Just found something amazing!!! Must see!!!

[Pompeyboyz]

Ok running a test today to see how it work.

I am using Jermaine Defoe as my Guinea Pig and I have asked him to retrain as a DC. I am using Tugs 09 Training as well.

I have taken a screenie before I started the holiday mode and when I get home tonight(weather permitting) I will pull another from the end of the season to see what has happened. I will post the results for all to see.

I have made no changes in any editor so this is perfectly clean data being run with.

[HerrKaputt]

Ok running a test today to see how it work.

I am using Jermaine Defoe as my Guinea Pig and I have asked him to retrain as a DC. I am using Tugs 09 Training as well.

I have taken a screenie before I started the holiday mode and when I get home tonight(weather permitting) I will pull another from the end of the season to see what has happened. I will post the results for all to see.

I have made no changes in any editor so this is perfectly clean data being run with.

Great! Thanks a lot, glad you have a home computer to work with (I play FM 08 on my laptop, which I also use for work).

Remember that you have to run two games: starting in July 2007 (for example), run until June 2008 with a certain training schedule and retraining to DC. Then restart in July 2007, run until 2008 with the very same training schedule but don't retrain to DC. (I'm just making sure we're on the same wavelength here)

When you post the results, the most interesting ones will be those graphs of the evolution of attributes for the last 12 months that you can see under "Training" --> "Attributes". Please show us the graphs for the shooting category (finishing, composure etc) as well as the graphs for defending. You can select those categories in the top right corner if I remember.

Thanks again for taking the trouble to do this

[Pompeyboyz]

Well I was actually going to do 4 separate trials. The one posted above. One with the positional training and the general training. One with no positonal training and on Tugs training and one with no positonal and general training so that we'd get a good cross check of what effects there are and how it improves.

I am running each process from the start of the Enlgish game season through to 31st May.

[HerrKaputt]

Well I was actually going to do 4 separate trials. The one posted above. One with the positional training and the general training. One with no positonal training and on Tugs training and one with no positonal and general training so that we'd get a good cross check of what effects there are and how it improves.

I am running each process from the start of the Enlgish game season through to 31st May.

That's even better. Thanks again

[Pompeyboyz]

I will add that due to wanting to play the game as well as test it i will be posting the results on a day to day basis so tonight I will show the results on Defoe having Tugs training and the position training. Tomorrow I will run a secnd test while at work and I should be able to grab a couple over the weekend

[HerrKaputt]

I will add that due to wanting to play the game as well as test it i will be posting the results on a day to day basis so tonight I will show the results on Defoe having Tugs training and the position training. Tomorrow I will run a secnd test while at work and I should be able to grab a couple over the weekend

Can I make a humble suggestion? I would take the screenshots on a day to day basis, so you can play, but post them all together in one post. If you post the screenies at a rate of one per day they might end up in different pages, which will make comparison difficult. Putting them all in the same post would make comparison much easier.

[Pompeyboyz]

Can I make a humble suggestion? I would take the screenshots on a day to day basis, so you can play, but post them all together in one post. If you post the screenies at a rate of one per day they might end up in different pages, which will make comparison difficult. Putting them all in the same post would make comparison much easier.

Your suggestion has been taken and will be my goal.

[Pompeyboyz]

I am intrigued by this one as I have just checked back the results for the first test and Defoe has not taken the DC positoin and his stats are also identical except for the FRee kick taking dropping one place.

[Ljuba82]

I'd like to ask the original poster, Ljuba82, to do the following:

- Edit the original post saying that this is NOT a cheat and that you do NOT need to use the editor

- Also edit it with a step-by-step "recipe" to see the effects of this in FM. Have two such "recipes", one using the editor (which can be done quickly, if someone wants to see this) and one using just the game, for example something like

"Take the original FM database, size small, take Juventus, and immediately start training Ibrahimovic as DC, keeping the general training schedule. Go on holiday until end of season and check the evolution of his finishing. Then restart the game with everything equal but do not retrain Ibra to DC, just keep him on the general schedule, and go on holiday until end of season, and check his finishing."

This would make people stop coming here saying "big deal, you found a cheat", and would draw attention to the fact that it is not a cheat and can be done in the game itself without the editor or third-party programs.

If you don't do it, I don't blame you -- it would be a bit of work. If you wait a while, I will make such a post here and you can just copy/paste it to your first post.

