Team Statistics in real life vs Team Statistic in FM

[themadsheep2001]

2 minutes ago, Cap'nRad said:

It does use a way of playing that is, in this modern day, considered somewhat ancient. A few components of which are:

1) The ball rarely goes back to the defence in the buildup, especially on higher mentalities. This is due to lack of passing lane pressure and strikers defending.

2) Still an over reliance on crosses from wide positions.

3) Lack of curved runs means offsides are numerous and movement is only in a more static, forward only direction.

4) More defensive mentalities can hold onto the ball regardless of stats, also due to lack of passing lane pressure/ any kind of coordinated press.

5) Lack of passing lanes means that 1st and even 2nd block can be breached with ease.

6) General directness of play and abundance of random hoofing.

etc. 

As for player choices, just look at that 5v2 counter video posted a few weeks ago. Shows why bigger sides don't dominate as they should, the buildup and ability to regain possession is so poor at times. It's clearly a m.e issue.

That's not a way of playing. Those are areas to be tighted up or improveds (outside of pressing which need developing.

None of which stop the ability to create (or re-create) top tier approaches. Hence why many human users can do it relative easily. The AI isn't as good at making said choices. So when you say its clearly an ME issue, it's in fact, less so.

[Svenc]

42 minutes ago, Cap'nRad said:

As for player choices, just look at that 5v2 counter video posted a few weeks ago. Shows why bigger sides don't dominate as they should, the buildup, decision making, and ability to regain possession is so poor at times. It's clearly a m.e issue.

 

Almost forgot about that one -- even though I did similar in the past, the biggest scoreline I ever got was a 1-31ish or something, that was back on an iteration where such shipped goals anyway. I don't know how viable that is, as at such constant breaks the game basically triggers "counter attacks", which would likely influence a player's decision making. In such scenarios, you basically have an artificial scenario you never have out of a running game, a counter attacking in oodles / endless of time, as the defending team's players are artificially triggered to stay forward (by making them man marking the oppositions CB/s defenders). It's basically training matches of 2 vs x. So one for SI to gauge how useful it all is for current iterations.

I still was some curious and just did similar in a pre-season friendly, and to say there was all around decent decisions made would be exaggerating it (loads of poor ones made). If SI deem this useful as a "stress test" of final third decision making, all they would need to do was taking a side, fielding all the players bar two centre backs in attacking positions, and make the advanced players all mark the opposition centre backs, ideally. I'd be cautious with concluding from this that the general decision making wasn't much cop though, due to it being such an artificial scenario the players may be coded to have "no awareness of", as with any normal tactics, such artificial counter attacks would never transpire. As such, their decision making may always "assume" that players may be back in position soon, rather than endlessly staying forward, so they may be coded for some added urgency to finish the move off. May still reveal weaknesses in how counters are played out (counter attacks are traditionally imo one of the weaknessess).

[Cap'nRad]

54 minutes ago, themadsheep2001 said:

That's not a way of playing. Those are areas to be tighted up or improveds (outside of pressing which need developing.

It is a distinct style of play that is brought about due to those areas lacking. There is no need to recycle the ball among your defence when the 1st block is nonexistent, strikers don't defend properly and thus the midfield can be breached easily; which results in the more direct style of play the m.e uses. It is much harder to thread through balls when the runs aren't curved and thus it's very hard to avoid offsides, which leads to crossing preference. It is much easier for defensive mentalities with bodies in their own half to keep the ball due to lack of coordinated pressing. You see where I'm getting at. The style of play is a direct result of the lacking areas.

 

54 minutes ago, themadsheep2001 said:

None of which stop the ability to create (or re-create) top tier approaches. Hence why many human users can do it relative easily. The AI isn't as good at making said choices. So when you say its clearly an ME issue, it's in fact, less so.

I think some of the choices for recreation of these approaches are illogical in the context of the real world (and the in-game descriptions for that matter). But regarless of recreations, these distinct m.e traits will still exist. 

[Cap'nRad]

6 minutes ago, Svenc said:

Almost forgot about that one -- even though I did similar in the past, the biggest scoreline I ever got was a 1-31ish or something. I don't know how viable that is, as at such constant breaks the game basically triggers "counter attacks", which would likely influence a player's decision making. In such scenarios, you basically have an artificial scenario you never have out of a running game, a counter attacking in oodles / endless of time, as the defending team's players are artificially triggered to stay forward (by making them man marking the oppositions CB/s defenders). It's basically training matches of 2 vs x. So one for SI to gauge how useful it all is.

I still was some curious and just did similar in a pre-season friendly, and to say there was all around decent decisions made would be exaggerating it. If SI deem this useful as a "stress test" of final third decision making, all they would need to do was taking a side, fielding all the players bar two centre backs in attacking positions, and make the advanced players all mark the opposition centre backs, ideally. I'd be cautious with concluding from this that the general decision making wasn't much cop though, due to it being such an artificial scenario the players may be coded to have "no awareness of", as with any normal tactics, such artificial counter attacks would never transpire. As such, their decision making may always "assume" that players may be back in position soon, rather than endlessly staying forward, so they may be coded for some added urgency to finish the move off. May still reveal weaknesses in how counters are played out.

Aye I can understand this, although this would mean that during counters pre-coded behaviour takes preference over stats so attributes like anticipation, composure and decisions don't matter.  I don't think such situations should be considered artificial by the coders, as they do happen in real life; just not as often in the top tier leagues (obviously 6v2 is quite extreme, however 3v2, 4v3, and 4v2 are more common). I can understand how it's hard to code counters as there are different ways to do it in real life and there's always some awareness and nuance to the directness.

Maybe the m.e could be coded with counter attacking as the base, then as the space gets smaller and the bodies more, more possession based decisions are made, eventually becoming full blown possession football when playing a parked bus. At least that's how it's taught in the real world. Either way, I don't think the attacking and counter attacking play are reflective of real life, especially with bigger sides who have the ability to pull off lethal counter attacking moves and control possession.

[MBarbaric]

1 hour ago, themadsheep2001 said:

None of which stop the ability to create (or re-create) top tier approaches.

there is no way to recreate any kind of possession game that resembles real football.

Possession based game plan is born as a means to overcome the solid defensive block. In order to work, it needs movement. Coordinated movement between players of the team in possession. This movement of players and the ball unbalances the defensive unit allowing the team in possession to find holes in the otherwise solid block.

Play from the back is commonly accepted as a standard when creating a possession based tactic. When you chose that instruction in the ME it will instruct your players to play shorter and slower. However, that isn't how "play from the back" works. Short passing is a consequence, not the goal. The idea is to shift the defensive block and find holes in it. This is not primarily done with short passing but with movement and positioning of players on the pitch. Even the most basic movement in football, player rotation, doesn't exist in the ME. If there is no movement, there can't be a sound possession game. 

Now, someone might ask, how/why are we then able to create tactics that have 65% of possession? After all, if there was no movement, there couldn't be possession, right?

