Concerning tactical issue

[Rashidi]

14 minutes ago, sirgiorgio said:

I don't think he gave in to community pressure; I think he gave the chance to anyone (AI included!) to build decent tactics using "players in those advanced wide areas"

I agree with Barside, he gave in. Fact is the engine was a lot tougher to beat in its earlier iterations, when the wider players didn't always track back. Now I can even make players with low bravery track back and put in a challenge, you just need to know which player to pick. They may not win the tackle, but they will become an extra body to navigate. Overall though you make a valid point about each player bringing something different to the game. The AI though has had the field levelled in some respects with the Tactical Creator. Before we had the TC, we had people asking for all kinds of options to do things, human invention will more often than not trump what the AI can do. The AI is still a long way away from being able to play well against the best players, it still has an average transfer policy, though I do see it improving slightly from older additions. This game is more than just tactics, its about building a winning legacy, the challenge I'd like to see is elevating the AI to the point where it can not only play by making tactical changes based on what I do in the game, but it can also develop a squad around a style of play and maintain a viable transfer policy to lock that in place as a lasting legacy, Honestly I believe if the AI had a way of not only copying successful tactics, but setting up effective transfer strategies, then this game would get a lot harder. If we can make players choose to stay at successful clubs instead of leave for the next dollar all the time then this game might just get a bit harder.

[sirgiorgio]

2 hours ago, Rashidi said:

I agree with Barside, he gave in. Fact is the engine was a lot tougher to beat in its earlier iterations, when the wider players didn't always track back. Now I can even make players with low bravery track back and put in a challenge, you just need to know which player to pick. They may not win the tackle, but they will become an extra body to navigate. 

But it's the same for human teams and the AI's, isn't it? In a world where 40% of the AI teams played 4231 or 433 and their wide players spent their time talking with the ball boys, it was much easier to exploit their weaknesses. It's not that it doesn't happen anymore: all those crazy 4312 and 433 narrow are goal fests, but again, I wouldn't ask for a more balanced ME, I would ask researchers to never set those formations for AI managers, or for AI managers to use those formations only when they're abolutely desperate to score and not as one of the many possible options. 433 narrow is clearly an "extreme" formation, 4231 and 433 are like the norm nowadays. And they are the norm exactly because there are dozens of ways to adapt them to different situations. If I have a Mandzukic on the AM L spot, why wouldn't I want him to run up and down like a wingback and tackle like a BWM? I certainly don't expect him to do what CR7 does. And if I have CR7 on the same spot, why would I want him to do the Mandzukic?

 If you're saying that there's some sort of imbalance that kind of forces an attack minded player on attack duty on the AM strata to always act like an out-an-out midfielder in the defensive phase, I personally haven't witnessed it (at least it doesn't seem that extreme to me) but I believe you. If you're saying that you can make your wide players as good in defense as Mandzukic and as effective in attack as CR7, that I would call it good management.  

[Guest]

1 hour ago, sirgiorgio said:

 I would ask researchers to never set those formations for AI managers, or for AI managers to use those formations only when they're abolutely desperate to score and not as one of the many possible options. 433 narrow is clearly an "extreme" formation, 4231 and 433 are like the norm nowadays

This is exactly what I would expect & until recently it's the approach that the German research team appear to have taken.

4231, 433 & such like should be set as the preferred attacking formations with the norm being the base setups of 451, 4141, 4411 which can easily hybrid into the the more attacking formations when in possession within the current TC. The stronger sides, with the right players should still start matches using  the attacking formations because of their dominant position but it would be less common than has been over the last 5 or 6 FM's

That's not what happened though, researchers started setting attacking formations as the default to the point where far too many managers has the same default & attacking formations.

This is most likely due to broadcasters starting to show formations in their possession form (451 doesn't portray an attacking ideal) which led to more fans thinking that was how teams actually setup & the inevitable complaints about lazy wide players started to appear, the end resulted were ME decisions that accommodated the poor interpretation of football tactics rather than correct the research data & attempt to educate the average FM'er.

[themadsheep2001]

31 minutes ago, Barside said:

This is exactly what I would expect & until recently it's the approach that the German research team appear to have taken.