It's not about work, I have spent lot of time trying to explain the best I can, so if you have some suggestion or if you are willing to write something, you can send me to PM, a nd I'll copy that to my 1st post

[Ljuba82]

Having just spent the best part of 2 hours at work reading the whole of the thread I understnd the situation better and retract my initial belief that this would cure the 2 ST tactic bug as a whole but in simple terms it should make my ST better players. The tactic bug will still mean they don't perform as well as they should.

PS Ljuba82......I hate you for making my head hurt with all the information on this thread.....I need to lie down in a darkened room for a few days :D:D

...heheh

...mate I still don't see connection between this and the fact that 2 ST don't perform equaly.

I am intrigued by this one as I have just checked back the results for the first test and Defoe has not taken the DC positoin and his stats are also identical except for the FRee kick taking dropping one place.

Remember that player has to reach DC rating of at least 11 for this to happen. Maybe 1 year is not enough for that!!!

Also, bear in mind that the younger and more versatile player is, the easier is to reatrain him. Defoe is not so young (26yo) and his versatility is only 6.

[Pompeyboyz]

I wasn't able to run a second test today due to the missus not being out of bed in time for me to set it up so will try and run it over the weekend.

Remember that player has to reach DC rating of at least 11 for this to happen. Maybe 1 year is not enough for that!!!

Also, bear in mind that the younger and more versatile player is, the easier is to reatrain him. Defoe is not so young (26yo) and his versatility is only 6

I will run a two year test with the same conditions but run it with Defoe and One of the young guns at Portsmouth

[jayahr]

Now as we are wiser than before, I wonder, how it affects the way we play.

Are we retraing strikers to defenders or vice versa? Are we still buying D/AM Rs?

I refuse to do the former as it felt like cheating to me and thus would take the fun out.

However the latter I do. If those guys are there I do not hesitate to buy them, though in all honestly I feel a little itch in my conscience. Still, I just ignore it and play the way I would have played without knowing and sign whoever I would have signed anyway. Now I just know why players with great stats are cheaper when versatile.

The most awkward thing I had was when buying a really great striker for my Wolves team for very cheap. Only after signing him I saw he was 'competent' as AM RLC, MRC, DM, WRB, DRC as well.

How do you other guys behave now?

[Pompeyboyz]

I carry on as normal. The retraining thing I tested and saw it took 2-3 seasons before the SC had DC at a decent enough level so it wouldn't have too much of an effect

[Majkee]

Now as we are wiser than before, I wonder, how it affects the way we play.

Are we retraing strikers to defenders or vice versa? Are we still buying D/AM Rs?

I refuse to do the former as it felt like cheating to me and thus would take the fun out.

However the latter I do. If those guys are there I do not hesitate to buy them, though in all honestly I feel a little itch in my conscience. Still, I just ignore it and play the way I would have played without knowing and sign whoever I would have signed anyway. Now I just know why players with great stats are cheaper when versatile.

The most awkward thing I had was when buying a really great striker for my Wolves team for very cheap. Only after signing him I saw he was 'competent' as AM RLC, MRC, DM, WRB, DRC as well.

How do you other guys behave now?

I am playing in the same way as I have played it before. I know, which is my weakest position in team and needs upgrade, so I send scouts to the world to find someone. If he is good enough and I have money, I will try to sign him without inspecting his positions and level of his skill on these positions.

[SnUrF]

I read through 3 pages but got kinda bored. But if I understand it corretly this would mean that..

1. If I'd have MC whos preferred foot would be right only, and would have good crossing, dribling and good wingers attributes in general. Normally I'd be tempted to retrain him as right winger. But doing so, with the weighting of his preferred foot and winger attributes changed, I'd actually risk him getting worse. Thus no point in retraining?

2. If I had right footed AML, and left footed AML with similar CA. The right footed more suitable for cutting inside, and the left footed more of a byline crosser type. This would mean that the right footed would in general have better attributes (and be a better player) because of the weighting of the preferred foot? (This would mainly just affect the AI in my games since I don't use any editors or such so wouldn't know the CA of a player anyways).

Infact it would affect my games also, as it would mean my scouts would consider inside cutting wingers to be poorer than the crossing wingers.

Sorry if these topics have already been discussed here, but as I stated earlier, I got a bit bored after 3 pages (that I read right after reading the whole of the 'how important is 7 star coaching' thread).

-SnUrF

[Tomo]

How do you calculate CA of a player? Which program do you use? And what position retraining should I avoid? If I understand it well left footed AML with strong right foot will lose his attributes if he is trained for AMR (which I use quite often because I want my wingers to swap position)? How about winger for ST and vice versa?

