the starting database fits well with the 2050 database, et cetera.

Not really... A well-designed system will ensure unlikely situations are unlikely, and as long as the spread of results is reasonable, it is fine. Unusual things can happen in games anyway - as long as those things remain unlikely, nobody complains. You might get a 15-0 Premier League result one day - not a concern unless it is commonplace.

Under the Central Limit Theorem, as long as there are enough soaks, a random process is just fine.

The various suggestions I've seen in this thread would ultimately mean "Player in good team will turn good, player elsewhere will not".

No, I've said that facilities are just one ingredient in development. Things like talent, injuries (or lack of them), adaptability, attitude, amount of first-team football, how good they perform (in the first-team and at youth-level) all play a part. Having world-class training facilities helps, of course, in the sense that Manchester United's gym and staff will help you develop better than dumbbells in your bedroom, but it is by no means the only factor.

I maintain that the current system is pretty good. Perhaps it is too easy to spot and develop good talent, perhaps some attributes should develop differently over time, perhaps some attributes can be tied to current/potential ability differently. But the core system works well. The Luca Tonis, Giggs' and Sheringhams are rare, very rare. So to use them to beat the current system over the head may be somewhat unfair, although I agree it's difficult in to maintain old players without physical decline. I think what is needed is slight modifications to the current system, not throwing it all out and starting afresh with something new.

It's not a radical change - it's just removing training wheels from a bike, but learning how to ride a bike better.

Removing PA also removes CA and this removes attribute weighting and weak-foot weighting - all pluses in my opinion.

All players have a certain talent, and very few players can turn into professional footballers, let alone be Man Utd, Barca or Real Madrid quality. You simply cannot turn any player into a good Premier League or world class player. It can't be done. Big clubs have worldwide scouting networks for a reason. I think the current system reflects this quite well. I certainly don't want to go through 3-4 versions of FM with a whole new "PA" system before it gets reasonably stable and well-tested. And I can guarantee you that a change like this will take years to even approach some balance and stability.

Never say never. You never know - as you said, Toni is "rare".

And I'm not saying SI should definitely run into this quickly without stabilising - but I would like them to think about it anyway.

[milnerpoint]

I'm going to detail another example to get my point across.

Say that one day, someone in SI Towers decides that it would be a good idea to pre-determine the results of a match before it is played in-game. For example, Manchester United play Arsenal, and the game decides that the "maximum score" is 0-2 - i.e. Manchester United cannot score a goal, and Arsenal can score no more than two goals.

This is a bit like determining PA - the game determines the maximum score beforehand, like "God".

Here are the mappings between CA/PA and the game:

Current state: CA right now/current score at any one time

Maximum future state: PA/maximum score

Expected future state: Some value < PA/some score < maximum score

Late-bloomer: CA-PA gap high, CA suddenly goes up dramatically/0-0 after 85 minutes, Arsenal score twice

Early-bloomer: CA-PA gap becomes low at an early age/Arsenal score twice within the first 10 minutes

In other words, no matter what happens between the first minute and the 90th minute, you cannot get a scoreline better than 0-2. No matter if you do something entirely unexpected that will have a clear difference on the scoreline, like put Rooney in goal, or Almunia in goal.

There is a key reason why this is not acceptable - both managers can do things to influence the score. 0-2 is an expectation; a generation of some random variable, but does not say why it thinks 0-2 is a sensible maximum scoreline. During the game, 0-2 could look quite silly as an expectated maximum - say Arsenal go down to 9 men by half-time, and Almunia is in goal, although the score is still goalless.

I challenge you to think this is acceptable, and map this back across to PA.

Thats not even close to being the same situation, you cannot compare the two because your talking about two completely different things.

You have to accept we all have our limitations, none of us have infinite potential especially not when it comes to something as specific as playing football and you cannot expect that anyone with the right attitude and conditions can become a top player, otherwise scouting would be pointless, you would just group a heap of people together, pick those with the best attitude and boom with good training and game time you have a world class team.

The room for human error is made with each new input of data, there is no other way around this, plus it doesnt really matter if the game doesnt represent how players turn out in life, because as soon as you click continue it ceases to become related to real life, it becomes the world of FM.

What does count is how the system works with generated players and it works almost perfectly, it accounts for almost all senario's, late bloomers, early developers, world class players, lower league players, players never good enough to make it, players good enough but had issues with injurys, players who have one good season and drift away, anything you can think of can happen to a newgen player.