The reason is this: You don't need any kind of movement to retain possession in the ME because there is no solid defensive block that you need to by pass. Possession based tactics are born to overcome that defensive block, so if there is no block, there is no need for extensive, coordinated, intelligent movement. You can by pass the second line of defence with simple short passes as the defending team doesn't act as a unit. As it should.

here's an example of defensive organization in real football. It isn't particularly successful but you can clearly see there is an idea behind it and what that idea is. In FM, there's nothing even close. 

 

[Totalfootballfan]

I decided to check how things stand in Seria A so I picked Benevento and quickly finished the 1st season.

I uploaded the game save at FTP - Seria A Test.fm

 

Here’s the league table at the end of the season

 

 

 

 

Juventus got only 62 points  and finished the 6th place, scored 59 goals and conceeded 39 and didn’t even manage to get Champions League spot for the next season.

Napoli got only 59 points and finished the 8th place didn’t even managed to get into Euro Cups the next season, also it scored only 62 goals in the season when in real life it usually scores 80-94 goals in a season.

I don’t know about you guys but to me all that looks extremely unrealistic and disappointing because at least during the 1st season in FM I expect to see things happens somehow close to real life.

 

 

 

 

At the same time I was checking Spanish La Liga, I picked Girona there and was sacked just before few weeks the end of the season.

Here’s the league table at the end of the season:

 

 

Barcelona got only 76 points and finished the 3rd place, also, Barcelona scored only 55 goals  in the season when in real life it scores 110-120 goals in a season.

Real Madrid got 87 only points and such small amount of points was enough to win La Liga , also, Real Madrid scored only 64 goals in the season when in real life Real Madrid usually scores 110-120 goals in a season.

Atletico Madrid scored more goals than Barcelona and Real Madrid in the season.

[russell9]

After managing teams like Barca, Real, Juve, PSG by myself, I can conclude that this problem is due to bad AI manager employing ineffective tactic.  All the top teams I managed easily scores 120+ goals  and about 60% avg possession in a season.  The AI manager for top team is simply not up to par.

[Guest]

3 hours ago, russell9 said:

After managing teams like Barca, Real, Juve, PSG by myself, I can conclude that this problem is due to bad AI manager employing ineffective tactic.  All the top teams I managed easily scores 120+ goals  and about 60% avg possession in a season.  The AI manager for top team is simply not up to par.

 

Has the AI deteriorated since FM2017? Sounds unlikely, but those stats above made me think about it.

[Svenc]

7 hours ago, Totalfootballfan said:

 

 

Barcelona got only 76 points and finished the 3rd place, also, Barcelona scored only 55 goals  in the season when in real life it scores 110-120 goals in a season.

I'd be interested in the SOT rankings, as goals (randomly) fluctuate in a very natural way: Even if we would assume that the forward was completely unmarked/unpressured on each of those (even on FM's level we can rule that out I reckon, no less because it's been traditionally oft easy to "spam" the more difficult shots on target, e.g. headers from set pieces), there is a fat chance for missing each of those (which is why the long-term shot conversion is pretty constant at around every 3rd-4th).



The same goes for teams actually scoring those 110. In the above run, Eibar was the only team that scored 30 goals/17 matches. Ronaldo scored 12 from 15 (injured now), at 3 attempts per match (real life 7). Real Madrid unterperformed now too in the purely point rankings, sitting 9th place with a WDW record of 7-5-3. In real football, if such teams drop a significant amount of points, it is usually in parts influenced by (bad) finishing streaks (just look at the match stats of matches which Real haven't won this season, or Juventus early 2015/2016 season with a bad finishing streak that lasted 10 matches) -- and if you want to gauge this at its most rotten, look at Dortmund 14/15 season, unbelievable. Now on FM for individually matches that may still be the case. But also, as argued, no hugely dominance transpires anymore. Naturally, this is all idenitifiyng symptoms at best.

[andu1]

55 goals scored by Barcelona? that is terrible.

[Mitja]

11 minutes ago, andu1 said:

55 goals scored by Barcelona? that is terrible.

no that must be mistake, a typo. probably Messi scored 55 alone?

[Mitja]

17 hours ago, MBarbaric said:

Now, someone might ask, how/why are we then able to create tactics that have 65% of possession? After all, if there was no movement, there couldn't be possession, right?

The reason is this: You don't need any kind of movement to retain possession in the ME because there is no solid defensive block that you need to by pass. Possession based tactics are born to overcome that defensive block, so if there is no block, there is no need for extensive, coordinated, intelligent movement. You can by pass the second line of defence with simple short passes as the defending team doesn't act as a unit. As it should.

here's an example of defensive organization in real football. It isn't particularly successful but you can clearly see there is an idea behind it and what that idea is. In FM, there's nothing even close. 

 

a bit harsh imo, of course there is a defensive block, the biggest problem imo is that it doesn't shift enough (especially midfield, which should be seperate unit much more mobile than currently), doesn't congests space, forwards offer almost nothing except pressing the centrebacks and most importantly players should mark much closer especially in own third. I must say pressing was improved a lot from 16 to 17, you could spot 3 even 4 players pressing one opponent but of course wide players didn't offer any support to centre midfield on 17. 

it is beyond me how top AI teams can't dominate posession, but most probably it is also related to the long term issue when attacking mentality meets very catious one and every team on earth being able to play short passing game (which is topic on it's own and again linked to poor defending). I find frustrate opposition tactics completly (contain, defensive even counter) unrealistic, irl no manager would try to aproach a game in such suicidal manner as general match plan. but on fm it works.

[Mitja]

20 hours ago, Cap'nRad said:

Unfortunately it's not this simple. The direct style of play the m.e uses does not fit the current style of modern big teams who prefer to control the game, not only attack-wise but also defensively with more complicated pressing methods. As well as the ability of mediocre sides to hold onto possession regardless of technical and mental stats, the whole m.e configuration is not really geared towards these big teams massively over performing without the use of some interesting tactical choices (from a real life point of view).

not only that, it is quite possible to eliminate opposition posession when they employ contain or defensive tactics but you need to maximise the defensive line, pressing, employ tight marking and use hard tackling. but it looks AI is not awere of that.   

[andu1]

1 hour ago, Mitja said:

no that must be mistake, a typo. probably Messi scored 55 alone?

only if it was true...  

[tajj7]

Personally I have always thought pace and acceleration is over rated by the match engine when it comes to forward players and how they get into goalscoring spots and this causes some players in game to perform better, especially when it comes to scoring goals than others. Examples are the likes of Benik Afobe, who has a return of 10 goals in 47 games for Bournemouth in real life, to in game I have seen him in several saves score over 20 league goals for Bournemouth. Walcott seems to get back into the Arsenal team very easily and scores goals, I see countless examples like this of players who have high pace and acceleration attributes but average in most other areas performing like forwards. 

Now whether it is something to do with a direct style working better for the AI or the AI not managing how to best use strikers who don't rely on pace I don't know.  This might explain the AI managing Messi, Ronaldo etc. maybe the AI just isn't able to get the best from teams that are set up to be more possession bases or controlling but can get teams that are more direct and counter attacking to perform better. 

There is a guy in the careers forum who has got to top 7 in the PL with Cambridge using a very direct system, basically hoofball and whilst fair play to him, I am not sure teams playing this way in real life would survive even in the championship. 