[phnompenhandy]

CA isn't a calculation - it's a figure. You need an editor to see it - the pre-game editor shows the initial CA; an in-game editor will show current CA.

[Kewell08]

1. If I'd have MC whos preferred foot would be right only, and would have good crossing, dribling and good wingers attributes in general. Normally I'd be tempted to retrain him as right winger. But doing so, with the weighting of his preferred foot and winger attributes changed, I'd actually risk him getting worse. Thus no point in retraining?

That's the killer. A lot of this doesn't have much effect without deliberately aiming to achieve this, but your example proves exactly why this needs to be fixed. The fact that retraining someone to a position which they have the stats for makes them worse... is ludicrous.

[Granty]

I am a developer and have a great alternative for the people who find these small problems ruin the game for them.

My game is Fussbal Runner and although we have secured no rights to the English football league we do have full rights to the Kazakhstani, Mongolian and Fijian National leagues while having all of your favorite clubs almost but not quite there, you just have to imagine, e.g. we have Kydneypool with Fernando Tornado and Steven Gerhart from Germany, also we have Manbreaster unsighted for whom we have Shreck Gooney and Tall poles he's from poland in our version you understand we must take care not to infringe any copyright laws.

Ok you got me maybe there is no game just thought this thread needed a little lightening up and to point out that while there has been a vital flaw highlighted that does not mean the integrity of the game has been spoiled so long as you continue playing the game in a realistic way and so long as you avoid the editor it is so easy to do, I am glad that FM continues to evolve and so long as they stick to their promise and sort this out for FM10 then this will not have chance to affect too many people in that short a time, hey it could be worse as I am sure a game of the above quality is possible to whack together within a week, anyone interested? Lol

well done Ljuba too for highlighting this

[crafty bison]

^ What on EARTH are you on about?

[Dan-Erik]

What i now would be interested in would be the weightings of the ratings for each position?

Diablo-esque number crunching is right down my alley .

[muncherdave]

Granty you are a random soul, but thanks for bumping this thread, it was a great read other than the trolls on page 1 waiting to jump on a foreign person because he got excited with his typing that is. Props to Ljuba82 for discovering this

[leroybrown]

well ive read the entire thread and from what i gather the original poster claims by simply retraining a player in to a new position until he is effective at said position, the stats from the original position should begin to increase again through training as CA/PA points will be allocated differently since he is no longer just a one-position player...mmm?

Does make sense in theory but putting it into practise could prove different

ill try and retrain a bunch of LLM nohopers who have CA=PA give or take a few points and see what happens, i noticed someone did post they were testing but no further word

Oh im playin 08 btw, 09 runs too slow on my pc for me so i uninstalled it

Ill use Tugs training for some L2 side and report back

Is any of the main posters in this thread still interested in its findings?

[moj]

This has been well known for some years if you think about it, look at players who can play MR, ML, MC, AMR, AML, AMC, DMC and FC, guys like Wilshere, they turn out to be amazing.

But a very good find none the end, another flaw to add to the calamity-like flaws that FM has, and now anyone can have a world-class player with 150CA.

EDIT: Check out Fabregas's and L. Diarra's stats, Fabregas has the higher CA + PA, but Diarra is the more effective player in FM09, just compare them.

[zeusbheld]

...

Is any of the main posters in this thread still interested in its findings?

not a main poster, but interested.

[KingLara]

I tried to do a simular thing. my idea was to create a player with nearly all his stats 20 and put him in a championship team and see if he could get them promoted on his own. Unfortunatly his stats were 15 and 16 when i put his ca200 pa 200 and nearly all stats 20.

Haven't tried it since then.

[NepentheZ]

Did any of the tests from this thread become complete? I'm very interested in the outcomes...

[Blanchflower1]

I tried to do a simular thing. my idea was to create a player with nearly all his stats 20 and put him in a championship team and see if he could get them promoted on his own. Unfortunatly his stats were 15 and 16 when i put his ca200 pa 200 and nearly all stats 20.

Haven't tried it since then.

I too faced the same problem!!

[slipknot67]

Is any of the main posters in this thread still interested in its findings?

I'm not a previous poster on this thread, but I'd like to register my interest in the findings of this experiment:thup:

[PaulC]

Just to keep people up to date, we are tweaking the weighting system in Fm2010 to take account of very versatile players which will hopefully see a better balance of stats in their profiles.

