Is FM a simulator or an illusion?

Hi everybody!

I have never felt the need to post anything on a football manager forum before, but over the years I have had this growing, nagging feeling that I need to get ventilated now. And oh, sorry for the long post. I hope you can bear with me! :o

Is Football Manager really a proper simulator, or is it just a lot of unconnected buttons and wires while the Wizard of Oz gives us the show we want to see?

Quite frankly. The options and subtleties are endless, but in my experience there is not much you can do that really makes a difference. You can get higher skilled players. That makes a difference. But everything that would seperate a good manager from a bad, knowing the right tactics depending on player material and opponent strengths/weaknesses and so on, has very little effect on the actual scores. If any. Still, when I look around in forums like this, it seems that everybody is convinced about the importance of good tactics.

To back my nagging feeling with some actual proof, I did a small test. I replayed the same game five times with strategy set to overload. Then I replayed the game five more times with strategy set to contain. All other tactical settings are at default values. I also played five "reference" games with my normal strategy: control.

Overload Contain Control

3-3 1-1 2-0

1-1 1-0 1-0

1-1 1-0 0-1

0-0 1-3 1-1

0-1 0-0 5-3

Okay, it is not a massive statistical set. But it does show that the most aggressive and most defensive strategies are not very far apart. If a real team pushes desperately forward for 90 mins I would expect more goals, but also that the team would lose most of the times. For a 90 min contain strategy I would expect mostly 0-0 or 0-1.

Maybe the players are restricted by the "style" setting. So I did the experiment again with the style set to very fluid. This should allow attackers to join the defence in the contain scenario and the defenders to join the offence in the overload scenario.

Overload Contain

1-0 1-1

3-0 0-2

4-0 2-2

1-1 0-0

1-0 0-2

Still no major differences. Just letting in 1 goal in 5 matches with a head over heels crazy offence tactic seems suspicious.

Until now I actually went on holiday during the matches, with the "Use current match tactics" and "Use current team" checked. But maybe my assistent manager changes the strategy during the game? So I did a final test where I actually played the games myself (with nothing but commentary on though).

Overload Contain

1-0 1-1

1-2 2-0

0-0 1-1

2-4 0-1

9-1 1-3

More or less the same results, except for one crazy win by 9-1 (it seemed like the other team gave up in the second half). The 2-4 result may also have been tilted by a red card on my team.

So what am I trying to say here?

I am not naive. I know that it is extremely hard to make a complex simulator. I can understand that SI wants to keep the main probability algorithms simple enough so that there won't exist any tactical loop holes that eventually would be discovered by players and abused to achieve unrealistic results. But I do want them to at least try. They claim to make a simulator and they give us all these settings to tamper with. Then I want the settings to actually do something as well. This is my challenge to SI for FM2015: Give the player some credit for learning the tactical depths of the game. It is after all a game. It should be possible to aquire skill to sort out good managers from bad managers.

There is of course the possibility that FM IS realistic. A football manager is perhaps an extremely overrated guy with minimal influence on the team's performance. But in that case, screw reality! In a football manager game, the manager should be an important factor for the team's success! I am not talking extremely dramatic changes, but it should at least feel worth while to try the overload strategy when chasing that equalizer in the last minutes, even if it probably means another goal in the back. It should feel worth while to take out Messi by heavy marking, even if it means giving Iniesta a lot of open space.

What do you other guys say? Does anyone agree with me that playing FM is like riding a bike when you are 4 years old with dad holding on to the back tyre keeping the balance? As long as you don't look backwards it may seem that you are cycling. Or am I wrong? Is the impact of tactical settings useful enough?

Cheers!

I understand bits of what you mean.

After all the tactical "advice" from people like cleon & others in the tactics forum that's only really needed or SHOULD be only needed for your first few seasons, after that whatever league you play in you will have good enough player to out weigh tactical intelligence.

