leftfoot

Just as a disclaimer this is not a discussion on free roles in the real world rather how the free role is assigned by AI managers in the game.

Inspired by Abramovic's AI thread I decided to try and figure out Rijkaard's tactics as they were proving very successful for him as manager for Arsenal in my game in 2011. In his home attacking formation 4231 I noticed that he had the left attacking midfielder on free role but not the right which I found odd as I would have thought that this would create an imbalance.

I decided to investigate this using FM Modifier (I only use it to investigate the game I swear) and it turns out that he actually assigned the free roles based on the hidden free role attribute.

I saved the game, modified the free role attribute to be 20 for the AM and strikers, and upon viewing the game found that Rijkaard gave all of the attacking midfielders and the striker a free role. I reloaded the game, modified the free role attribute to 1 for all the players and in the same fixture against the same opposition using the same formation Rijkaard didn’t assign any player a free role.

From changing the values what I discovered was that there appeared to be a threshold value of 10 whereby for Rijkaard his preferred tactical setting is to give any AM or striker within his 4231 formation a free role if their free role attribute is greater than or equal to 10. The opposition in this case was inferior. Against stronger opposition he used a 41221 and assigned free roles to the AMs and striker using the same greater than 10 rule.

I tried the same test for Ferguson at Man Utd and found the same threshold value of 10 applied in this case. In his 442 against weaker opposition (Greek team away) if the players in the winger position and striker roles had a free role attribute greater than 10 then the free role box was ticked. In his 4411 the free role was ticked for wingers, AM and lone striker. Against equivalent opposition (Real Madrid) at home the same principles applied. In defending a lead away against Barcelona the free role for strikers remained but was removed for the wingers in the 442 no farrows formation.

For both managers, regardless of the free role attribute of the central midfielders the free role box was never ticked. This applied regardless of whether the CM was an attacking player, i.e. forward runs set to often, or a holding player forward runs rarely.
Also Rijkaard has a hidden free role attribute of 16 and Ferguson’s is 12. Pako Ayestaran, assistant manager of Liverpool, has a free role attribute of 4 so I ran the test with him and he followed exactly the same pattern of >10 free role attribute for a player then give him a free role when playing as attacking midfielder/striker/winger in an attacking mindset, and remove free roles when defending a lead.

What the tactical implications are of this are best left up to those who can judge what tactics are in place from watching the match engine, because as my screen name suggests I suck at this. I have an idea about what the free role means from understanding football but within the game itself I find it difficult to see the impact of different instructions on player behaviour despite running numerous test games which essentially left me more confused. The only thing I noticed that seemed consistent with the free role was that those with a high free role attribute seemed to be more adept at finding space and tended to drift away from the opposing player covering them ( although I accept that this can also depend on off the ball, decisions, anticipation but to test it to this extent would take a lifetime).

It does throw up a question mark over the free role attribute and its hidden nature as there seems to be little link between visible attributes you would expect a good free role player to have (flair, creativity, anticipation, decisions etc.) and the value that they have for the free role attribute. In particular if assigning an attack minded player with a free role, who has a free role attribute greater than 10, results in a better all round performance than surely the AI has a distinct advantage in this scenario as it obviously has access to hidden values we can’t see without using ‘cheat’ tools like FM modifier.

Just as an example within a 4222 I had two wingers set to free roles, yet despite their similar attributes one always performed better than the other. After discovering this free role assignment AI behaviour I checked the values my two wingers had for free role attribute. The one who performed better had a higher free role attribute, although I would be cautious in attributing his better performance solely to this.

For those who don’t like to cheat or use FM Modifier or Genie Scout then maybe you can post players you use in free roles who don’t seem to perform as well as others. This way I can check it out myself to see if there is a link between the free role attribute and general performance in a free role, particularly in strikers/AM/winger roles (although there are a multitude of other variables at play it might give a general indication).

P.S. If somebody already discovered this and their post is buried in the forum, apologies for the re posting of known information.