4231, 433 & such like should be set as the preferred attacking formations with some the norm being the base setups of 451, 4141, 4411 which can easily hybrid into the the more attacking formations when in possession within the current TC. The stronger sides, with the right players should still start matches using  the attacking formations because of their dominant position but it would be less common than has been over the last 5 or 6 FM's

That's not what happened though, researchers started setting attacking formations as the default to the point where far too many managers has the same default & attacking formations.

This is most likely due to broadcasters starting to show formations in their possession form (451 doesn't portray an attacking ideal) which led to more fans thinking that was how teams actually setup & the the inevitable complaints about lazy wide players started to appear, the end resulted were ME decisions that accommodated the poor interpretation of football tactics rather than correct the research data & attempt to educate the average FM'er.

This absolutely nails it for me. In those t iterations of the match engine, the German researchers methods was the best way of doing it, and i would have open up some of the roles attached to AML/AMR up to the ML/MR strata. Even now i still think that's the right way to approach it

[Guest]

Having  more AM & ST roles also available in the strata below would be ideal, maybe lock it to the support duty options to reflect a more conservative focus.

[Rashidi]

1 hour ago, Barside said:

4231, 433 & such like should be set as the preferred attacking formations with the norm being the base setups of 451, 4141, 4411 which can easily hybrid into the the more attacking formations when in possession within the current TC. The stronger sides, with the right players should still start matches using  the attacking formations because of their dominant position but it would be less common than has been over the last 5 or 6 FM's

Agreed, it annoys me when I see so many sides now using the 4231 as the base when the base should actually be a 4411, which is what they are actually playing. The last thing people should trust are the graphics from broadcasters. Sometimes these graphics are done more out of convenience than accuracy.

[Guest El Payaso]

But using 4-1-4-1 instead of 4-3-3 or whatever you want to call it would more than likely influence negatively on amount of goals. A MR in my experience doesn't rush forward in the same way as a AMR does and rushing forward already at times seems too slow. This could also influence in team selections as the AI might not select players to those positions where they 'can't play'. 

With 4-1-4-1 it's quite easy to shut the door for the opposition and keep clean sheets. Imagine how it would be like if there were two AI teams playing against each other with 4-1-4-1. I would think that the amount of 0-0s would increase drastically. 

[themadsheep2001]

Just now, El Payaso said:

But using 4-1-4-1 instead of 4-3-3 or whatever you want to call it would more than likely influence negatively on amount of goals. A MR in my experience doesn't rush forward in the same way as a AMR does and rushing forward already at times seems too slow. This could also influence in team selections as the AI might not select players to those positions where they 'can't play'. 

With 4-1-4-1 it's quite easy to shut the door for the opposition and keep clean sheets. Imagine how it would be like if there were two AI teams playing against each other with 4-1-4-1. I would think that the amount of 0-0s would increase drastically. 

Scoring goals from 4-1-4-1 isn't a problem 

And there German researchers already had AML/R players with ML/R ratings 

[Guest]

I had no issues scoring when using 4411 or 451, it's all down to role/duty selection.

Edit: Of course the AI managers would need to be taught this but to be brutally honest that's what SI & the tactics testing community (I'm being critical of my own contribution on this) could have spent the last half decade doing rather than try to get a somewhat flawed base logic to work.

[HUNT3R]

Yeah, I've used both a 4-4-1-1 and 4-1-4-1 this year as weaker and stronger clubs and they can be deadly.

[Rashidi]

13 minutes ago, El Payaso said:

But using 4-1-4-1 instead of 4-3-3 or whatever you want to call it would more than likely influence negatively on amount of goals. A MR in my experience doesn't rush forward in the same way as a AMR does and rushing forward already at times seems too slow. This could also influence in team selections as the AI might not select players to those positions where they 'can't play'. 

With 4-1-4-1 it's quite easy to shut the door for the opposition and keep clean sheets. Imagine how it would be like if there were two AI teams playing against each other with 4-1-4-1. I would think that the amount of 0-0s would increase drastically. 

As opposed to this?

We have played 27 matches so far and united have scored 79 goals, no complaints there, but relegation bound Bournemouth played a 4231 against United? There are times when I wished some sides would default to playing something a bit more conservative when they are up against the league champions. OK so they wanted to play a 4231, why not a deep 4231 with 2 DMs? I reckon what we need is some form of decision making matrix that assigns a team a more conservative variant of the same system if they are to play against a side that is superior to them.  These things didn't happen in the first season of my save, it took me a few years to get to the premiership and I have only noticed them now, as I attempt to reign in this runaway giant.