Once my team wins the league once I feel like I can pretty much change to whichever style & system I want and still win (easily) because of the players and people will try to defend that but they can't just look at Manchester city last season it's not easy to do. And even guardiola trying to convert barca from 4-3-3 to a 3-4-3 ended up with mourinho taking a title and that wasn't even changing the style just the formation with some of the very very best players in the world.

I have been aggrieved by this issue for a long time. I can never tell if anything I did made any change or not and even posted a few threads along these lines.

However, if the manager had very little ability to control or change the result, then by logic, everyone would get random results and definitely quite a few losses/ draws. But some forum users (see Cleon's thread on Sheff. United for example) have been able to get amazingly consistent results.

The only way I can fit together these two completely antagonistic observations is by hypothesizing that the manager does have control but these hardcore players have developed these techniques over 15+ years of trying things.

Looks like fm 14 is going to be even more depth (with AI adapting to tactics) so the casual players will keep getting flummoxed in trying to figure why the hell we getting these seemingly random results.

There is certainly an ME framework defined in which you are instructing. I.e. closing down behavior roughly behaves like this (prior to FM 2013 it was pretty much always one player engaging the ball carrier, which made for a limited diversity in closing down and defensive behavior full stop). The same goes for defensive lines, etc. Even the most extrem settings won't push your side to play hollywood balls all the time, as the code dictates that players may opt for more and sensible options regardless here and there (looking on how football is played, probably for the better of it). So in that, you are correct.

The problem I see with your test is that you look at strategies in isolation. This is a matter of documentation, it is crucial for SI to finally provide a documentation proper rather than relying on community work to pull this off. The problem, or one problem is none of the strategies inherently have a huge effect on the shape of your team. For instance, you might have set your full backs to defend duty, which means they will provide cover in all of the strategies you picked. As such, going overload in isolation would make your side adopt a more risky, forward-pushing kind of play (more forward balls), more tacklings, a more aggressive d-line, etc. But it might not necessarily expose the side as much as you'd think it would do when going forward, in particular if you're playing bonafide holding roles in midfield on tops, your defenders are reasonably quick and/or the opposing attacking players aren't quick to start with. Furthermore going a riskier strategy doesn't equal more goals straight by definition, as going more risky obviously also has the side-effect of chances being more rushed, and then there is also the play influencing all those games you loaded up. The AI adopts to score lines, for instance, usually. When it sits deep and keeps men back (perhaps after opening the scoring), going overload might only encourage to play into its cards.

Looking at the entire statistics of a match might be more telling than looking at score lines in isolation, as they tell very little about how the matches are actually being played. I did this same of this in FM 2012, I think, and going contain/defending with a team that was able to reasonably contain (i.e. one being better/not inferior to the opposition and being able to play a bit keep ball even when under pressure) caused a lot less shots taken on either side of the pitch, for instance, whilst going the opposite, well, vice versa. Actually the few shots my team then had were almost exclusively after set pieces. The end results might have been often similar, but without context, score lines lose meaning. Similarily, in my last FM 2012 save, I went to Norway, Frigg, to employ a bit of an Egil Olsen style there. After a couple seasons and wheeling&dealing the required players, that is reflected by the statistics as well as the play. Point being there are score lines, and then there are the matches that lead up to them. Looking solely at results you will never make much sense of this.

That said though, having a team of great players has always been making things easy on FM. Having a go in La Liga with Real or Barca it's borderline impossible to finish out of the top two unless you're going completely crazy formations, and the similar is true if you get to that point in a long-term save by developing and nurturing your own players. At one point the players are good enough that they even make up for instructions that see them in isolation (which in itself arguably is realistic to an extent). Their frequently won dribblings, headers and one on ones (see the stats) will open up attacking space all by itself, their good decision making will do the rest. Overall I think though it's not that realistic, as such clubs should provide their own challenge, and the few times one of the stars slightly objects to warming the bench in my opinion doesn't cut it. This is arguably more related to man management, board/fan/press pressure and of course an AI that isn't all that in long-term squad building.

