Positional training - wrong? (long post)

Just a few things...

I decided to train one of my young regens, a right-back but two-footed so I retrained him at left-back as well (always a good thing to do when seeking out loans too). After a season, he became accomplished in that position. I then sent him out to Birmingham on loan where he played 12 times at left-back, 28 at right-back. However, the player does not train this new position while on loan (and no player does) so halfway through the loan, he "forgot" the position, dropping down to "awkward" (despite playing a fair amount there anyway).

This strikes me as wrong because:

1) You forget things a lot less readily than you learn them

2) You only forget something entirely after learning it to near-perfect levels after several years, possibly decades

3) Although he never trained his new position on loan, there isn't much else he needs to learn and if he's still playing there, that should be enough to keep the knowledge there

The whole positional thing is weird in-game anyway - there's "position" and "new position" as FMRTE calls it. "Position" is the position either set in the database (for real players) or generated upon regen creation (for regens) and is taken as the true positional sense of the player. "New position" is where a player learns new positions in-game, and goes up-and-down but the "new position" knowledge never goes below the "position" knowledge as "position" is the true positional sense. So let's just say Rooney is ML 12 in the database - if you retrain him to accomplished, say 18, then take him off the retraining, it will drop back to 12 but no lower.

In addition, while you can retrain a player to "natural" with a bit of work, start retraining to a new position and after a while it drops to accomplished (as "new position" > "position").

So in other words a player is always constrained by his original positions. I think this is really wrong. I would take the following "guidelines":

1) The more you learn, the harder it is to forget.

2) You never forget something fully unless you never learnt it properly in the first place. And even this may take time.

3) The game needs to make sure that you can actually learn different positions while retaining old ones without penalty.

4) Even if you stop learning something, but continue to practice (i.e. play in that position), you will keep your (imperfect) knowledge intact.

1) implies that if a player retains to natural, then they will not forget it over a season. Arguably they may never forget it at all. So I would recommend something like:

- If a player stops retraining and never plays in a natural (20) position ever again, 20 will go to 15 in about 4 seasons, then to 10 in about 8 seasons

- If a player stops retraining and never plays in a accomplished (18) position ever again, 18 will go to 13 in about 3 seasons, then to 9 in about 6 seasons

- If a player stops retraining and never plays in a competent (15) position ever again, 15 will go to 10 in about 2 seasons, then to 7 in about 5 seasons

- If a player stops retraining but plays there continuously anyway, the positional knowledge should remain as-is, possibly even increasing but slowly (as learning > learning on the pitch)

- If a player stops retraining but plays there occasionally, the positional knowledge should remain roughly as-is, possibly dropping by 1-2 points or still possibly increasing but slowly (as learning > learning on the pitch)

2) implies that if I learn Spanish properly, I will never forget some things - even if I will be restricted to numbers and ordering McDonalds; the same cannot be applied to someone who half-heartedly learns Spanish who well remember a lot less down the line. I recommend that, say, the game should log the maximum positional knowledge for a position and the positional knowledge for this position should never drop below half of this rounded down. So if I learn a position to natural (20), it will never go below 10, 18 -> 9, 15 -> 7, 10 -> 5, 2 -> 1 etc. So if a player is a natural ST but doesn't play there for two decades, it may drop to 10, but he will know somewhere deep down how to play there, but relearning this position will be substantially less painful than someone who knows nothing about that position at all.

3) is rather simple, although it must account for loans not learning new positions. This is where we need much slower forgetting of positions. My regen above doesn't need to "learn" he's a left-back as well as a right-back - he simply "is" a full-back capable of playing on both flanks. He shouldn't need to learn anything at Birmingham in the short-term to keep this positional knowledge up. Perhaps loanees should lose positional sense even slower than normal because their main objective is to make it for the first-team, not the loan team.

4) is a bit like 1). If I play a single game in a position, that could be considered enough to refresh my knowledge of a position. So it might simply reset a decay counter of some sort (i.e. in 3 days you will drop positional knowledge by 1 - but playing resets this counter to 100 days).