[Guest El Payaso]

12 hours ago, Totalfootballfan said:

Barcelona got only 76 points and finished the 3rd place, also, Barcelona scored only 55 goals  in the season when in real life it scores 110-120 goals in a season.

Real Madrid got 87 only points and it was enough to win La Liga , also, Real Madrid scored only 64 goals in the season when in real life Real Madrid usually scores 110-120 goals in a season.

Atletico Madrid scored more goals than Barcelona and Real Madrid in the season.

Las Palmas conceding 31 falls into the same category as they should be conceding about double that amount if not more.  <- Would add a smiley here but it's simply sad, not a good simulation at all. 

In my save in terms of Barcelona the situation seems encouraging in the beginning as Barcelona have 10 goals scored in 3 fixtures but Real Madrid are being woeful. I watched some highlights of the Madrid derby and Real were really uninspiring and totally lacked movement up front and it was just an endless Modric losing the ball as he had no options and totally unable to shield the ball. There should be better one-two playing from players: when you pass the ball you move along higher up the pitch to allow a passing option.

Personally I have always thought pace and acceleration is over rated by the match engine when it comes to forward players and how they get into goalscoring spots and this causes some players in game to perform better, especially when it comes to scoring goals than others.

This for me has always been the case on FM as the ME is most of the time unable to create anything that resembles real football. The intelligent side of the game still isn't there, especially when you look how defenders play in the game: put a little bit of some called closing down on them and they will clear the ball while in real life even defenders most of the times look for the best possible option and especially with instructions like 'play out defense' should always be looking to pass the ball short and take risks doing that.

Atlético by the way against Real Madrid were quite uninspiring against Real as they only created something from set pieces and those random long balls where defender wins the ball back either by tackling or intercepting and immediately hoofs a long ball (simply hate this behavior) and there was Gameiro randomly through on goal two times I think. Has nothing to do with real football or tactical choices. Simply just random stuff like those repetitive throw-in chances also are. 

[Cap'nRad]

4 hours ago, Mitja said:

a bit harsh imo, of course there is a defensive block, the biggest problem imo is that it doesn't shift enough (especially midfield, which should be seperate unit much more mobile than currently), doesn't congests space, forwards offer almost nothing except pressing the centrebacks and most importantly players should mark much closer especially in own third. I must say pressing was improved a lot from 16 to 17, you could spot 3 even 4 players pressing one opponent but of course wide players didn't offer any support to centre midfield on 17. 

Well, I guess you could say there is something trying to resemble a defensive block, however the whole defence is off. And it all starts with the strikers. Lets assume the defensive formation is a 442.

This is how defence would look on fm: Due to our strikers' poor defensive positioning, the opposition cb's are able to breach the 1st line of defence far too easily and find their deeper midfielders. And since in fm players close down the ball carrier automatically, this forces one of our midfielders to step out and close him down. This leaves just 1 central midfielder trying to hold a good position. And since our wide midfielders don't tuck inside to the extent that they should, parts of our halfspaces and center are left exposed. Which means a simple pass to their deep striker is possible, ans since the cb's don't mark and follow strikers who come deeper to get the ball, he is allowed time to turn and dribble into dangerous areas while the defenders back away.

In real life, with a 442 our strikers would be engaged behind the opposition's defence, usually with one marking the deepest midfielder and the other closing down using passing lanes with the clear intent of guiding the ball to the flanks. Once on the flanks the whole defence shifts to that flank, with the opposite wide midfielder coming centrally to maintain the defensive block. With the aim being to win the ball or at least stop progression on the flanks, and keep the ball wide, away from the midfield. In an ideal defensive situation, the ball would be shifted from fank to flank with no progression past the first block for the full 90 minutes. 

So as you can see, starting with the strikers, the problems with fm's defence are quite numerous (lack of striker defending, always closing down ball carrier, lack of tight marking in certain areas, etc) so I can understand when @MBarbaric says there's no defensive block. Like I said there might be something trying to resemble it, however it doesn't come close to succeeding. How can it when the defence is exposed from the first pass?

And this ties into the topic because the possession numbers would be much greater for teams who try to break down a good defence and succeed, mainly because their choice of passing would be different, and player quality would shine through more (Ball-playing defenders, in the real world sense, and deep lying playmakers would be more important due to their ability to handle/escape pressure from the first block and play the right pass).

 

[Guest]

7 minutes ago, El Payaso said:

it was just an endless Modric losing the ball as he had no options and totally unable to shield the ball. There should be better one-two playing from players: when you pass the ball you move along higher up the pitch to allow a passing option.

 

Modric is small, and he is not particularly strong. This is why he constantly loses the ball. In my experience, this is pretty much all there is to it.

 

This has been a problem with FM for years: size and strength, both individually and together, beat skill. In real life, it's quite hard to get the ball from Coutinho, and almost impossible to steal it from Iniesta. But in FM, both players lose the ball all the time, because they are small and not very strong.

 

This, incidentally, is precisely why I have never had problems beating Barcelona either home or away: they are a lot smaller and weaker than my Liverpool, and I have a tactic which quite easily crushes them, 4-0 at home and 1-5 away. That's the way it is. It's not good, but that's how it is.

[Guest El Payaso]

3 minutes ago, xzar_monty said:

Modric is small, and he is not particularly strong. This is why he constantly loses the ball. In my experience, this is pretty much all there is to it.

Yes and this is a good example on the fact that the ME cannot produce intelligence as what makes Modric a great player in real life is most definitely his intelligence and by footballing intelligence I don't just mean creative passing but also that he knows how to retain possession even though he is not strong. 

When I look at my team and especially Jonathan Viera and Sergi Samper I notice that with them I do not see intelligence as so many of their pass attempts are totally mindless and without any idea. Not expecting them to pinpoint every pass or to put in 10 through balls a game but there always should be an idea when they pass the ball and they should stand out more in terms of intelligence. 

[Guest]

2 minutes ago, El Payaso said:

Yes and this is a good example on the fact that the ME cannot produce intelligence as what makes Modric a great player in real life is most definitely his intelligence and by footballing intelligence I don't just mean creative passing but also that he knows how to retain possession even though he is not strong.

 

It's true. However, this is where things become seriously problematic for the programmer. The intelligence you describe could also be called human creativity, i.e. the ability to improvise within a given set of circumstances. Anyone who can genuinely do that with AI basically has a Nobel prize waiting for them.

 

Someone could argue that it can't be so difficult, because modern computerized chess and go! games easily can be made unbeatable for humans. The difference is that each individual situation in a chess or a go! game is static. Football is never static, even in set pieces.

 

I am not qualified to say whether it's realistic to expect drastically better AI from this game, but I agree that there are glaring problems with the current one. Size beats skill, there is no real intelligence, players don't understand each other's size in relation to each other, nobody ever learns from mistakes, and so on.

[Guest El Payaso]

6 minutes ago, xzar_monty said:

 

Yeah there is something seriously wrong in general with how teams play and create chances if Real Madrid can create four shots with 58 ball possession. I don't see a problem with Atlético winning 2-0 against them at home as this is a possible yet currently unlike scenario but if it happens that way that I described then it's not okay anymore. Both teams looking really uninspiring and all the chances from random hoof balls and set pieces. Same with Barcelona 5-0 against Athletic Bilbao: 3 goals from crosses scored by Dembele and Messi (smallest men inside the area with a team that doesn't rely on crossing) and two from messy after initial set piece marking and Barcelona recording 52% ball possession and 12 shots in total. 

