Year 2012: The Rise of the Machines - The Super Newgens in FM 2012

[whilewolf]

I’m kind of just recapping what needs to happen within the game world to keep it balanced and working correctly while the Regens kick in.

1: The number of players in the Db and spread of CA/PA needs to remain roughly consistent.

We know it isn’t this year due to a bug that means inactive nations produce too many high PA players

2: The player’s attributes must accurately recreate the types of players we see in the real world. There must be big strong Target men and fast mobile Poachers.

It’s getting better but there are still a few problems. Physical attributes are too likely to be high in high CA players and low in low CA players.

A lot of attributes have no weighting these Tendencies occur too randomly things like bravery need to be tied to player’s roles so it’s highly likely to be high in DMs and more random in AMs.

Hidden attributes values are allocated far too randomly. SI needs to create templates for Mental attributes so Regens have the kinds of temperaments that researched players do.

3: AI managers must develop the young talent in their squads so they naturally take over as the older researched players decline.

I think this is a real problem the AI teams don’t give youth a chance and play high reputation players as long as they can with the result that a lot of players who are twentyish when the game starts never get much football until their late twenties.

This is a harder problem to solve then the other two because it requires the AI to show some forethought and plan for players declining and blood youth accordingly.

The AI also needs to be able to revaluate the importance of competitions on a round by round basis so if its mid-table with two games to play its more willing to give its prospects a run out.

[BiggusD]

I’m kind of just recapping what needs to happen within the game world to keep it balanced and working correctly while the Regens kick in.

1: The number of players in the Db and spread of CA/PA needs to remain roughly consistent.

We know it isn’t this year due to a bug that means inactive nations produce too many high PA players

2: The player’s attributes must accurately recreate the types of players we see in the real world. There must be big strong Target men and fast mobile Poachers.

It’s getting better but there are still a few problems. Physical attributes are too likely to be high in high CA players and low in low CA players.

A lot of attributes have no weighting these Tendencies occur too randomly things like bravery need to be tied to player’s roles so it’s highly likely to be high in DMs and more random in AMs.

Hidden attributes values are allocated far too randomly. SI needs to create templates for Mental attributes so Regens have the kinds of temperaments that researched players do.

3: AI managers must develop the young talent in their squads so they naturally take over as the older researched players decline.

I think this is a real problem the AI teams don’t give youth a chance and play high reputation players as long as they can with the result that a lot of players who are twentyish when the game starts never get much football until their late twenties.

This is a harder problem to solve then the other two because it requires the AI to show some forethought and plan for players declining and blood youth accordingly.

The AI also needs to be able to revaluate the importance of competitions on a round by round basis so if its mid-table with two games to play its more willing to give its prospects a run out.

A game like FM must have the entire game built around the AI. It is the entire point of the game. Many developers give up doing this and instead go for online solutions, but SI tried this and failed.

It is clear that SI has not thought of it this way, because many of the features in the game are not used by the AI - and others are used wrongly. It shouldn't be a matter of ease to master any feature in the game better than the AI, but it is. I think that the first thing that the development team needs to do is to sit down and program an AI manager that uses every feature in the game perfectly. That is the only way the AI manager attributes can be meaningful tools in the database.

5 simple design steps

1. How is the feature supposed to work

2. What is the correct, maximized use of that feature

3. What does the game need of information to allow the AI (and the user) to use the feature correctly

4. Make sure that the AI, when the attribute governing the use of this feature is maxed out, uses the feature correctly

5. Create a range of output that lets the use of the feature be "scaled down" according to the attribute range. (As such, step 4 above shouldn't be available in that attribute range).

This way, the use of for instance training schedules would let the AI have as a governing body the divinely correct schedule for any player. The human manager who has a lot of patience, skill and experience, could of course figure out that schedule too, but he could never make a -better- schedule since the entire training feature was designed around that schedule. As it is now, the guys at SI does not know what the best schedule for a player is and consequently the AI don't either. The input is cloudy and the output is wizardry and magick for them as well.

Needless to say, programming an AI is a frustrating task when it tries to imitate life (which is impossible of course), but a computer game is not an infinite set of variables and nothing is down to random chance; so the solution is to let life imitate art, and then create the illusion that life is really imitiating life.

[Thomasmc135]

My issue is with all the terrible regens. I mean, why would there be regens with a PA below 5, let alone hundreds? It just takes up unnecessary space in the game world. I don't think regens should be allowed to have a PA of less than 35.

[Seismica]

There is a very simple solution; reduce the number of leagues loaded and run a smaller database.

Five seasons into my Aston Villa game (Which only have the English leagues loaded) and there is only around 7 regens with a PA above 185, and only one above 190 (According to Genie Scout). The one above 190 has severe personality flaws which makes him not a viable option.