All in all, I point to scrapping "position" and "new position", keeping one positional "grid" and another grid with maximum positional knowledge gained. Positional knowledge should not fall below some minimum value based on the maximum positional knowledge gained (i.e. floor of half). Positional knowledge lost needs to be lost much, much slower - over several seasons, not months; depending on how much you learn too. Loanees need to forget positional training learnt at their original club much slower.

For maths-type people, the rate of decay should be inversely proportional to the amount of positional knowledge and maximum positional knowledge, and the minimum positional knowledge should be proportional to the maximum positional knowledge.

Some good thoughts here. I'll make sure SI see it

The positional training annoys me so much that I have even taken to using an Editor to change their positions once the training is complete.

I currently have a ML who I decided to train as DL to give me some cover, he now plays regularly as my DL sub, playing maybe 30 minutes in most matches (as my first choice DL is cream crackered by half time).

I then had an injury problem and started playing him in DM and training him there, he promptly started forgetting how to play in the DL position.

I have a problem with this sort of rapid decline, he has lost ability in DL twice as fast as he learnt it.

I do hope this gets sorted in a future patch/version, your opinions are pretty close to mine.

The other thing I forgot to mention is that if I play a player in a new position but he didn't learn beforehand, he will still gain "some" knowledge of that position, perhaps awkward but something nevertheless. We see this with centre-backs being stuffed at full-back or defensive midfield in emergencies, sometimes multiple times. Learning a new position in this way should be slower than if you were learning through training of course, possibly even capping itself as if you want to learn it properly, you need to do it through training which can cover all the bases.

So if a D C has no knowledge of D R, if I play him at D R in an emergency because of injuries, he could perhaps gain positional knowledge 1 at D R, and if I keep playing him there, he may well become "competent" after a whole season playing there (depending on versatility of course) - even if I don't retrain him.

Another idea is "total knowledge" which is dependent on the player's versatility. With this sort of training a player could well end up being accomplished in tons of positions. With "total knowledge", it could cap the total positional knowledge (perhaps weighted) over all positions so if you learn too many positions the player will forget some others as you are overloading him.

Another idea is that if you learn D L to accomplished and it takes x days, then you will learn D R in slightly less time than x days because it's just a mirror image of D L - the rough idea of playing as a full-back is there, just that your mind needs to mirror itself too.

Another idea is that, say, if a player is accomplished at D L and M L, then he should be able to learn WB L and AM L quicker than usual as they are similar positions to what he already knows.

There are some very good points here.

In response to your very last point: I think regens and indeed current players need to be given positions far more liberally. It annoys me when I see a DL with no rating for WBL, DC or DR, more so when I see a DM/AMC with no rating for MC...

With real players, I think researchers should give ratings of 5-12 for positions a player may be some use in, even if he hasn't actually played there. I think regens should have "position sets", rather than positions, and 95% of these should include 2 or more positions at some rating.

For example, we could see more players with DR 20, WBR 16, MR 14, AMR 11, or ST 20, AMC 18, AMR 15, AML 12, MC 10, ML 8, MR 8. Wingers and full backs should have some ability on the other side, MCs should have some ability in one or both of DM and AM.

This is an area that has improved massively in the last few versions, but I'd like to see it come on more.

Really good set of posts. I hope that it's something SI will consider looking at and implementing in some fashion.

Great post.

I do think this an area that could do with improving, and as SCIAG said it already has come on greatly and seems to be going the right direction.

The weirdest regens I've seen are those with 20 rating for WB but nothing for M/D. Such a limited player would never exist surely? Similarly I've seen far too many AMR unable to play MR or AMC unable to play MC, and with positional training often needing to be a continuous process to be effective something seems amiss.

Great post.

I do think this an area that could do with improving, and as SCIAG said it already has come on greatly and seems to be going the right direction.

The weirdest regens I've seen are those with 20 rating for WB but nothing for M/D. Such a limited player would never exist surely? Similarly I've seen far too many AMR unable to play MR or AMC unable to play MC, and with positional training often needing to be a continuous process to be effective something seems amiss.