Edit: A rather simplistic look at this, but still a case in point:

Here's the heatmap showing the average positions of Lazio facing Napoli away. I changed the "auto" duty of the contain full backs to attack, as else they'd be on defend (on overload they're on attack). You could argue that goes against what containing a match is about, but I did this just to keep instructions a tad more equal, since the initial test was on all accounts done by changing but the strategy, not anything else. This is no cherry picking, but the first matches being played.

Here's the match played with the overload tactics, which ended in a 1-2 loss. http://www.pic-upload.de/view-20757736/overload.png.html

Here's the same match reloaded and re-played with a contain strategy, which ended in a 1-4 loss, actually (again, don't focus on results without context) : http://www.pic-upload.de/view-20757737/contain.png.html

As you can see, the contain and defensive strategies encourages players easing off, timing their runs later, dropping deep, playing with less width. The opposite goes for more attacking ones. I didn't take shots of interceptions, as more aggressive strategies encourage your team to also win back the ball further up the pitch, and vice versa for the less aggressive ones, and that showed in interceptions and tacklings in Napoli's half of which there were more when going overload. The contain match saw fewer of each than the overload one. Now those are two fairly similarily matched sides. If you take a team like Real for instance, and have them have a home match against most of their league competition, they will likely dominate play and shot count and not get pinned back even with applying a contain strategy. The reason this is the case could be quickly seen if you took a look at the play: There might be less forward and risky balls being played, but the opposition is being outplayed regardless, and eventually the far superior home side will end up in the final third regardless, and take a shot. This is also influenced by PPMs, you won't get Ronaldo's dribblings and attackining moves out of him, for instance, or similar traits out of other players.

Strategies are but one concept in the Tactic's Creator. In broad terms, they are meant as an instruction that determines whether your team should focus on defending/mix risk and reward/focus on attacking. This is also influenced by basic formations, a side heavy up-front naturally will find it harder to merely keep it safe than one that has players and options deep, and more crucially, there are roles and duties too, the latter of which for almost every role dramatically influencing shape when in possession. Likewise, going contain and applying a couple shouts such as "play with even less risk" will stop pretty much any kind of attacking risk whatsoever. Simply decreasing mentality settings to zero, which just picking the strategy does, for instance, doesn't cause wingers to avoid dribblings altogether, for instance. These are a good thirty seconds of containing a match, drastically:

Nothing of this might show in the score lines necessarily. In the play, it does.

Svenc,

I agree that there is some effect. But it is not enough, in fact I think it is next to negligable. And if the tactical variations show in play but not in score lines, then what is the purpose?

One reason I think why SI does not really want to explain how things work, is that it would reveal some of the illusion. I guess a lot of the playability of FM is because we see things and connections that simply aren't there. The simulator is not as complex as we think it is. I may be out of line here, but this is my conclusion after three years of playing. It's still an addictive game though .

Seeing as most scores tend to be similar in real life, I'm not really sure what point you're making.

partially agree with OP, to the extent that for me it's always been difficult in FM13 to spot exactly why things are or aren't working. It's difficult to put my finger on just what I'm doing right or wrong. I do think it will be better in FM14 though.

Seeing as most scores tend to be similar in real life, I'm not really sure what point you're making.

The OP is trying to determine if a user change such as match strategy results in any significant change in the results. He shows that it does not and hence you could logically argue that changes that a player makes have little impact on the results.

I don't necessarily agree with it but with the fogginess that SI maintains about a match result, its hard to tell if its an illusion or not.

No, it won't lead to significant changes in the results, not without all the other aspects of the game (such as motivation, player quality etc) factored in as well.

These sorts of scores:

1-0 1-1

3-0 0-2

4-0 2-2

1-1 0-0

1-0 0-2

will be seen regardless of your own tactics* because so much of the rest of the game, including those factors affecting the opposition, have just as big an impact.

*making the assumption that you're using good balanced tactics

The best managers are better able to react quickly enough in a match situation and correctly assess what changes need to be made to provide the biggest chance of success - but even the best tactical manager will find themselves in trouble if they ignore the rest of the game.

I agree that there is some effect. But it is not enough, in fact I think it is next to negligable.