[Guest El Payaso]

I have no doubts that us AI users, well maybe excluding myself here, will get the same result out of 4-1-4-1 but as you notice I was entirely talking about the AI. 

@Rashidi don't the AI managers also have the 'defensive formations'? For example Barcelona usually line up with a 4-5-1 but when they played for example against us away from home the formation was 4-1-4-1 which was set as the manager's defensive formation. Well needless to say that we basically witnessed a dull 0-2 game where Barcelona instead of winning through two defensive errors were really uninspiring as the distance between Suarez (up front) and Messi & Neymar was too big and the support was either too slow or non existent. I feel that also with my team at times but during this version I'm not really that bothered to start sorting it out as the system usually gives me those 1-0 or 2-0 victories. 

[themadsheep2001]

11 hours ago, El Payaso said:

I have no doubts that us AI users, well maybe excluding myself here, will get the same result out of 4-1-4-1 but as you notice I was entirely talking about the AI. 

@Rashidi don't the AI managers also have the 'defensive formations'? For example Barcelona usually line up with a 4-5-1 but when they played for example against us away from home the formation was 4-1-4-1 which was set as the manager's defensive formation. Well needless to say that we basically witnessed a dull 0-2 game where Barcelona instead of winning through two defensive errors were really uninspiring as the distance between Suarez (up front) and Messi & Neymar was too big and the support was either too slow or non existent. I feel that also with my team at times but during this version I'm not really that bothered to start sorting it out as the system usually gives me those 1-0 or 2-0 victories. 

As we've said, the Bundesliga was that way when the German researchers we doing it. It played some very well rounded football. AI selection in those shapes was no more an issue there than in a 4-2-3-1 etc. 

[sirgiorgio]

14 hours ago, Barside said:

Having  more AM & ST roles also available in the strata below would be ideal, maybe lock it to the support duty options to reflect a more conservative focus.

Yeah, that's a part of the problem (and again, it is not ME-related). The perfect example would be Thomas Müller, who even has a position named "in his honour" (Raumdeuter), yet the AI plays him in the MR/L spot most of the times.

But there are two more things to consider:

1. Attributes: If I had to hang on to a result, I'd very much prefer an AM R/L with high teamwork, workrate, tackling etc. than a M R/L with low teamwork, workrate, tackling etc. In other words, it's a bit too simplistic to say that a 4411 is by default "more defensively solid" than a 4231.

2. If I play with three at the back and two wingbacks, I might expect my wingbacks to be more or less attack minded, but I mustn't expect them to be out-and-out fullbacks or out-and-out wingers. I think the same can be said for attacking midfielders. Defenders, midfielders and strikers are what they are, defensive midfielders and attacking midfielders are supposed to be a bit of both, and be more or less involved in one phase of play than in another depending on a number of factors. But they shouldn't ignore their "midfielding" or their "attacking" unless they're explicitly asked to.

I was quite vocal in explaining that 352 doesn't mean "3 defenders and 5 midfielders" and I'm happy to see that the researchers and the users now set the "correct formation"; but the same can be said for 4231: you can describe it as 4411, 4231 or even 424 depending on how cool you want to sound, because attacking midfielders can be seen as midfileders or forwards; but they are attacking midfielders, just like wingbacks (formerly known as DM R/L) are wingbacks.

[themadsheep2001]

18 minutes ago, sirgiorgio said:

Yeah, that's a part of the problem (and again, it is not ME-related). The perfect example would be Thomas Müller, who even has a position named "in his honour" (Raumdeuter), yet the AI plays him in the MR/L spot most of the times.

But there are two more things to consider:

1. Attributes: If I had to hang on to a result, I'd very much prefer an AM R/L with high teamwork, workrate, tackling etc. than a M R/L with low teamwork, workrate, tackling etc. In other words, it's a bit too simplistic to say that a 4411 is by default "more defensively solid" than a 4231.