[el sid]

Just to keep people up to date, we are tweaking the weighting system in Fm2010 to take account of very versatile players which will hopefully see a better balance of stats in their profiles.

Good to hear

[slipknot67]

Just to keep people up to date, we are tweaking the weighting system in Fm2010 to take account of very versatile players which will hopefully see a better balance of stats in their profiles.

I was just wondering Paul, if the isssue of regen two footedness is something that is also being adjusted for the next game, this seems to be one of the hottest topics on the forums atm and although slightly off topic, as it concerns attribute weighting I thought I'd ask. Apologies if this has been asked and answered elsewhere btw.

[xx11]

I was just wondering Paul, if the isssue of regen two footedness is something that is also being adjusted for the next game, this seems to be one of the hottest topics on the forums atm and although slightly off topic, as it concerns attribute weighting I thought I'd ask. Apologies if this has been asked and answered elsewhere btw.

been answered many times in the 2 threads in recent days about 2 footedness.

The weighting for 2 footed players is being re-assessed on a position by position bases and changes are being investigated. Currently weighting is the same no matter what position, however it may be made different for each position (this part my guess, keeper = hardly any weight, Defensive positions = small weight, offensive positions = similar to currently with just a slight reduction on weighting)

[slipknot67]

been answered many times in the 2 threads in recent days about 2 footedness.

The weighting for 2 footed players is being re-assessed on a position by position bases and changes are being investigated. Currently weighting is the same no matter what position, however it may be made different for each position (this part my guess, keeper = hardly any weight, Defensive positions = small weight, offensive positions = similar to currently with just a slight reduction on weighting)

Thanks for loooking that up for me xx11:thup:

[NepentheZ]

Did any of the tests from this thread become complete? I'm very interested in the outcomes...

Yeah, what I said back then

[OugaBagga]

There are players that used to be strikers as youth players and switched to DC or others that used to be DC as youth players and switched to being strikers. A big example would be Gallas.QUOTE]

Or Rio Ferdinand

[Ljuba82]

I hope this bug is fixed for FM10...

[addicted86]

Of course there are a few rare examples of players that have changed position over the years, like Nesta (striker --> DC), Gallas (striker --> DC), and Barthez (striker --> GK) but it is safe to say that they are all not much of a goalscorer anymore. Like MSCCG said.

this is not a rare situation at all, you would be able to find a plethora of examples like this. in fact i think you'll find its extremely common. e.g. matty taylor was a striker but when luton picked him up they turned him into a LB/LM. my brother is another (but less high profile) example

fact is, in youth football, where do the majority of the best players play? ST or AM. therefore i agree that this is a huge flaw, and needs to be looked at. a natural continuation of this is that we should also have more freedom as to what postions we could teach 14/15/16 year old regens to become proficient in.

[lance101]

Hi guys!

The findings in this thread are very interesting, but unfortunately also point to a major flaw in the way the game is designed. I was thinking about this for a while and I think I devised a solution that would in one hand solve the current "exploit/bug", as well as improve the match engine in a different way, while, hopefully, not making the game run much slower.

First I will outline how I THINK the game works, and then I will provide my alternative.

In FM08 and 09, I think the CA-attributes adjustment model works in the following way:

1 - Player i has his attributes in accordance with his CA.

2 - Time passes (two weeks? a month? I wonder if anyway outside of SI Games actually knows this)

3 - After a certain period passes the game looks at the player's performance, number of games played, age, mental attributes, training intensity, quality of facilities, etc and adjusts the CA up, down, or leaves it unaltered.

- If the CA goes up, the attributes in general will go up, with the increases "shaped" by the player's position - in case of human manager players, this change will probably take into account the human manager training schedule.

- If the CA goes down, the same applies as before, but the attributes in general go down.

- If the CA stays the same, the attributes stay unchanged ,in pc managed players, while possibly are reshaped according to training schedule in human managed players.

Now, at some moment, before or after the change in attributes, the game as to check if the attributes are too high or too low for the CA... How do I know this? Because if you edit a player using a savegame editor, making for instance all attributes 20, the game will lower the attributes automatically, even if the CA of the player goes up or is unchanged. So, currently, the game has to calculate:

CA = SUM over all attributes [attribute i * SUM over all positions {weight of attribute i in position j * proficiency in position j / Sum of proficiency in all positions } ]

4 - Time passes...

etc

Now, how would my system work:

1 - At the beginning of the save the computer calculates the implied CA (ICA for short from here on) of a player in each position, taking into account the attributes and the positions proficiencies.