The quality of regens here compared to the quality of regens on another save I have which has Germany, France, Brazil, Argentina etc. loaded on a large database, is significantly lower.

The number of high PA regens seems to scale to your database size. This does make sense; but the ease of being able to find the best regens (Even without external tools that reveal hidden attributes) means that if the player has a large database with many high profile leagues/countries running; it's no surprise that they are going to have an abundance of regens available to them.

The problem lies with the limitations of the game; the skill/ability of a player is a quantitative value. When a newgen spawns their CA/PA are usually proportional, i.e. within a certain range of eachother. You're not likely to find a player with 195 PA in a player with 30-40 CA. They're more likely to have a CA of 90-110, but you can't say the same about a player with 140 PA; their starting CA is likely to be lower. Even without external tools, you can identify the best players by looking at their starting stats, and then cross referencing that with the PA recommendation from your scout (Taking into account the facilities of the club they are at, which has an effect).

You cannot tell me that the game is producing too many players with the potential to be world class; because that simply isn't the case. You are just able to identify them and train them too easily. But beware, 'fixing' this could reverse the problem. You might end up not getting enough quality later in the game. I remember in Total Club Manager 2004/2005, this was a significant problem as the players who retired simply weren't replaced, and the most talented youngsters couldn't be trained up to the same level. As for identifying them, the game already provides the option to hide player stats. You could further reduce the problem by hiring reasonably good scouts, instead of ones with PA judge of 19-20.

[kepap]

My annoyance is that there aren't any Inzaghi-type of regens or similar. Meaning ones that have very specific skills and are suited only for one role. Most great regens follow the same blueprint: they excel in two whole attribute CATEGORIES(technical, mental, physical) and are "only" very good in one. They are too versatile. They should be more role-bound. Of course there are great allrounders like Yaya Toure IRL but they are quite rare.

[Frank Abbondando]

The newgens on my current network game (I very rarely play on my own) are absolutely awful. We're 4 seasons in and afaik the host isn't using any external tools. The best one I've seen is a RB from France but even his stats are only mediocre.

If I do play on my own, I create a bunch of players in the editor but leave every attribute to chance and restrict myself from signing any of them for the first 10 years. Makes the game a bit more unpredictable for me.

[Mittens2317]

TBH, I actually reckon this is quite realistic. Look at the standard of footballers from 40 years ago, and compare it to nowadays.

I have this argument with older supporters all the time. The likes of Colin Bell, George Best, etc. would struggle to play in League 1 if they stepped through a time machine. Footballers nowadays are like genetically engineered super athletes, and this trend will continue, the gulf will widen on an almost constant basis until we reach the peak of human conditioning. Whilst it would probably take twice as long for these youngsters to get to that level, these superstar newgens are far from unrealistic, and represent a very possible footballing future.

If anything, I'd say that the "ratings out of 20" are unrealistic, and they should up it 100. On top of that, instead of giving a concrete figure, you should have to consult your backroom staff, who give their verdict. Instead of a finishing attribute of 15 (out of 20), it should be "somewhere between 76 and 89 out of 100", making the stats a lot more vague, bringing your managerial and observational skills into play so that it'd actually you who determines whether or not a players good enough, as opposed to an easily read page of 100% accurate attributes.

[Guest]

TBH, I actually reckon this is quite realistic. Look at the standard of footballers from 40 years ago, and compare it to nowadays.

I have this argument with older supporters all the time. The likes of Colin Bell, George Best, etc. would struggle to play in League 1 if they stepped through a time machine. Footballers nowadays are like genetically engineered super athletes, and this trend will continue, the gulf will widen on an almost constant basis until we reach the peak of human conditioning. Whilst it would probably take twice as long for these youngsters to get to that level, these superstar newgens are far from unrealistic, and represent a very possible footballing future.

If anything, I'd say that the "ratings out of 20" are unrealistic, and they should up it 100. On top of that, instead of giving a concrete figure, you should have to consult your backroom staff, who give their verdict. Instead of a finishing attribute of 15 (out of 20), it should be "somewhere between 76 and 89 out of 100", making the stats a lot more vague, bringing your managerial and observational skills into play so that it'd actually you who determines whether or not a players good enough, as opposed to an easily read page of 100% accurate attributes.

It's not realistic because as has been mentioned before the greats of today make up a very small percentage just like the greats of 20, 30 or 40 years ago did, because the game needs a defined upper limit to work (PA) then it is hard to argue that a marked increase in the number of very top quality players is a reflection of natural trends because with every generation of footballers there will only be 2 or 3 that represent the very pinnacle, not the 10 or 15 that happens in 2045 FM.

It's also worth noting that the ratings are currently based on a 1-200 scale.

[Mittens2317]

Maybe the best solution would be a "stats downgraded due to advancements in athletic capability" thing every five seasons, in the same vein that training facilities are downgraded because of technological advancements.