2. If I play with three at the back and two wingbacks, I might expect my wingbacks to be more or less attack minded, but I mustn't expect them to be out-and-out fullbacks or out-and-out wingers. I think the same can be said for attacking midfielders. Defenders, midfielders and strikers are what they are, defensive midfielders and attacking midfielders are supposed to be a bit of both, and be more or less involved in one phase of play than in another depending on a number of factors. But they shouldn't ignore their "midfielding" or their "attacking" unless they're explicitly asked to.

I was quite vocal in explaining that 352 doesn't mean "3 defenders and 5 midfielders" and I'm happy to see that the researchers and the users now set the "correct formation"; but the same can be said for 4231: you can describe it as 4411, 4231 or even 424 depending on how cool you want to sound, because attacking midfielders can be seen as midfileders or forwards; but they are attacking midfielders, just like wingbacks (formerly known as DM R/L) are wingbacks.

There is no real reason (any coding issues aside) why the AML/R and ML/R shouldn't have the same roles. Certainly something that needs to change 

[Svenc]

4 hours ago, themadsheep2001 said:

As we've said, the Bundesliga was that way when the German researchers we doing it. It played some very well rounded football. AI selection in those shapes was no more an issue there than in a 4-2-3-1 etc. 

The typical formation here in the BL tends to be the 4-4-1-1, except for Guardiola, here FM 2015. (4-1-4-1), with Lewandowski oft kinda hanging thus because of the duty selection  (frequently switching that to a two forward formation during matches, apparently if results didn't meet expectations then). There's also got a bit of a reason why on FM 2016 too he oft struggled to score like more than 50/60 goals from 34 matches... which is precisely the kind of "competition" where I fully agree with Rashidi. The elite should be a tad more competitive. Of course, now that he's in England he's immediately adopted a fav formation where there's a heavy front load (the research seems to have switched it pronto).




Apart of that, the German research were the only ones back then who took a thread by THOG on that completely literally. This highlights a couple of issues:

1) Synchronizing the research. One part of it running in one direction, the other running into another [speaking of which, the German research in my opinion is the only one that handles "forward runs" PPMs as they should. Those are the definition of (positionally) illdiscipline imo considering that all of that can be achieved via roles/duties etc., and more importantly, a guy on defend won't stick to his job as much

2) Long-term ME development. Not saying that there isn't but firstly there's got to be a long-term goal, and secondly for any of that stuff rechecking with their football contacts is the way to go. Feedback is vital, also in terms of actual playability from the player base. But to exaggerate slightly, if Paul had too "soft" a heart here, in two iterations there would be 10-10 scorelines universally, as somewhat decent chances, one on ones, etc. are finally converted at rates that don't cause as much frustration anymore

3) This thread runs dangerously close to needing [AI] spoiler tags  (I'm reluctant to peek anymore myself, and I've experimented how attributes link to their decisions with data experiments et all before).

[sirgiorgio]

29 minutes ago, themadsheep2001 said:

There is no real reason (any coding issues aside) why the AML/R and ML/R shouldn't have the same roles. Certainly something that needs to change

And there's no reason why they should, unless it's made perfectly clear that formations represent the defensive shape of the team. Which is debatable, both from a gamer's point of view and from the point of view of someone who watches football, reads about football and possibly even works in football. And it would be a sort of a revolution for FM, involving every single aspect of the game.

If I had to choose, I'd prefer an ME that adapts (as much as it can) to the common football jargon than a game that becomes more and more esoteric in an attempt to follow its own ideal of football (or, even worse, to hide potential limits in the ME).

 I'd seriously ask experts (players, coaches, or, worst case scenario, journalists) if they are more puzzled by FM's idea of "positions", FM's idea of "formations" or the ME's "representation" of certain positions and formations.

[themadsheep2001]

18 minutes ago, sirgiorgio said:

And there's no reason why they should, unless it's made perfectly clear that formations represent the defensive shape of the team. Which is debatable, both from a gamer's point of view and from the point of view of someone who watches football, reads about football and possibly even works in football. And it would be a sort of a revolution for FM, involving every single aspect of the game.

If I had to choose, I'd prefer an ME that adapts (as much as it can) to the common football jargon than a game that becomes more and more esoteric in an attempt to follow its own ideal of football (or, even worse, to hide potential limits in the ME).

 I'd seriously ask experts (players, coaches, or, worst case scenario, journalists) if they are more puzzled by FM's idea of "positions", FM's idea of "formations" or the ME's "representation" of certain positions and formations.