At the same time in Premier league players like Iheanacho and Afobe score goal a game as the ME favors that type of players. Players without real football intelligence but lot of pace in their hand. While in real life both are struggling to even get to the team. You could argue that hey Defoe has been banging them in but Defoe is quite intelligent player also especially with his movement and this has been improved him throughout experience and he is also a really solid finisher in his older days. 

[Mitja]

can't comment on 18 but I must say last couple of MEs improved player inteligence vs speed or strength a lot and I like to watch full games on 2D. I don't know but it seems to me when my team meets better one, I'd say it's exactly the intilgence and skill that seperates my team to the other.  try to think about MEs from a deacde ago when speed was truely overrated.  I quite like 17 ME, it would be very solid one if the wingers and strikers could defend, there are too many offsides, crossing is a little weird, passing too good, first touch is insane but I enjoy my save.

but the reason I like it so much is because I edited some manager attributes in the league. I changed all 442 and 352 formations to 4411 and 4231. I improved all smaller team managers in tactical knowledge, gave them more attacking mentality and tempo and some other minor stuff and bum...I have the league and ME to my taste 

you won't believe maybe but there are actually many less shots in all league matches and football looks very good, too good. I even think the overall scores are a little lower but I didn't do any testing. 

 

 

 

[Guest El Payaso]

1 hour ago, Mitja said:

I quite like 17 ME, it would be very solid one if the wingers and strikers could defend, there are too many offsides, crossing is a little weird, passing too good, first touch is insane but I enjoy my save.

Would be very solid... And then lists about ten things wrong with it.  Well in general I would like to say that intelligence might have been better in FM 2017's ME. Some of the stuff I see in the 18 ME is really absurd. Like:

- How often defenders especially make the worst possible choice like: okay I just won the ball back, have three or four easy options available to actually profit my team and my manager saying to me: play out of defense. What do I do? Hoof the ball straight to opposition with first touch. Or: I'm under a bit of pressure and have to clear the ball. I could clear it out for a throw in or just along the touchline. What I do? Put it out for a corner or straight into the middle where the opposition midfielder is receiving the ball. Or in closing down: the opposition winger has just passed the ball to the midfield and there is already one of my midfielders going to close him down. What do I do? I will go there too and leave the winger totally open and let him put a free cross in. This list is endless...

- Or long balls. The ideal scenario would be that basically always both attacking and defending team have a player going to challenge for the first ball but so many times on FM there is only one player moving to the ball which sees either a striker or a defender collecting the ball freely. When I used to play football one of the first things I was taught especially as I played defender: you have to challenge for the first ball and even if you don't win it you will make it harder for the opponent to control it and set other people up. This applies both to high and low balls played.

- Throw-ins. These for me would need thousand different scenarios but I see the same defensive flaws every year. Throw-ins for me actually are one of the best scoring chance outlets. I see my DM getting a scoring chance or two or three every game from throw-ins. Teams never seem to learn to defend throw-ins and same applies to all set pieces as there are highly repetitive scenarios in all of them. "Oh it's that goal again".

- Offside awareness. This most definitely has lot to do with intelligence. The numbers are high and it shows bad intelligence both from the player passing the ball and the one who is going to receive it. Makes especially comprehensive highlights really frustrating to watch.

- First time shots. Just like the thing with defenders clearing with first touch attacking players seem to have the tendency to shoot it first time even with an open path to the penalty area ahead of them. Balances things a bit but I bet that this is more likely to be fixed than defenders hoofing needlessly.

SI for me have failed to improve the intelligence notable for years and this is simply where they should be concentrating on during this version and in the future. I think that this is a bit off-topic but of course this lack of intelligence especially is shown within big clubs as they should mainly step out of the masses in this department. 

[BlueAnderson]

7 hours ago, xzar_monty said:

This has been a problem with FM for years: size and strength, both individually and together, beat skill. In real life, it's quite hard to get the ball from Coutinho, and almost impossible to steal it from Iniesta. But in FM, both players lose the ball all the time, because they are small and not very strong.

Hazard has this problem on FM at times too. I do agree that player intelligence needs a huge focus. Shame it won’t be this iteration though. 

[Svenc]

13 hours ago, El Payaso said:

- How often defenders especially make the worst possible choice like: okay I just won the ball back, have three or four easy options available to actually profit my team and my manager saying to me: play out of defense. What do I do? Hoof the ball straight to opposition with first touch. Or: I'm under a bit of pressure and have to clear the ball.

 

Certainly not coded to happen, in particular as SI have tried to introduce a bit more risk attached to such "build nicely from the back" encouragements in more recent (as was also outlined by THOG when he was still around). As outlined, this is an assessment that has puzzled me as to prior iterations already, as you could comfortably sit on the ball even with lesser teams. My playing experience with 18 is still limited, but I already see my modest Mainz side building it cleanly from the back and making "risky" choices even straight after opposition attacking throw-ins, with all their players still pushed up in the final third of the pitch, the lb passing it short inside to the cb, who passes it to the midfielder, rather than any clearing. Currently running a match in the background where they thus keep the ball from Bremen all throughout, which personally, I considered as part of the weakness (and traditionally is naturally connected to a lack of actually pressure put on players).

Players just randomly hoofing it has never been a thing in more recent iterations. It's the opposite if you truly insisted on it, on any level, which you could then (to an extent) exploit due to the lack of pressure on the ball carrier. Retaining possession has always been a fairly effective tactics after taking leads on any level on this. That lower league players (lesser mental traits) are meant to prefer more "direct" options under pressure could be seen by older patch notes though. To which extent they actually did -- massively debatable.

[Mitja]

well said Svenc retaing posession is too easy, actually one would expect more direct football and more clearences after (smaller) teams take lead.  

expect more hoofing on more defensive mentalities and especially on lower creative freedom.

[Guest El Payaso]

2 hours ago, Svenc said:

 

In general this is true but there is that persistent behavior from defenders when they win the ball back either by tackling or intercepting and then hoof it long with first touch and by that nullifying their (possibly) good effort when they win the ball back. I see this happening couple of times every game and also noticed that at least in FM 2017. 

I would say that both 'play out of defense' and 'retain possession' as instructions still are too rewarding as it is too easy for teams to pull it out successfully. At the moment I'm seeing some possession lost because of those instructions but mainly because of some mindless ball across the field which is easily intercepted. Reaching the midfield and even more advanced areas by passing the ball short is generally really easy for any team as there isn't any kind of high press in the engine and also the midfield is not that well defended. 

[Scrench]

Instructing the keeper to distribute to centre backs is such a powerful instruction that I even use it when my centre back are as intelligent on the ball as a handful of dead butterflies. It can easily mean +5-6% to posession numbers at least, as there is no pressure on the ball unless the opponent fields 3 strikers which in turn completely destroys their deep defending.