I think those are weird too. With the lack of wing-back formations and wing-backs in general, there are no pure wing-backs. Even the likes of Roberto Carlos had to adapt their game somewhat for their clubs as they didn't always play wing-back formations.

I think youth players should sort of generate players based on what you already play - so if you don't play wingers, you should get less wingers coming through or the wingers that do come through are capable of playing in various positions like midfield or full-back.

My latest batch included 2 - yes - 2 - left wing-backs.

I really like what I have read here and would like to add a thought. I coach high school IRL, so I recognize that I am not a professional coach; however, once the game starts, the only time a player is actually "in his position" is after a restart (i.e., after a goal, halftime, set piece, etc.). The rest of the time, he covers space or a man dependent on the coaching philosophy. If a player can learn those marking areas for attacking/defending purposes, it just doesn't matter how well he plays WB/L v. DL because we are now talking about minute nuances that have more to do with tactics than positional knowledge.

I think this idea is fantastic. My only concern is what is to stop players being trained to every position?

I think this idea is fantastic. My only concern is what is to stop players being trained to every position?

There could be a cap for 3, 4 or 5 different positions. We can probably all think of someone who can realistically play at several different positions without much drop-off in skill.

I think this idea is fantastic. My only concern is what is to stop players being trained to every position?

The "total positional knowledge" idea I put up will solve that, which also doesn't really need a cap on positions which is very iffy. For example, you could imagine a player who is awkward in every single position except one (where he is natural).

However, there are players who are flexible enough to play in tons of positions and I'd argue this is where versatility can be used more effectively.

I wouldn't cap positional ability. Maybe you could do it in combination with versatility- 20 versatility means you can potentially have 100% position points (not including GK), 19 means you can have 95%, 2 means 10%, etc.

One of my biggest problems with this issue is what happens when I retrain a player who has already reached his PA. I have 24 year ol who is a natural DC and accomplished DL, and has reached his PA. The idea that he would somehow have to become worse at crossing or shooting or pace or some other attribute to become a "natural" at playing DL seems absurd.

One of my biggest problems with this issue is what happens when I retrain a player who has already reached his PA. I have 24 year ol who is a natural DC and accomplished DL, and has reached his PA. The idea that he would somehow have to become worse at crossing or shooting or pace or some other attribute to become a "natural" at playing DL seems absurd.

I agree with this. Also tieing CA/PA to a player's ability to use their weaker food seems equally absurd.

I agree with this. Also tieing CA/PA to a player's ability to use their weaker food seems equally absurd.

So if you have two players, identical in every way, except one of whom can use both feet and one can only use one, they'd both be equally useful?

So if you have two players, identical in every way, except one of whom can use both feet and one can only use one, they'd both be equally useful?

No. The two-footed player would be better player because his technical attributes would not be penalized. The player with a weaker foot would have his technical attributes suffer when attempting to pass, shoot, dribble, etc. with the off-foot.

However, I believe with the current system if a player is two-footed or is strong with his 'weaker' foot, then it takes 'points' away from his CA/PA. This would mean that if you have 2 identical players with 170 PA, the player who is only strong on one foot would have the potential to become a better passer, dribbler, scorer, etc. on one foot than the other. The two-footed player would never have the potential to have a foot as strong as the first player.

Very good post, I concur

You're looking at it the wrong way. Players are not given a CA and then attributes, the attributes "add up" to give the CA.

Does a striker with high finishing have points "taken away" from his pace, or technique? Nope.

SCIAG > Everyone else.

SCIAG for mod.

SCIAG > Everyone else.

SCIAG for mod.

stop licking his boots dude its lamme

stop licking his boots dude its lamme

What's lamme?

I actually think you shouldn't be able to retrain to many positions. It should be capped at 1, possibly 2. Otherwise we'll end up with players who are natural or accomplished in most positions. That would be utterly unrealistic.

It should take longer to get from Inefficient to Accomplished (and Natural), and the decline should be much slower as well. But we shouldn't be able to retrain players to a wide range of positions.