It has been made clear the formations are the defensive shape in FM. This is the very crux of the issue. Which is why the Wide strata's should be blended as one aera. As someone who did football match analysis, they are a mish mash, which is where the confusion comes from. But they certainly are not the attacking shapes generally.

FM's use of formations is incredibly static, and until we start showing all 4 phases, the German Researcher way is one of showing it

[westy8chimp]

9 minutes ago, themadsheep2001 said:

It has been made clear the formations are the defensive shape in FM

Not sure that's true. As mentioned... putting players in their preferred role would pretty much mirror exactly how SKY/BT etc pundits show a team setup.

I can't see anywhere in the game that states the formation is defensive shape (it may well be there, but certainly not 'clear') is it there?

My brother and I have been playing every iteration of CM/FM... I only joined this forum this year and was the first time I'd seen it mentioned. When I txt my brother his response was "that's ridiculous... so I should be playing my striker as a CM?"

[themadsheep2001]

Just now, westy8chimp said:

Not sure that's true. As mentioned... putting players in their preferred role would pretty much mirror exactly how SKY/BT etc pundits show a team setup.

I can't see anywhere in the game that states the formation is defensive shape (it may well be there, but certainly not 'clear') is it there?

My brother and I have been playing every iteration of CM/FM... I only joined this forum this year and was the first time I'd seen it mentioned. When I txt my brother his response was "that's ridiculous... so I should be playing my striker as a CM?"

Si have stated that's the defensive shape. Again the fact that you don't actually see it mentioned is part of the problem.

That formation you see tells you nothing about either their attacking or defensive shape. They will look very different in attacking and defensive transitions.

FM doesn't do well in showing you either.

[Svenc]

7 minutes ago, themadsheep2001 said:

Si have stated that's the defensive shape. Again the fact that you don't actually see it mentioned is part of the problem.

That formation you see tells you nothing about either their attacking or defensive shape. They will look very different in attacking and defensive transitions.

Except if you had everybody on defend/hold position, upon which the guys would just move forward/backwards en bloc (not that anybody's  doing that usually).

[westy8chimp]

9 minutes ago, themadsheep2001 said:

Si have stated that's the defensive shape. Again the fact that you don't actually see it mentioned is part of the problem.

That formation you see tells you nothing about either their attacking or defensive shape. They will look very different in attacking and defensive transitions.

FM doesn't do well in showing you either.

on top of that though... it's one thing to 'not mention it' but then they compound it with;

• players prefered position/role (bar's/icons/stars/tactic advice)
• auto training trains players in their preferred position... not a 'defensive position' i.e. if I leave my training to coach with Arsenal.. he would put Sanchez training as IF, not a wingback or wm
• player promises "will play in preferred position" striker always asks for a striker role, wingers always ask for winger role
• Many AI formations appear same as you see on TV (balanced/top heavy)

These are all issues we have covered... but my point here is ... if SI have decided something so crucial to the game... why have they then given us all the other gameplay visuals and mechanics directly contradicting that paradigm.

It seems not just an oversight or unclear process.. it is completely contradictory.

Logical thought would be - our tactical creator is the defensive shape > portray this through (all of the above)

 

EDIT: Sorry should also add to the rant that all the descriptions of the roles are therefore misleading... and that of the mentalities.

 

[MBarbaric]

i think all the confusion starts with people not grasping that in defensive phase all players become defenders while on attack they are attackers ;D 

to clear the confusion TC should be divided into defensive and offensive phases as minimum. defensive shapes and attacking shapes, team instructions, individual instructions, team philosophies...  instead of current one shape fits all approach.

 

[Guest]

1 hour ago, MBarbaric said:

i think all the confusion starts with people not grasping that in defensive phase all players become defenders while on attack they are attackers ;D 

to clear the confusion TC should be divided into defensive and offensive phases as minimum. defensive shapes and attacking shapes, team instructions, individual instructions, team philosophies...  instead of current one shape fits all approach.