But you're only looking at end results, not what is actually going on, how your instructions influence play. On average one side being more open might result in more matches with a significantly higher amount of goals on average (in your little sample it does, btw. though we still don't know HOW open you actually are, this is more than just a pick of strategy) It is still by no means a good way of a) learning to play the game and b) getting the gist of what is actually going on. And it's not proving much of anything. If the two sides you picked for your small amount of filters are evenly matched, those results aren't particularly surprising, tbh.

One reason I think why SI does not really want to explain how things work, is that it would reveal some of the illusion.

I was more talking about the TC settings, which will see a re-work and will replace tactics sliders entirelly for FM 2014. Strategies aren't the be end of all when it comes to influencing a match and turning matches into your favor, far from it. A counter strategy can be very attack minded, a straight attacking strategy a tad more probing, etc. That's going to be the documentation problem of the future. It is nicely addressed in these forums, for most of the part, but not very much officially by SI. The strategies also are not as static as some of their name tags appear. One of my personal pet peeves for instance is that "contain" isn't quite as what I'd personally envision containing to be. By default not anyways.

For the play you see in the video above, which really aims to see a match out and not much else, I had to additionally change duties so that players don't push forward so much and prove options for deep play, change to a formation employing two DMs for enough passing options deep for likewise reasons, tweak the wingers so that they would avoid dribbling as much as can*. A successful dribbling for instance causes play and the ball advancing deep into the opponent's half, which in turn leads the entire team pushing up, which in turn frequently turns into something I wouldn't like going that extreme (which is avoiding attacking pretty much altogether): the team eventually approaching the goal mouth and taking shots. I think FM 2013 is actually a tad more forward-pushing and less lateral in regards to play in general, which doesn't make sense as matches in FM 2013 are about 20 minutes of play longer and SI don't need to rush things to get to realistic results in any kind of way in which say FIFA Soccer would. But that may be an influence of SI being bloody Brits mostly, who knows.

* Whilst I adore PPMs loads -- in which management game have you ever seen somebody selling Robben because he's sick of him dribbling, cutting inside and taking pot shots?? -- I think researchers should be more sensible when approaching them. Whilst some players may have moves that frequently would break their manager's plan and cause frustration sometimes (Robben's cutting inside, Lucio pushing forward), how many players are there that do so actually? But that is the current weight they are given in the ME, they reign supreme over anything else, from my experience anyways.

You need to look in terms of results and points, as well as goals scored.

That's quite a trend! According to this, overloading without considering the team's shape is a bad idea. Making the shape more fluid and being prepared to make changes has netted more than 3 times the points. And double the goals.

I don't think you should be so quick to dismiss that 9-1 either. You designed a tactic to get goals - it got goals! Contain scored 2 goals twice, Overload scored 2 (and then some) 5 times.

And here we see the exact reverse trend. Total Football and Conservatism don't go so well together. And your AM clearly doesn't know how to make it work!!!

As Ackter says, if you use more balanced tactics the differences will be more marginal. Then again, the difference between West Brom and Wigan last season was 0.34 points per match (roughly an extra win every 10 matches). West Brom finished in a comfortable 8th, while Wigan got relegated in 18th. Spurs missed the Champions League by 0.05 points per match!!! Marginal differences can make or break a season.

And just to point out the difference in goals scored over 15 games of the test:

Overload scored 28 and conceded 14

Contain scored 12 and conceded 17

Two of those tests were played with a fluid style, so in my opinion the goals conceded with the Contain strategy is artificially high.

In terms of goals scored though, tactics clearly made a difference!

Or his players and tactics just don't suit the control strategy - its one of the hardest to use correctly.

The question is, does the game replicate the problems facing a rl manager?

If it does are the solutions (or lack of them) represented in a rl manner?

I've always felt the real missing factor in FM is the impossibility to converse and that's not going to change any time soon.

Playmaker,

I see what you are saying. But the few games I played are probably not enough to make any conclusions about minor trends. Maybe I will do a more thourough investigation sometime, but it takes so much time, so I thought I would stick out some probes here first to see what the general views are .