[Guest El Payaso]

3 hours ago, Scrench said:

Instructing the keeper to distribute to centre backs is such a powerful instruction that I even use it when my centre back are as intelligent on the ball as a handful of dead butterflies. It can easily mean +5-6% to posession numbers at least, as there is no pressure on the ball unless the opponent fields 3 strikers which in turn completely destroys their deep defending.

Yeah if there is not going to be high block then teams should be even better at defending the final third but what happens is that teams allow the team in possession reach the midfield really easily and also there still are really big issues with strikers being able to link up especially in support duty. The movement from both attacking and defending players should be more dynamic: if an attacking player for example makes a move to come deep to get the ball the marker should also be equally active in terms of following that movement. What currently happens is that DLPs especially are summing 100-150 pass attempts, a CM without playmaking duties ~90 and strikers something between 40-60 which clearly tells you a story about low block "working" in the ME and meanwhile something like a ball playing defender is useless as they are hardly touching the ball. Much more dynamic movement is needed which would force teams to retain possession more in defense and would make for example slow tempo attacks more risky if you try to move the ball like this in the final third or even middle of the pitch. More aggression and a sensible defensive shape would really be appreciated and I hope that SI will drop strikers to defend during FM 2018.

[MBarbaric]

2 hours ago, El Payaso said:

Much more dynamic movement is needed which would force teams to retain possession more in defense and would make for example slow tempo attacks more risky if you try to move the ball like this in the final third or even middle of the pitch. More aggression and a sensible defensive shape would really be appreciated

This!!!

[Guest El Payaso]

1 hour ago, MBarbaric said:

This!!!

Had to watch one game from Barcelona and actually make some report of these to the bug area. I did this previous year on all of those issues that I am seeing and didn't really get agreement on them (or even opinions) from SI. Hopefully this time round it goes better.

Some of the stuff that I saw from Barcelona was actually good looking and even though they ended up with 0-0 against Alavés I saw a lot of potential on them actually performing. Just a couple of issues that should make it better if these are 'fixed':

- The effects of low block. Defenders (centre backs and full backs) need to be more active in terms of taking space away from the player who is receiving the ball. Saw Bojan running the game against Barca as Umtiti and Pique were just hanging around instead of one of them working as a stopper and actually bullying Bojan out of the play. Same with Alba especially on left flank: always letting Wakaso to be an easy outlet for passing for them. Free collect and he can pass the ball to one of the midfielders. The pressure has to come before the player receives the ball or even better: marking should make players unplayable at times aka create a feeling that "I cannot pass the ball to him as he is marked and we will lose the possession". 

- Offsides: the offside awareness is awful and Barcelona with world class players recorded 10 in total. So many of their attacks with potential were ruined by this. There must be better awareness and I hope that people who actually complete full seasons will provide save games for SI to inspect as I saw a worrying message on the bugs forums where someone from SI told that they are "happy with the amount of offsides currently".

- Midfield positioning. One of my favorites for years. A little bit of moving the ball around and it becomes a mess. Alavés midfielders summing 100-130 pass attempts against Barcelona is simply just awful. Sakai(?) a player with 12 off the ball was getting more than hundred I think in an attacking role when Barcelona has Busquets there to "mark" him. 

- Finishing. Barcelona didn't create anything that easy but from those two CCCs and other chances with their quality they should finish one, two or even three out of them. They get those chances here that they are able to convert often in real life. 

You can watch the pkm if you want. I will tag it to this post. Recommended to be watched on comprehensive.

AI Deportivo Alavés v FC Barcelona.pkm

[Svenc]

22 hours ago, El Payaso said:

- Finishing. Barcelona didn't create anything that easy but from those two CCCs and other chances with their quality they should finish one, two or even three out of them. They get those chances here that they are able to convert often in real life. 

You can watch the pkm if you want. I will tag it to this post. Recommended to be watched on comprehensive.

AI Deportivo Alavés v FC Barcelona.pkm

Even looking purely at the shot map, if that ever came to pass, every match on this would end 5-5, or Barcelona barely ever drop a point (an instance where I am glad that SI seem to take research and studies into account, rather than YouTube clips or anecdote).

The two Dembele shots significantly favor the keeper from that angle (comparable to El Shaarawy's first time shot against Sweden last week, technically gifted as it was, the keeper had but his arm to stick out). And the one in central space and range from Suarez he was pressured enough, same to the keeper having closed the angle already -- I may be some concerned with how he finished it off though, perhaps. That's naturally all a bit subjective, but a draw looks a totally game result to me (even if this were actually football). But then my argument has always been that the actual issues that existed in particular previous were that the game would/could create too many weak shots, rather than the other way around. On older releases they were also readily awarded a clear cut tag (whilst vice versa, something like this which is far harder for the keeper as he's handicapped by a) the assist from out wide b) the ball changing instantly direction with the shot c) which is also made first time, e.g. off the first touch isn't flagged a CCC or anything), ... which will all naturally remain as behind the scenes this is obviously a completely semantic debate. Like: "Should a header finish ever be allowed to be a clear cut". When it should be: What did our "shot models/ calculations say how likely it was all to go in?" I'd ditch the thing altogether, no least because 9/10 players don't have a clue what that stat even is (at its most "ideal", which hasn't proper worked in a decade now), but that's another topic.

[Guest El Payaso]

Might be. Hard for me to see but for example the only shot from Messi is the one that he often does score IRL even though it is not even a HC, of course he also misses lot of them. With better offside awareness they would have though created those easier chances that they need. The only things I'm really worried about when viewing that game is the ineffective defensive shape that Barcelona create: they leave so much space for Alavés to build from the back and even to create something dangerous and the complete lack of intelligence in terms of offside awareness. 

Also one thing to note about that also applies in general is: how goals are spread. Even the AI might get realistic amounts from players like Benzema and Suarez who play in the middle as they are constantly in those areas where they can score while I for example notice in this game that Messi is wandering completely out of play here and he probably is an inside forward so he should have the freedom to wander more on the pitch and be there where he can score and set others up. He mainly seems to do that when he is dribbling with the ball. If you set someone to "roam from position" he should do it a lot more especially intelligent players like Messi. 

[Rooneye]

On 27/10/2017 at 19:18, Neil Brock said:

So to clarify did you run the league in full match or quick match engine? As that'll have a substantial bearing on the result? 

And if you run a test like this surely you've come to a conclusion? If you think it's a problem then please raise it in the bugs forum. But as said we'd need the above information. 

 

On 27/10/2017 at 19:36, Neil Brock said:

Did you get sacked by either team? Why don't you raise an issue in the bugs forum with relevant save games if you think there's an issue? Thanks. 

 

Do YOU think this is an issue? Lets assume you did a test and the test was did correctly and these numbers are what is coming out, what conclusion would YOU come to? Would/is it a problem you think would/does need looking into?

[Svenc]

Is there a more comprehensive review going on of AI on the purely tactical side? To me it also looks they have no concept of the footedness/PPMS of players. So Bayern play top heavy on occasion with Robben cutting inside from the right (PPM anyway), Coman from the left as right footed. Whilst centrally, there's already Müller, and not seldom, Lewandowski dropping off. On the one occasion, an individual skill move may still see them goalside. On the other, they may simply meet and have a picnic. Still nothing against the decisions that would make sides off narrow formations not advancing a single back to at least support the midfield some. Or those curious in-match Allegri switches to 4-3-3 narrow, "everybody boxside please!", with the entire midfield getting emptied and the backs flushed up the shores too. I'm no guru, but doesn't that seem iffy to anyone?