 

As I posted in the thread TFF started I do not think this woudl work with with current level the AI is at but as an alternative have two formation screens that the user flips between which show the expected attacking & defensive shape based on the selections made in the TC

[Kcinnay]

In that matter, the formation screen like it used to be, the hybrid with the arrows, was much more like a real life tactics screen. I used to be a choach. When I put my line-up on the chalkboard, I used forward, diagonal and backward arrows to show which players moved forward in the offensive phase or moved backwards in the defensive phase. Like almost all coaches do. Lots of teams play a 4-2-3-1, defend in a 4-4-2 and attack in a 2-4-1-3 or a 2-4-4. That's how it works.

I've said it before and I'll say it again: I despise the 'formation screen is the defensive formation' dogma. It's illogical and inconsequent, like illustrated by other users before. Following that logic, in some games, I should play 4 players in the D-strata, 4 players in the DM-strata and 2 players in the M-strata. Which looks awful and hasn't got anything to do with football logics or positions or roles.

I've been quite addicted to CM and FM since I was 12 (so for 16 years now), but this is the first time I've actually truly stopped playing the game for a long time. Never been so successful as I was in FM 17, but I truly hope that in a tactical matter, FM 18 will have had a massive tactical (!) overhaul, closer to real life logic, not to FM logic.

[Wells]

3 minutes ago, Kcinnay said:

I've been quite addicted to CM and FM since I was 12 (so for 16 years now), but this is the first time I've actually truly stopped playing the game for a long time. Never been so successful as I was in FM 17, but I truly hope that in a tactical matter, FM 18 will have had a massive tactical (!) overhaul, closer to real life logic, not to FM logic.

Seconded that.. I felt disappointed when I saw FM 17 has not even a single improvement in tactical department.. I think they must do improvements with the next version, they can't really let it go.. we'll see it though... I will just hold on before buying FM 18.

[MBarbaric]

3 hours ago, Barside said:

As I posted in the thread TFF started I do not think this woudl work with with current level the AI is at but as an alternative have two formation screens that the user flips between which show the expected attacking & defensive shape based on the selections made in the TC

well, there are engineers and programmers who should think about how to do that over certain time span. nobody thinks it could be done in a year or two. but there should be some road map that slowly moves FM logic towards football logic. some basic plan which things need to move first and what is to follow.

separate defense and offense in TC

check coaching manuals and see how defense works in football

-implement basic positioning

-introduce zones

-implement pressing

check coaching manuals and see how offense works in football

...

ahh, and get rid of team shape. that absolutely makes no sense in football terms and confuses the hell out of people. team shape is formation in any football context and nobody knows what it actually means in FM.

[noikeee]

16 hours ago, Kcinnay said:

I've said it before and I'll say it again: I despise the 'formation screen is the defensive formation' dogma. It's illogical and inconsequent, like illustrated by other users before. Following that logic, in some games, I should play 4 players in the D-strata, 4 players in the DM-strata and 2 players in the M-strata. Which looks awful and hasn't got anything to do with football logics or positions or roles.

I agree. It's the whole root of the issue. I understand it was a way of settling things for the introduction of roles, and an easy concept to work as the basis of coding them, but the end result has become counter-intuitive for the user, and inconsistent with so many things in the game.

What I want, is for the formation screen to be the AVERAGE formation. Then show us up extra panels for the defensive and the attacking formation.

[Preveza]

All these points are constructive and is very pleasing to see. I feel it will take SI years to develop all this and more. They have just recently transitioned away from sliders (FM13) and onto tactical roles and responsibilities. For these changes to be made, it will take time, and lot's of it. Game needs huge improvements in other areas aswell. Far too many areas of the game have been untouched for a long long time. (Years)

[sirgiorgio]

While I was reading this, I thought that maybe a simple redistribution of roles would be a huge step forward. For instance:

1. Remove duplicate roles for different positions: I want a WB (role), I use a WB (position). I want my wide midfielder to stay close to the CM, I use M R/L (position) and WM (role). I want my wide midfielder to be a classic winger, I move him to the AM L/R spot, where the winger position would be available. The same logic can be applied for duplicate roles on the CM/DM strata.

2. Review certian roles and combinations: I want my second striker to come deep to get the ball and help the team in transition... well then maybe I don't want a striker? The SS v F9 example above is fitting: if the false 9 is a "fake striker", then it shouldn't be a striker (position), but an attacking midfielder "acting" like a striker, which is more or less what a SS does in a striker-less formation. Proof is, that role was frequently occupied by Fabregas, who isn't a striker (and not even a "natural" AM, but certainly more "adaptable" as AM than he is as ST).