What is obvious from the few games I played though, is that there is no significant difference between the two most extreme tactics. And note that we are talking tactics that "should" only be used in the end of a match because the nature of them "should" normally backfire over longer periods of time. I would expect both overload and contain to make you consistently lose games against an equal opponent (with occasional exceptions of course), if those tactics were really implemented they way they are described in the game. In my mind, overload should result in average scores like 3-5 or something, and contain should result in scores around 0-1. And if the difference in outcome from the two most different tactics is so small that we can't be sure if we see any trends at all in five matches, then imagine if we would try to detect the influence by more normal use of tactics.

As I said, this is only a small test. But I made it since after three years of playing I started to realise that all my hours of researching and tactical fine tuning didn't make much difference in the end. And the test didn't really contradict that feeling.

You're thinking way too simplistically about the tactical variations, completely missing out the intricacies of how those would interact with the other elements of the game, or even the opponent's tactics.

The changes are seen on the pitch, not in the final scores. Just like real life.

a very intelligent and focused post! you can say that the empiric evidence is too small, and that other control variables like moral and fitness etc should be added. but i feel he is on point in that it is difficult in determine causality in many aspects of the game that are supposed to have clear effects. the instructions on playing style is one such thing that you would think has a big effect.

this leads me to what i have been discussing before about a lot of things seem to be cosmetics. like the training system; i feel no control there no matter how i mirco manage it(the most important thing is just to avoid injuries and give youngsters playing time).

could it be that the producers are so concerned of getting unrealistic results, that in their incessant simulations and subsequent corrections toward what is percieved as realistic, that it has affected our ability to have impact on the game? while i know that there are alot of factors that determines success in FM, like giving the team time to adjust tactics and blending in players(thus not changing tactics too often or use too many new players), and the quality of players, the weather and pitch contidions, moral, etc. but the potential to be a true tactical wizard is not really there.

You're thinking way too simplistically about the tactical variations, completely missing out the intricacies of how those would interact with the other elements of the game, or even the opponent's tactics.

The changes are seen on the pitch, not in the final scores. Just like real life.

his hypothesis is very clear, it's not simplistic thinking. given a certain amount of matches, the changes he made had little significant effect. that's cause and effect. you can argue about other variables having impact, but then you are speculating. he should have told about other things that might influence his result, but that is a more complex study.

i don't get your last sentence at all.

His hypothesis isn't simple thinking, but his idea of how tactics work appears to be so.

Last sentence is simple. A team's approach to a match will never be the be-all and end-all, there's far too much else to consider to make it a worthwhile hypothesis.

A team on contain should not be limited to 1-0's or low score lines. It can work that way, but it definitely should not be the norm. Teams sitting back invite pressure, some teams can penetrate a deep lying team better than others. There are so many variables to take into context, the contain strategy is just one of them. Focusing on one aspect, and expecting exact results is always likely to fail in this game.

I'm a Contain player, and I regularly win 3-0 +

The test isn't very good "science" since the lack of difference between the two extremes only shows that the base tactic doesn't make a huge difference one way or another.If a tactic is very bad or very good, its user assumes an inherent advantage or disadvantage compared to the opposition and its tactic. If it is neither, its usefulness is restricted to allowing the user to play matches at all, while the result is subject to chance unless the user controls the outcome with manual input during the match. In this case the tactic seems to be of the latter sort since the three experiments had little variation in outcome.

In other words; the base tactic seems to be insensitive to strategy changes, which is what an experiment on the sensitivity of strategy changes showing little sensitivity will invariably show.

The test isn't very good "science" since the lack of difference between the two extremes only shows that the base tactic doesn't make a huge difference one way or another.If a tactic is very bad or very good, its user assumes an inherent advantage or disadvantage compared to the opposition and its tactic. If it is neither, its usefulness is restricted to allowing the user to play matches at all, while the result is subject to chance unless the user controls the outcome with manual input during the match. In this case the tactic seems to be of the latter sort since the three experiments had little variation in outcome.