Speaking about tactics. I see that @Totalfootballfan seems to spread the notion that "defensive tactics" on this would be harder to beat. What we partially witness here may be also part result of a few tighter defending, yes. However, same as on FM 17, were your tactics were weak against 5-3-2 sitting deep and packing it in particular: You still are on your central overload schtick, which is  -- on average -- pretty effective due to ME weakness, not through "tactical nous". I don't think you will ever get the miracle results of the tactics you usually share against defensive AI, as teams just keeping an army behind the ball will be always a few harder to break down. And it usually paper cracks also over a few defensive issues (wide midfielder defending on FM17, overloading weakness on FM 18), which you always find an exploit to. However, the reason why you struggle to score more with tactics like these is because -- it's your tactics. Not even AI would regularly play like this when targeting a result, though some of the stuff comes near, and it encouraged point drops from top sides on FM17 already imo. Flushing the midfield empty/pushing every player boxside plus keeping play narrow and compact doesn't positionally stretch deep defenses anywhere. It would be pish easy to defend in  football. It is reasonably easier to defend in-game when teams sit deep, as you noted tends to happen if you're on a higher reputation.

There may be a connection -- are defenses too "tight"? However, what you experience specifically remains connected to your tactics. They wouldn't make sense in football in that scenario. It is encouraging that they struggle for openings in-game. The reason you experience this is that you ask yourself what is currently the most "efficient" inside the match engine. For this scenario it may disappear if you would ask yourself what teams do in football when trying to challenge massed defenses, as it's the completely opposite to this. Now about that AI...

[Svenc]

@ElPayaso More recent I've taken a look at a match you had uploaded some.  A derby between Atletico and Real Madrid where Madrid (both AI managed) didn't get any more than 4 shots off. I've always liked the idea of a match engine that would the guaranteed high shot counts of older releases in particular down -- even top sides on the day can struggle against inferior opposition to get hugely much in terms of quantity going. On FM it used to be in such a way that you (or the AI, which is naturally relevant to the topic at hand) merely had to field the likes of Hazard -- they'd run circles around defenses by their very presence. However, naturally, it's a balance. If SI tweak defending, they would need to tweak attacking, both can be also related to tactics, too, naturally. If they wanted to, they could produce a match engine where every second move would end in an easy tap-in in loads of space for the forward, and 10-0 scorelines. The vice versa, likely, too.

I have huge respect for guys who provide detailed match reports. I did report some stuff in the past too, but found it sucked the fun out of the game for me from a player's perspective -- at some point, you're only looking for flaws, or what you perceive as such. The premise of FM is a "make believe" simulation. That entire suspension of disbelief thing. However, two things that stuck out immediately straight from the first 40 seconds of the match to me.

 

1) Attacking positioning
2) Related to this: Arguably things that make life difficult for SI, as FM models the width of the pich based on the centre of the pitch, when arguably it should be the position of the ball
3) Both easily isolating players, not merely on the flanks, but in the centre too.

Sorry for the bad Painterly skills, they make mey eyes bleed to. The first shows a sequence of play where: Modric (10) presumably on a roaming playmaker type of role starts drifting out to the left right from kick off, leaving oodles of space unoccupied. This in turn leads to Benzema running out of options, and him getting easily dispossessed in the last picture, as due to Modrics roaming out to the left, he doesn't occupy his "natural zone", and also occupies the same space as the other guys. That was 14 seconds in, Real's first attempt to transition the ball towards the Atletico Box (failure). You can also see the distances between the wide players and the central guys opening up (second to last picture meant to mark it)

The reason this stuck out ot me is because it happened exactly the same way in the 2nd attempt too. This is 24 seconds later.




The roaming of Modric in action, where he tends to occupy the same space as the other guy throughout, and none of them providing options for easy ball retention on the flanks too (easily dispossed).




As FM models the width on the pitch not in accordance to where the ball is, but in relation to the pitch, that's arguably also arguably one reason why, same as in defending, you don't have teams positioning intelligently as units, thus isolating players in general. If players are isolated, they cannot possibly construct decent moves. Or, at the very least, are much easier to dispossess/run out of options.



Naturally, FM18 also seems to have nerfed "dribblings" some (Ronaldo, Hazard etc.). so isolated "skill" moves won't open cans of space all by itself, but that may be my perception in parts. Guys seem to be more readily "doubled up" in defending on the flanks too. The kind of stuff I'd watch out for some upon further patching tweaking, perhaps. I didn't look through the entire thing, but that pkm looks interesting.

[Guest El Payaso]

+1 to this definitely @Svenc I think there are serious issues with link up play too which is for me for example limiting the amount of goals for other positions but centre forwards. Centre forwards are scoring realistic amounts if you consider players like Benzema and Suarez while there are also longer list of CFs that seem to score far too many like Afobe, Ben-Yedder, Jermain Defoe etc. They are always on those areas where they can score regularly while the link up play from other positions lack and for example Lionel Messi seems to be completely roaming out of the game as a right inside forward. I noticed this when I was watching some comprehensive highlights from Barcelona and am seeing this too with my team. I never use wingers but still my wide players seem to tuck too much to the flanks when given the option to roam and having no instruction to stay wider. I would expect a lot better attacking positioning depending on where the ball is. Also something like CM(A) seems not to be willing to link up well with the striker even when having chance to break and an attacking role (if not for example AP(A)) should be aggressively getting above the ball instead of holding back for several seconds. And with roaming I would expect him to roam and offer easy passing options for the wide players too and even getting higher up the pitch to offer them.

Someone like Eden Hazard for example with option to roam will roam from one flank to another and is in the middle lot of time: he roams on those areas where he can have the ball and same most certainly applies to many other players like Lionel Messi and Cristiano Ronaldo. Ronaldo on my save has declined into an assist machine while Messi luckily is so far scoring more than goal per game in La Liga. At the same time most of the teams seem to have a regular goalscorer in CF position and many CFs are scoring much more goals than they do IRL.

[MBarbaric]

I don't think there's a specific problem with that space left by Real in the middle. Teams do this kind of things all the time in an attempt to overload certain zones, so Modric moving out of his position is fine (Guardiola would probably push his FB into that space to provide more passing options, but he is obsessed with possession...). The real problem is with, i assume, Kroos (8). It is his inability to recognize he should move away that breaks the move. 

And that is the source of all problems in ME. Players aren't working together. The roles decide what players do, but they don't "communicate" between themselves. While in football you have an aim, i.e. overload one side and switch the point of attack towards the other flank to isolate your winger/FB in 1v1, in FM you chose the roles and hope role combination will produce what you want to see. that is why so many people feel so lost as football produced isn't really what the user intended, but since it works anyway it is fine.