Choices would be made (is a DLP a DM or a CM?), new roles would be introduced (e.g. "attacking fullback" or something) and existing roles would probably acquire new meaning and/or effectiveness (e.g. DW). 

Unfortunately, It would also be a pain in the neck for researchers and whoever creates "default formations" (the ones you find in the DB) because, following this logic, very few managers would use perfectly symmetric formations... 

[Keith Ang]

Hey Wells, im just too lazy to read the 4 pages long comment. so if im repeating this im sorry. 

ok here is what i think. Now i dont think that its bad AI. the Winger stays out wider usually to check runs from opposition winger or marauding full backs that usually uses the byline. in this case it looks really space out because u r playing against a narrow formation team. hence u see the disparity. the other reason could be your team shape which is Structured then they tend to stay a little wide as thats was their initial function as a Winger and thus not breaking the formation. 

442 traditionally is a balanced formation but as time flies new football philosophy/formations appears which were meant to outdo the older ones. the 4132 has the extra players in the middle making short and through passes a potent threat against your 442. hence thats why u will be left open at times. So basically 442 inherent weakness is the wings. but ofc it doesnt mean u cant tweak the formation to mitigate that weakness. Arrigo Sacchi, Diego Simeone to name a few has shown us the way.   

example, im playing newcastle using a 4411, ayoze just intercepted the ball but we were on the backpedal. see how far wide the wingers are? im playing a standard mentality with the most narrow option i can select on the width selection. btw  the mentality does changes the width depending on how aggressive you are in attacking. 

So in short i dont think its bad AI. i think its meant to be that way and even in RL. u can tweak and make changes to overcome the weakness your formation have. or in any cases switch the players role. i hope my 2 cents helped.

[Svenc]

Beating the engine, defensive AI and a few general curiosities

This was touched upon before, but FM 2017 due to that overhaul has also introduced a really strange match dynamic as per defensive AI. Defensive AI = oft both wide backs on defend duty. It's the kind of thing that entirelly stats/ratings focused players would previously rage about. Like in long-term saves where your side got the BIG TEAM TO BEAT in the game world and every AI went defensive, which was prone to trigger the reaction: If I concede at all, it only happens from either a set-piece, long shot, bug or fluke, bad ME/cheating AI. Being primarily stats focused meant nobody would pick up this might be because half the opposition teams just started sitting behind the ball outside of precisely those set pieces, and most of the attacks thus being easily defended. For better and worse, on FM 2017 this can be different on the occasion.

If defensive AI goes this defensive on here, AI v s AI matches as well as your own, it channels a bulk of the play through the middle where this season's weaker areas are. And naturally, 4-4-2s are the most vulnerable, as they don't defend as they once used to.

Evidence A: Ball just transitioning into the final third, the two chains actually keeping it reasonably compact for the defending team, the attacking teams backs stay back, so play tends to go through the middle naturally a wee bit with limited wide outfield balls in the final third too (this is all AI).



Evidence B: The right-hand players breaking it apart, opening big big big channels

Evidence C: Guy into the channel being passed to. Rather than any midfielder having closed those (like those guys sitting miles wide), it is the centre back who is forced to step in.



Evidence D: Guy plays a neato through ball, goal. Naturally that defending itself is nothing "new" as for this thread, so please don't flame me. I still have a suspicious this would have likely ended in a few more of those fluke shots on prior iterations, overall having a far lesser chance of resulting into a goal (as a logical result of AI being focused on keeping it tight, not as a sign of an inherently broken ME).




Naturally, this defending of central areas is where this season's big exploit tactics lie. Probably he most effective ones in years.  Weakest area on the pitch virtually or otherwise because: a) shortest distance to the virtual goal and b) once successfully overrun, no more d-line. If you look around, you'll quickly find a few reviews about how you just could plug stuff in, and roll over. Naturally, some enjoy this, others point it out as a sign of a broken ME also in various reviews, such is life/preference. Personally I just came in 2nd in the BL with Freiburg not doing anything but hitting continue, had the matches on "commentary", not made a sub ever. Technically, this is still "playing" I guess. How do the good guys do this? Keep it central, stupid.