In other words; the base tactic seems to be insensitive to strategy changes, which is what an experiment on the sensitivity of strategy changes showing little sensitivity will invariably show.

There definitely isn't enough difference between the extremes of play styles imo - but its not the difference in results that bothers me, it's difference in the actual play styles. I'm hoping for an improvement in this area for FM2014.

Anyway to answer the OP simplistically it's a simulation and is developed as such.

In other words; the base tactic seems to be insensitive to strategy changes, which is what an experiment on the sensitivity of strategy changes showing little sensitivity will invariably show.

What is the "base tactic"? Changing style to Very Fluid and strategy between Contain and Overload should pretty much change the tactical behaviour of all my players to the maximum possible extent. That's what those choices are there for. I could of course change the line up between 5-5-0 and 0-5-5 or something to get it a little more extreme, but that is not the point.

his hypothesis is very clear, it's not simplistic thinking. given a certain amount of matches, the changes he made had little significant effect. that's cause and effect.

Since nobody had a look at the matches at all, but only the results, nobody can tell. The results are also skewed by playing the same team again and again -- if the odds are even, there won't be much gulf in results likely, and if the odds are skewed toward's the human team's favor, i.e. the human team was superior, even going contain for 90 minutes won't invite much pressure, and going overload against an underdog might result running battle-ram against a team sitting deep.

The tactics also weren't drastically changed, all was changed was the strategy, and philosophy. Those are but pieces of a puzzle. If the full-backs weren't on auto-duty, for instance, this means the team shape was comparably similar when going forward. When they were on defend duty, and when a lot of holding roles in midfield were employed, even going overload won't suddenly turn that into a side exposing itself and rushing head first towards the opposing goal.Again, this is down to lack of documentation that leads to misconception. Still I have already shown simply changing strategy has of course an effect, a big one, but as real-life football, this isn't like a strategy game, in which you pit unit X against unit Y and you know the outcome will be Z. This isn't like that, there is ambiguity to it. In influencing play, (which you do, as does the AI who may change course throughout a match), you are dealing with probabilities. Facing an opposition pushing men up, it might well be worth-while sitting deep and hitting them onto the break, in particular with speedy forwards, etc. You DO have an influence over this, as shown. Sitting extremely deep against far superior opposition might well cause your side to totally crumble, etc. Going up there are two possible routes: keep on attacking or playing it more safe, but this decision is yours to make. And this is rather basic, you could build entire teams around more physical and direct or more possession based play, currently I do the former, and there are loads of goals scored this way when compared to the league average (currently this is an FM 2012 save I might add, as physical players are much more favored in FM 2013). Insisting everything to not have any effect (which, as shown, is false), will lead to exercises in frustration.

If you don't recognize such, and keep on looking solely at results, you will never ever make much sense of this. FM ain't perfect at simulating this all, and there are limitations, but it is apparent it's been heading towards that general direction. This isn't say FIFA or PES Manager, where nothing you will do significantly changes play any, as both the engines as well as the structure dictate that teams need be

What is the "base tactic"? Changing style to Very Fluid and strategy between Contain and Overload should pretty much change the tactical behaviour of all my players to the maximum possible extent. That's what those choices are there for. I could of course change the line up between 5-5-0 and 0-5-5 or something to get it a little more extreme, but that is not the point.

The base tactic would be the formation plus the role and duty selection at Balanced and Normal. And no, you will still attack with Contain and defend with Overload, just differently. You are actually more likely to be countered upon when using Contain even though the Defensive Line is minimized because standing off all over the field and having 10 players refraining from running forward will still see the team push up when the opposing team have the ball on their third. Then they whack the ball over your defense for an easy goal. Similarly, on Overload you won't see that kind of counter-attack because all the players will chase the ball like dogs all over the pitch leaving them no time on the ball to do so. However, if they manage to play calmly through your lines they can counter that way. On the flip side, on Contain your team will deliberate carefully before finishing an attack, so efficiency will be higher than the happy-go-lucky attacks on Overload. The end result may well be the exact same score-line.