[scott MUFC]

not many newgens in my blackpool save

but I havn't seen newgens to the effect I see here ever, first database you got far more randomly high PA players, but very very very few could get anywhere near first class.

I have only ever seen ONE at 200, that was in the old database.

to me it makes no differance, it is where the stats are that makes the player

[RBKalle]

There are several separate issues with newgens IMO.

1) Too many awesome newgens from obscure nations. The "Youth rating" taking into account population is completely skewed and that leads to the likes of Egypt and USA churning out insanely good players year after year.

2) Great newgens are too good from the start. Most of the aforementioned wonderkids are world beaters at age 19 already, with attributes that would make them regulars at every top club in the world. That's not very realistic, or at least not as rare as in real life.

3) Average newgens are terrible, and even those who COULD be "fixed" with training and tutoring fall victim of AI's inability to develop players.

4) Newgen templates always have at least one fatal flaw and are quite unrealistic. It's not uncommon getting Crouch-sized wingers with amazing pace and dribbling, or tiny CBs with woeful physical traits... So in such scenarios even a worldclass PA won't help much because the original CA points are so poorly distributed the remaining ones can't really remodel the player into a good one.

4b) Newgen templates lack variety... Most strikers are either tall, lumbering oafs [whose best skill, heading, is useless with the current ME] or tachnically rough speed merchants [who'll score plenty despite single-digits mental and technical skills becasue "All You Need is Pace"]

Same goes for pretty much every other position...

Bottom line is: newgens are either Champions League material at age 19 or are doomed to a career of mediocrity. There's no middle ground... We get more than enough Top Players, but we're lacking the whole "middle class" of footballers here. That's the reason Top Clubs can still be competitive after some years, while mid-table acts turn into crap.

[yifanwang99]

Yep the huge difference between the current crop of real youngsters and computer regens are staggering.

I'm not having particularly a huge problem with the NUMBER of the high-pa regens provided that only a realistic number of them reached the potential. I.e. I don't mind 10 people with 185+ (Ronaldo/messi like) PAs comes out worldwide every 3 years provided that less than 1 in 10 of them actually reaches their potential where the others eventually faded either because of injury proneness, lack of professionalism/ambition or lack of proper training/recognition by managers/scouts.

In fact, the current 'real' youngsters should've been given a much higher PA. For instance, there's little doubt in my mind that players such as Mario Balotelli (duh, always him) has a PA of 185+ but it's unlikely that he'll reach it unless his attitude changes drastically. Wigan fans might remember Amir Zaki, who I also believe can easily achieve a PA of 175+ only if he tried hard enough.

[scott MUFC]

using PA for newgens on my curent save

current best ones are:

a striker at zaragoza with 193PA - however he only has 68 CA

Attacking midfielder (LC) at Real madrid 191PA however again low CA at 87

Attacking Midfielder (LC) at Hajduk 190PA - 90CA

all are 15

next one

Attacking Midfielder ® at AC milan 189PA - 98CA - this guy is 17

then we got a few at 180-182

still not that many

[Smac]

Quick question mostly for the OP:

Your database itself isn't expanding, is it? Because if you had database creep as a result of using the ADD/Remove leagues feature you'd get more top players simply due to having more players. I doubt that's the case but not bad to be sure...

[alfieham]

I'm only in the 2019 in my save but now I've made my way to the premier league what is annoying me are the insane physical stats.

I appreciate this has most likely been mentioned,I just played Liverpool who had two players with 20/20 pace/acc. I struggle to think of players that quick IRL never mind two in the same squad.

[Romanista.]

using PA for newgens on my curent save

current best ones are:

a striker at zaragoza with 193PA - however he only has 68 CA

Attacking midfielder (LC) at Real madrid 191PA however again low CA at 87

Attacking Midfielder (LC) at Hajduk 190PA - 90CA

all are 15

next one

Attacking Midfielder ® at AC milan 189PA - 98CA - this guy is 17

then we got a few at 180-182

still not that many

how is that not many? you got 3 15 year olds with the potential to become as good as messi. that is pretty good and realistic.

[scott MUFC]

messi is 199.

and those players have far too much work to get to that point. but it is hardly a huge wave of 195+

[Guest]

using PA for newgens on my curent save

current best ones are:

a striker at zaragoza with 193PA - however he only has 68 CA

Attacking midfielder (LC) at Real madrid 191PA however again low CA at 87

Attacking Midfielder (LC) at Hajduk 190PA - 90CA

all are 15

next one

Attacking Midfielder ® at AC milan 189PA - 98CA - this guy is 17

then we got a few at 180-182

still not that many

Looks like you two have missed the last part of his post, I get the impression that he's suggestion the numbers in his game are okay.