In particular with the wide mids, it is impossible to mark everybody in those narrow strikerless narrow wing back tactics. Only a few AI formations could (and do) somewhat plug those holes to counter. Namely: A third back getting involved that provides a spare man to cover any of the many times one of his colleagues is forced to step in and close down. This plus a few additionally central mids/DMs that somewhat balance this out. The reason I bring this up is, it would perhaps be of use if SI contacted some of the better engine exploit tacticians, in particular those who don't merely trial&error&copy/paste. Not primarily to stop people "beating the ME", though that may be a concern for SI too (those tactics only ever work short-term, undermine the multiplayer, plus prevent people from learning how to play). But because any such can highlight issues that may be otherwise flagged "minor" in the testing or may not get picked up at all, undermining a few of the hard work and progress made.

[crouchaldinho]

On 10/05/2017 at 17:07, Wells said:

Normally in 4-4-2 it should look like this

-----------W-------CM-------CM------W-----------

but it looks like this

--W----------------CM---CM-----------------W--

 

 

 

 

Apologies for bumping an old thread but I was wondering if this issue was fixed in FM18?

The issue in FM17 was that wingers/wide midfielders stayed far too wide defensively and you could not keep a proper defensive shape no matter what instructions you gave tactically. The MR/ML never moved infield to help centrally. If, for example, you were playing a 4-4-2 formation, the central midfielders would stand around holding hands in the middle while the wide players stayed out as wide as possible doing nothing. The wide players refused to help centrally even when the opposition overloaded the central areas. 

I'm thinking of picking up FM18 and it would be good to know if it was addressed and fixed for the current game.

[crouchaldinho]

On 10/05/2017 at 17:07, Wells said:

Normally in 4-4-2 it should look like this

-----------W-------CM-------CM------W-----------

but it looks like this

--W----------------CM---CM-----------------W--

 

 

 

 

Apologies for bumping an old thread but I was wondering if this issue was fixed in FM18?

The issue in FM17 was that wingers/wide midfielders stayed far too wide defensively and you could not keep a proper defensive shape no matter what instructions you gave tactically. The MR/ML never moved infield to help centrally. If, for example, you were playing a 4-4-2 formation, the central midfielders would stand around holding hands in the middle while the wide players stayed out as wide as possible doing nothing. The wide players refused to help centrally even when the opposition overloaded the central areas. 

I'm thinking of picking up FM18 and it would be good to know if it was addressed and fixed for the current game.

[HUNT3R]

10 minutes ago, crouchaldinho said:

 

Apologies for bumping an old thread but I was wondering if this issue was fixed in FM18?

The issue in FM17 was that wingers/wide midfielders stayed far too wide defensively and you could not keep a proper defensive shape no matter what instructions you gave tactically. The MR/ML never moved infield to help centrally. If, for example, you were playing a 4-4-2 formation, the central midfielders would stand around holding hands in the middle while the wide players stayed out as wide as possible doing nothing. The wide players refused to help centrally even when the opposition overloaded the central areas. 

I'm thinking of picking up FM18 and it would be good to know if it was addressed and fixed for the current game.

Wide midfielders will defend narrower than wingers.

[crouchaldinho]

4 hours ago, HUNT3R said:

Wide midfielders will defend narrower than wingers.

So are you saying that they are no longer hard-coded to behave like they were in FM17?

In FM17, regardless of role, they would move into positions as in the screenshot above. Has this been resolved?

 

[HUNT3R]

7 hours ago, crouchaldinho said:

So are you saying that they are no longer hard-coded to behave like they were in FM17?

In FM17, regardless of role, they would move into positions as in the screenshot above. Has this been resolved?

 

 

 

 

As you can see here, it's not nearly as bad as it was in FM17. There are times when you'll see them staying a bit wide when it doesn't make sense to, but it's not too much of an issue.

 

The same highlight, just further down the pitch:

 

[crouchaldinho]

Thanks for that. Judging by those screenshots, it certainly looks like an improvement over FM17. 

I might give it a go then. Craving some FM right now.

[HUNT3R]

7 minutes ago, crouchaldinho said:

Thanks for that. Judging by those screenshots, it certainly looks like an improvement over FM17. 

I might give it a go then. Craving some FM right now.

If you're unsure, you can always try the demo or watch a lets play video or two.

I've seen a few pre-world cup special on FM, so don't take too long to decide.