Your still ignoring the point that removing PA will not remove the issues you bring up, because for real players there will ALWAYS be a human element that could be wrong, and to make the system balanced a limitation would have to built in, you have still to actually expain what this would be and without it the whole thing falls apart.

[Cougar2010]

I'm going to detail another example to get my point across.

Say that one day, someone in SI Towers decides that it would be a good idea to pre-determine the results of a match before it is played in-game. For example, Manchester United play Arsenal, and the game decides that the "maximum score" is 0-2 - i.e. Manchester United cannot score a goal, and Arsenal can score no more than two goals.

This is a bit like determining PA - the game determines the maximum score beforehand, like "God".

Here are the mappings between CA/PA and the game:

Current state: CA right now/current score at any one time

Maximum future state: PA/maximum score

Expected future state: Some value < PA/some score < maximum score

Late-bloomer: CA-PA gap high, CA suddenly goes up dramatically/0-0 after 85 minutes, Arsenal score twice

Early-bloomer: CA-PA gap becomes low at an early age/Arsenal score twice within the first 10 minutes

In other words, no matter what happens between the first minute and the 90th minute, you cannot get a scoreline better than 0-2. No matter if you do something entirely unexpected that will have a clear difference on the scoreline, like put Rooney in goal, or Almunia in goal.

There is a key reason why this is not acceptable - both managers can do things to influence the score. 0-2 is an expectation; a generation of some random variable, but does not say why it thinks 0-2 is a sensible maximum scoreline. During the game, 0-2 could look quite silly as an expectated maximum - say Arsenal go down to 9 men by half-time, and Almunia is in goal, although the score is still goalless.

I challenge you to think this is acceptable, and map this back across to PA.

----------------------------------------

Jeez thats an essay x42.

The quote above is really not a great example of your point of view. God does not limit the scoreline of football matches in real life.

I do understand your point of view btw I simply disagree with it.

Your saying that everybody has the capacity to improve if they are playing well and enjoying their football. I sort of agree with that point but there is always a limit that is reached which is the PA. The difference between real life and FM is that players in FM reach their PA too easily and too often whereas in real life very few players reach their PA hence the capacity is there with most players to improve to some degree.

The issue is still CA and the way it develops and not PA.

[Biscotti]

This is a bit like determining PA - the game determines the maximum score beforehand, like "God".

The difference is, a person's potential is determined by "God" in the real world. A match result is determined by the players and managers involved. Hell, if you want to go the "God" route, you could argue that match results are determined by that too.

The reason your conclusions actually work is because you are working on the basis that we all already know the PA of every player (which to a certain extent, we do, and that is one of the problems already highlighted with the system as it is), but if you didn't know the PA of any player in the game it is entirely plausible for them to be written off as "average" for the first 10 years of their career and then to suddenly become amazing.

[whilewolf2]

Perhaps I’m missing something but under x42bn6 proposed no PA system how will player ever peak, plateau then tail off with out suffering an injury, enforced lay off, age or physiological crisis.

If a player gains CA by playing well then wont players who play consistently well continue to develop until a change in circumstances stops their development? .

What about players like Gary Neville or Tim Cahill who achieved an impressive level of performance stayed at the same team with the same facilities, played regular first team football and didn’t suffer an injury or dip in form but didn’t get any better instead they plateaued playing at the same level for several seasons. Indeed isn’t this the most common career trajectory for players?

Perhaps their performances were not good enough to warrant a CA increase under x42bn6 proposed system but if that’s the case a very high level of performance must be necessary to gain CA. Now young players aren’t going to perform well enough to gain CA unless you give them a boost because of their young age. But one of your primary objections to the current system was that it didn’t adequately account for late bloomers. If a significantly better then average period of performances is necessary to gain CA what will make mature players suddenly perform much better then their average late in their careers?

[Scoham]

PA changing throughout a players career? Ridiculous. Look at what it stands for Potential Ability.

The maximum a player can reach if all goes well. Fine as it is, everyone has a limit no matter how hard they work.

As said players like Pedro were simply underrated. It happens. However there are plenty of players that are rated very accurately that go unnoticed.

No doubt the way players develop (or not) their CA can be improved, but PA should not be changing every season. Could argue PA should be harder to reach, meaning researchers could be more generous with PA ratings. Agree also in some cases it should be harder to spot potential.

[CrazyEastboy1]

Well we can simply not call it PA then

[CrazyEastboy1]

The point that a lot of u are making is that every player would be good enough at good enough clubs, well how can they. With the system that I have talked about to become good your CA must increase, and at that same time the PA increase for that period should be maximum. So, it is less probable that everone would be a star NO