I think it makes a lot of sense that the game "penalises" players with a weaker foot compared to two-footed players. A player that can pass, shoot and dribble with both feet is of course an advantage over one who can't. So this should be reflected somehow. Don't know how the game does it, but glad it does.

You're looking at it the wrong way. Players are not given a CA and then attributes, the attributes "add up" to give the CA.

Does a striker with high finishing have points "taken away" from his pace, or technique? Nope.

Maybe I didn't make my post clear. Of course attributes 'add up' to give CA. The PA puts a limit to how many points a player can have.

If I understand this correctly and foot strength also adds up as part of CA, then wouldn't a player with 20/20 in feet have less points to award to the rest of his attributes than a player with 20/10 in feet strength before the CA reaches PA. Now I understand most players never reach their PA anyway, so this point may be moot I just think it's a rather strange system of determining a player's potential.

Maybe I didn't make my post clear. Of course attributes 'add up' to give CA. The PA puts a limit to how many points a player can have.

If I understand this correctly and foot strength also adds up as part of CA, then wouldn't a player with 20/20 in feet have less points to award to the rest of his attributes than a player with 20/10 in feet strength before the CA reaches PA. Now I understand most players never reach their PA anyway, so this point may be moot I just think it's a rather strange system of determining a player's potential.

You do understand more or less correctly.

Likewise, a player with 20 for decisions has less "points to award" than a player with 10 for decisions. The strength of a weak foot, much like decision making, balance or finishing, is a skill. Therefore, it comes out of the CA.

I actually think you shouldn't be able to retrain to many positions. It should be capped at 1, possibly 2. Otherwise we'll end up with players who are natural or accomplished in most positions. That would be utterly unrealistic.

I think it would be impossible anyway, because the time spent re-training and playing out of previously learned positions would mean the lad would forget his old position. I do see your point though, I could easily imagine an DC/DM/AM/CF (which actually sounds amazing).

Shouldn't the poition re-training depend quite heavily on stats? For example, you couldn't train a 5ft 8' winger with 5 heading and 3 tackling to play centre half, could you? I mean you wouldn't, if you had in common about you, but surely you couldn't either?

What's lamme?

It's Danish for Lamb, ask google.

It would be easy. Make sure the player's positional abilities lower when they are not played in them, even if they were natural in the first place. A lot of players started in a certain position and ended up in another. This shows that position is on a continuous and players can have their abilities manipulated to suit a new position.

I actually think you shouldn't be able to retrain to many positions. It should be capped at 1, possibly 2. Otherwise we'll end up with players who are natural or accomplished in most positions. That would be utterly unrealistic.

It should take longer to get from Inefficient to Accomplished (and Natural), and the decline should be much slower as well. But we shouldn't be able to retrain players to a wide range of positions.

I think it makes a lot of sense that the game "penalises" players with a weaker foot compared to two-footed players. A player that can pass, shoot and dribble with both feet is of course an advantage over one who can't. So this should be reflected somehow. Don't know how the game does it, but glad it does.

I think it should depend on the individual player, as people have said. Someone like, I dunno, Patrick Vieira shouldn't be able to retrain to more than 1 position if that - he's old and not versatile at all. On the other hand, look at a guy like Eboue. He was a centre back at Beveren, retrained to right back at Arsenal (and I'm sure he could play right wing back too), retrained again to right midfield (but still played regularly at right back), and he even played some central midfield / left back. If I wanted to retrain him back to a centre back, or to a defensive midfielder, I think that would be OK, even though he would know like 6 positions at that point.

Also, I like the idea of players "forgetting" old positions as they learn new ones. I think, personally, position retraining should be tweaked so that getting to competent / accomplished is quite fast, but getting to "natural" takes a year or two of constant retraining and/or playing at that position. Then, if you stop playing / retraining that position, your rating there should decline slowly (over the course of sevearl years), but *never* below competent or unconvincing (to use my previous example - Eboue still has a rating at centre half despite not playing there for many years).