- The position where the player has the higher ICA is thereafter considered the player's best position (and is quoted as such when you look at your assistant report), even if it is not the position the player is natural on!

- If the ICA of the best position is above the CA, the game adjusts down the players ratings, if the ICA of the best position is bellow CA, the game adjusts down, if it is the same, the game does nothing.

- The game also stores for reference (of our assistant managers, coaches, and of all the AI managers) the ICA of all other positions (or if that is too cumbersome for the game, just the ICA of positions where the players have more than 15 or 12).

2- Game starts.

3- Time Passes

4- Game adjusts the CA in the same way as before, but when checking if attributes are too high for CA, the game would need to do that calculation for each relevant position, so every thing is fine as long as:

CA >= Sum over all attributes [ attribute i * weight of attribute i in position j ] * quality in position j, for any position j, with the equation being in equality for at least one j

(if due to retraining or reshaping of the attributes a different position is now the one with the highest ICA, and that ICA is above the CA, the game should thereafter name the new position the player's best position, and reduce the attributes in order to the ICA of this new position to equal the CA)

Now, it might seem that there are more calculations to be done here (especially for players with many accomplished positions), but at least one thing would be simplified - each calculation would involve a smaller set of weights - so the game would probably be slower due to this, but possibly not as much as feared... Now, I think there is another area where this method would get some time saving, as well as improving AI tremendously.

Right now, when deciding to pick a player or not, and where to play it, (ignoring for a moment morale, form, match fitness, etc), the AI has to look at the player's CA and the player's positional proficiency. Admitting for a moment that the way the positional rating is implemented is very rudimentary, say that in the game Alex Ferguson looks at Rooney and thinks like this:

At FC he can play with quality 175 * 20/20 (CA * FC positional rating/20)

At AM C he can play with quality 175 * 14/20 (CA * AM C positional rating/20)

etc etc

So, now the game has to do that multiplication of CA by positional rating... if my method was implemented the AI would simply need to consider the CA for the desired position when deciding whether to use the player, which would be faster. This would also have the MAJOR advantage of having the AI understand what players are better for which positions (and for instance not think that Fazio would be as good a MC as a DC if someone changed Fazio to have MC = 20 in the editor)

Finally, if the speed changes I point out are not enough to compensate to the slowing caused by the calculation of more CA, I would say that it would be worthwhile to do the CA-attributes check less often, and have the game change the players less often - the increase in realism due to a better AI selection would be worth it in my view.

[L.McDaDe]

What level for the new position should you train him to ? COMPETENT ?

[RBKalle]

But...

If it takes a couple of seasons before the ST is competent enough as DC to actually exploit the flaw in the attribute distribution/weighing, won't he have peaked already at that stage?

If attributes go up when CA is still going UP, what happens if his potential has been fulfilled already and there are no more points to be distributed?

BTW, does training (schedule training, not positional one) actually have an impact on which stats go up or down? I've always thought so, but somebody in this thread said it doesn't...

[lance101]

This problem seems to have finally been sorted out in FM 2010. As far as I can tell until now, players with multiple positions do not seem to have better attributes than players with just one position, and attribute quality seems to correlate fairly well with CA, but I guess we will only be sure when the full game is released and we can play with the data editor.

[Mvriezekolk]

What a coincidence, both player have the exact same birthdate, height, weight, caps, goals, value and salary. Seems like a real time editor has been used on this player.

[dovEs]

^ you are being sarcastic aren't you?

[jayahr]

What a coincidence, both player have the exact same birthdate, height, weight, caps, goals, value and salary. Seems like a real time editor has been used on this player.

Of course there was

Please read through the thread. It's very interesting and should explain you why and for what purpose the editor was used.

[bjoggi33]

Does this still work in FM10?

[bjoggi33]

Anyone ?. ^

[Lordarod]

try it yourself but probably not because PaulC had the perfect solution.

[lance101]

I am not so sure it was the perfect solution... indeed multiple position players no longer benefit from having some low attributes for key attributes for one of their positions... but this was solved (at least according to the tests I have done for myself) by having an attribute take the max possible from its CA for any of the positions he his good at, in practice making players that are versatile worse than non versatile players.

[PaulC]

No it wasn't done that way. It basically reduces the average attribute slightly for unusually versatile players.