FM is more like cooking than football simulator. You have these ingredients (roles) and you mix them together in order to get a dish (certain behaviour from the players). IN football it works the other way around. You know what you want to produce, how you want to attack/defend and then you try to create situations that work towards that goal. 

oh, and those two strikers from Atletico, that really isn't how football is played

[Guest El Payaso]

Exactly @MBarbaric: teams should both attack and defend as a team. At the moment (and basically always on FM) the movement and activity in both attack and defense is really poor and by that the engine is not able to reproduce real football or anything near of that. I've watched my team, Barcelona and Real Madrid in this engine and with all of them the movement is really poor. The general defensive shape is lacking, defenders and midfielders when defending are not being aggressive enough and in attack the striker is always too far away from the rest of the team and that single handedly ruins lot of the attacking play. 

I think that SI need to concentrate on both of these: general defensive shape and how players step out of the 'banks' or 'lines' to mark and close down and in attack how players shift to the areas where they can be profitable. And most importantly: drop the bloody striker to defend! Even though that Atlético's 'fantastic' defensive doesn't seem to make them concede or lose games.

[Dave1990]

43 minutes ago, El Payaso said:

defenders and midfielders when defending are not being aggressive 

I agree with everything you have written except this passage. Defenders and midfielders seem to be aggressive, but in a rather stupid way, and this is correlated to closing down, set last year in headless chicken mode. The closing down is not choral but individual based on the position of the ball, FM matches from this point of view seems to me the children's matches where everybody chases incessantly the ball carrier. And in fact I often see two players close down the same opponent, but in areas of the pitch where doubling is absolutely useless indeed harmful.
 

[MBarbaric]

indeed, however, this is by no means a bug, glitch or however you want to call it. It works like that by design. It has been like that from the beginning and there is no realistic hope it will change. At least within the lifecycle of current ME. It could all go forward but only if/when the game moves on to another ME which will be the time when I buy new FM. In meanwhile, it will always lack something somewhere and SI will patch one thing opening up another gap always lagging behind how football actually works. 

 

38 minutes ago, El Payaso said:

I think that SI need to concentrate on both of these: general defensive shape and how players step out of the 'banks' or 'lines' to mark and close down and in attack how players shift to the areas where they can be profitable. And most importantly: drop the bloody striker to defend!

The trouble is, a lot of people live off the game (SI, youtubers ...). They depend on the game to keep their income. Any significant shift from current ME is a great risk for everybody as long as the game is profitable. That's why you won't see any FM youtuber criticising the game in any substantial way. They largely provide help on how to play the game, but not really how to advance the game forward. What is encouraging are the rumours that new ME is under development and this could finally bring in more football into football manager.

[Guest El Payaso]

3 minutes ago, Dave1990 said:

I agree with everything you have written except this passage. Defenders and midfielders seem to be aggressive, but in a rather stupid way, and this is correlated to closing down, set last year in headless chicken mode. The closing down is not choral but individual based on the position of the ball, FM matches from this point of view seems to me the children's matches where everybody chases incessantly the ball carrier. And in fact I often see two players close down the same opponent, but in areas of the pitch where doubling is absolutely useless indeed harmful.
 

Yeah basically they should defend more by positioning and by that making the player that they are marking either completely 'unplayable' or harder to reach. When they close down eventually it usually is most certainly the wrong thing to do. Currently for example playmakers and strikers are having way too much of the ball (strikers are passing the ball 2-3 times more than IRL) and this should be corrected by players stepping out of the lines and making those players marked so that the ball cannot be passed to them. Currently centre backs fold from so many long balls and stay too deep that the ME is really painful to watch at times. 

5 minutes ago, MBarbaric said:

The trouble is, a lot of people live off the game (SI, youtubers ...). They depend on the game to keep their income. Any significant shift from current ME is a great risk for everybody as long as the game is profitable. That's why you won't see any FM youtuber criticising the game in any substantial way. They largely provide help on how to play the game, but not really how to advance the game forward. What is encouraging are the rumours that new ME is under development and this could finally bring in more football into football manager.

The game has had better tendencies of those with the current ME and for example in FM 2013 strikers did drop to defend. Something like 'players stepping out of the line' or one player from each side always trying to win a long ball should be bread and butter in terms of realms of real football but I don't get why they cannot make those happen. I think that me and @CityAndColour have been reporting about strikers' free link up play ever since FM 2015 or even 2014 and have basically never had any positive response from SI. Anyone should be seeing that but there seems to be some unwillingness to try fix obvious flaws and this is strange from people who also are 'fans of the game'. 

[MBarbaric]

4 minutes ago, El Payaso said:

The game has had better tendencies of those with the current ME and for example in FM 2013 strikers did drop to defend. Something like 'players stepping out of the line' or one player from each side always trying to win a long ball should be bread and butter in terms of realms of real football but I don't get why they cannot make those happen. I think that me and @CityAndColour have been reporting about strikers' free link up play ever since FM 2015 or even 2014 and have basically never had any positive response from SI. Anyone should be seeing that but there seems to be some unwillingness to try fix obvious flaws and this is strange from people who also are 'fans of the game'. 

I don't think they are unwilling. Without a doubt, they know how poor the defensive phase in the game is. They know off the ball movement in the offensive phase is horrible, they know players act as individuals (or pairs at best) on the pitch instead of as a coherent team. There are plenty of people in SI that had or still coach and understand the game of football. As there is no significant change in the ME in last 3 years, I think they really have hit the wall with how much they can do with current ME. They try each year to improve one thing just to break the other and I believe they already have new ME being developed. If not, they can run this game as is for next 5 years and then they will have to find something else to do. Maybe long awaited "Chairman manager"

[MrPompey]

There will also be difference, some due to real life differences and FM differences in how the game compares to real life

2 Questions

1. Were you running the full match engine for all leagues that you collected stats on? or were the matches simulated . This is a key aspect

2. If you saw AI managed Leicester win the league in FM, would you (and I for that matter) put it down to silly ol' FM not replicating real life football

FM AI Manager Limitations

The AI Managed teams have limitations in how the adapt when losing late in a game e.g. they will likely only use the attacking formation thats associated with the AI manager. When FM switches formations it displaces players e.g. playing 4-2-3-1 wide with  say your AML and AMR roles filled by the best players in that position, going 4-2-4 often displaces one of those players and you may end up with 1 or none recognised strikers because FM jumbles the players with different formations, in addition if you switched to 4-2-4 with a sub on the bench i assume you may take off the AMC and put a recognised striker up front or even go to a back 3 putting a DC up front. Its these sort of tweaks which real life manager's do that FM can't do tactically. Additionally the real life manager would have many more attacking formations for such a situation than the single one the AI manager has

[Svenc]

1 hour ago, MBarbaric said:

If not, they can run this game as is for next 5 years and then they will have to find something else to do. Maybe long awaited "Chairman manager"

Even in five years from now and no touch, it would still be the most advanced management games sim on the market , despite its obviously imitations. Even if their would be several rival games springing up, inevitably start-ups anyway. This is talking 30 years of playing experience starting on Commodore 64/Amstrad 128k, and having witnessed pretty much any major attempt at match simulation. There's aspects where purely arcade sims a la Fifa are superior, arguably also in how they model zonal defending some. However their patterns of play are completely removed from football, as they have to be due to the huge compressing of playing time. They're dribble/arcade fests, like matches consisting of nothing but YouTube highlight reels. They'd also make for horrible management game engines thus, no least because they don't even simulate all that much. To me there were some odd decisions made in the past though, for instance the forwards, it's not as if they always were like that -- on FM 2015, they were even "customizable" to the degree initially that they would (unrealistically?) track back as far as your own box, further than any aggressive advanced midfielder. I don't think the game will be the same in 5 years time. Significantly rewrites have been made in the past, Every tech has limitations.

Still this thread is about adressing some of the balance in the current builds. It's not as if top teams ever were that "toned down" as they oft are now, were they? There were versions where you could put up totally schoolboy tactics, the individual players would still guarantee the points. If you forced the likes of Messi into having no options at all, they would "statistically" complete up to 20 dribbles per game, depending on which, as they would naturally try to start some each time that happened before getting dispossed.  For anybody who thinks this is nothing like football at all, which to me is a bit of an exaggeration, I don't think any much tweaking will fix it, ever. It may not even be purely technical limitations. It may relate to how SI view football, too, in parts. There's many aspects of this I prefer over prior releases, and the most puzzling decision for FM 17 (wide midfielders) is nicely adressed. However, the balance, which is also recorded in this thread, and may well relate to AI tactical decisions too, it hurts a bit the believability of the game worlds. All this is pretty repeatably on full sim after all, rather than purely one-offs. There too seems to be a struggle or a fear for kicking AI managers into the next gear, at least at the top, which the stats collected here are a sign of too after all. Some of their decisions seem erroneous.

[Guest El Payaso]

55 minutes ago, MBarbaric said:

I don't think they are unwilling. Without a doubt, they know how poor the defensive phase in the game is. They know off the ball movement in the offensive phase is horrible, they know players act as individuals (or pairs at best) on the pitch instead of as a coherent team. There are plenty of people in SI that had or still coach and understand the game of football. As there is no significant change in the ME in last 3 years, I think they really have hit the wall with how much they can do with current ME. They try each year to improve one thing just to break the other and I believe they already have new ME being developed. If not, they can run this game as is for next 5 years and then they will have to find something else to do. Maybe long awaited "Chairman manager"

Might be so as the engines between FM 2015 to 18 are pretty much the same with all of them having problems to produce anything sensible compared to real football and goalscoring/chances created scenarios along with stats are highly repetitive. I might as well toss FM to the bin or use commentary only until there is a new engine. Sad that I didn't realize this before this year's release. It makes the hope grow when you basically give up the hope on the previous version really early on and it creates an illusion that hey this year it's going to be better.

[tkg]

58 minutes ago, MrPompey said:

There will also be difference, some due to real life differences and FM differences in how the game compares to real life

2 Questions

1. Were you running the full match engine for all leagues that you collected stats on? or were the matches simulated . This is a key aspect

2. If you saw AI managed Leicester win the league in FM, would you (and I for that matter) put it down to silly ol' FM not replicating real life football

FM AI Manager Limitations

The AI Managed teams have limitations in how the adapt when losing late in a game e.g. they will likely only use the attacking formation thats associated with the AI manager. When FM switches formations it displaces players e.g. playing 4-2-3-1 wide with  say your AML and AMR roles filled by the best players in that position, going 4-2-4 often displaces one of those players and you may end up with 1 or none recognised strikers because FM jumbles the players with different formations, in addition if you switched to 4-2-4 with a sub on the bench i assume you may take off the AMC and put a recognised striker up front or even go to a back 3 putting a DC up front. Its these sort of tweaks which real life manager's do that FM can't do tactically. Additionally the real life manager would have many more attacking formations for such a situation than the single one the AI manager has

 In regards to question 2, this is why it's important to run numerous simulations and focus on various events within the game as opposed to getting too caught up in the outcomes.

 

I've got numerous simulations running (all fully simulated) but I've only gone as far as taking them up to game week 12 so far (in line with real life). For that reason I don't care too much about the fact that Southampton are top of the league in one simulation or that Liverpool are 15th in another. 

What I do care about however are things like the passing rates:

I've added a bug report related to the above, which gives some insight into the significant difference in passing numbers between the ME simulation and real life.

 

Some of the AI manager limitations you've highlighted aren't particularly insurmountable, such as throwing on more attacking orientated players from the bench when chasing a game.  i'd imagine that those kind of AI manager improvements aren't especially inhibited by the ME itself (as opposed to pressing behaviour of players), though it's perhaps foolish of me to assume as such  

 

[MBarbaric]

@tkg

I think it might be useful if you tried to run numerous tests for a full season and compare it to real life numbers. Not concentrating on individual teams as they change managers/styles but on overall numbers. especially if you could separate pass numbers in each third. I can provide these numbers from real leagues if you need it, but can't do it for the game as I don't have fm18.

[Svenc]

On that front, I think the shots / SOT FOR/CONCEDED would be more relevant. Not merely due to possible ME / AI limitations. But on FM and football, they should be fare more indicative of a team's generally performance. The shots conceded you would need to manually collect yourself. Without them thrown into the mix, that is half the picture. The season when Stuttgart where relgated, for a while they had as many shots (also on target) as sides being top of the tables. However, the also conceded a bucket load, all indicative naturally also of their then managers tactical style, which was quite risky. Come the end of the season they scored as if competing for Europe, but conceded as if they were worse than miserably HSV. 2nd Bundesliga, here we come.

I think comparing purely pass stats you're catching mainly noise / AI/ME quirks plus. At the end of the day, matches on this are about scoring goals and preventing them. If top sides in real football drop significant amounts of points, there is not seldom some finishing woes involved. In-game, that doesn't at all look like it, as far lesser sides get just as many shots going, pretty oftenly. Plus if top sides underperform, they do so via underperforming significantly in the shot rankings too. In real football, take a look at Ronaldo, even his purely shot data in the league is pretty impressive: 1 goal from 60 attempts. It's unlikely to last even without further tuning on Zidane's front, very likely as it's unheard of for his, even considering his adventurous decisions. At least it may not be a bug as such if the AI can get Lewandowski down to 3 goals from 15 matches in this.

[tkg]

I've collected the data for 3 simulations so far, all for the premier league. The overall average number of passes is 135% compared to real life (whoscored data).  135% of passing overall isn't too outrageous, but as my chart indicates the increase isn't spread uniformly across all teams. Weaker teams tend to have a much higher pass rate than real life, upwards of 175% for example.

Although it's only 3 simulations and not full seasons i'm seeing quite a tight range between the simulations, suggesting to me that although additional simulations will increase accuracy I likely won't see a dramatic shift from what I've already collected.

For example the average passes pg in the premier league this season (upto match week 11) was 449.7. My 3 simulations came to 581,587,572 (avg of 580)

Studying the specifics of these passes is naturally much more time consuming. My analysis so far seems to suggest that weaker teams or defensive setups are able to pass far too readily in the defensive and middle third. 

Interestingly i'm also finding that these increased pass rates aren't as highly correlated to possession as you would expect. It's entirely possible for a team to have 40% of possession in the ME but attempt in excess of 600 passes per game.