Since nobody had a look at the matches at all, but only the results, nobody can tell. The results are also skewed by playing the same team again and again -- if the odds are even, there won't be much gulf in results likely, and if the odds are skewed toward's the human team's favor, i.e. the human team was superior, even going contain for 90 minutes won't invite much pressure, and going overload against an underdog might result running battle-ram against a team sitting deep.

The tactics also weren't drastically changed, all was changed was the strategy, and philosophy. Those are but pieces of a puzzle. If the full-backs weren't on auto-duty, for instance, this means the team shape was comparably similar when going forward. When they were on defend duty, and when a lot of holding roles in midfield were employed, even going overload won't suddenly turn that into a side exposing itself and rushing head first towards the opposing goal.Again, this is down to lack of documentation that leads to misconception. Still I have already shown simply changing strategy has of course an effect, a big one, but as real-life football, this isn't like a strategy game, in which you pit unit X against unit Y and you know the outcome will be Z. This isn't like that, there is ambiguity to it. In influencing play, (which you do, as does the AI who may change course throughout a match), you are dealing with probabilities. Facing an opposition pushing men up, it might well be worth-while sitting deep and hitting them onto the break, in particular with speedy forwards, etc. You DO have an influence over this, as shown. Sitting extremely deep against far superior opposition might well cause your side to totally crumble, etc. Going up there are two possible routes: keep on attacking or playing it more safe, but this decision is yours to make. And this is rather basic, you could build entire teams around more physical and direct or more possession based play, currently I do the former, and there are loads of goals scored this way when compared to the league average (currently this is an FM 2012 save I might add, as physical players are much more favored in FM 2013). Insisting everything to not have any effect (which, as shown, is false), will lead to exercises in frustration.

If you don't recognize such, and keep on looking solely at results, you will never ever make much sense of this. FM ain't perfect at simulating this all, and there are limitations, but it is apparent it's been heading towards that general direction. This isn't say FIFA or PES Manager, where nothing you will do significantly changes play any, as both the engines as well as the structure dictate that teams need be

, and shoehorning more intricate management instructions into AI routines never meant to replicate a management experience first and foremost proves instructions lost in translation.

Exactly what he said, fantastic post.

The base tactic would be the formation plus the role and duty selection at Balanced and Normal. And no, you will still attack with Contain and defend with Overload, just differently.

I wish this was changed a little. wwfan always says that all the strategies would have both a defending as as well as an attacking element to it, bar the most extreme ones, but obiously this isn't quite this extreme. Whilst it is possible to see out a match to an extent and close shop, there needs be more backwards and lateral passing (for contain), less attacking moves in general, but also more hollywood balls for overload, maybe. Also the d-line instructions apparently are only applied when defending, that is when possession is lost.

Since PaulC can focus entirelly on the TC's concepts in the future, maybe the game can be better tweaked towards making not only the roles but also the strategies show more profoundly. That'd be 100% win-win for everyone, not only those who preferred toying arond with the TC anyways!

Anyway to answer the OP simplistically it's a simulation and is developed as such.

Yes, of course. I was thinking about making a post about my thoughts about the simulation as well, but I never did because it was a bit too much of a rant and I basically don't have any real grounds for what I am thinking. But since you got me started I can just outline my thoughts .

Warning: All of this is quite speculative so don't take it too seriously!

I guess, note guess, that FM is sort of a "top-down" simulator. A few major input parameters, like average team CA, home/away, current form and probably some hidden parameters (e.g. "bad luck"), etc, make up the base for the probability calculation of the score. Then the match is "rigged" to produce that score. However, there is some room for changes during the match, like team talks, shouts and substitutions (and to a very small extent, change of tactics).

The opposite would be a "bottom-up" simulator. In that case every single player has their own simulator. The match starts as a blank sheet and evolves when all player probability actions interact with each other. This is how most of us probably thinks a simulator should work. I guess.

For a top-down simulator the actions on the field are determined by the calculated score. For a bottom-up simulator the score is determined by the calculated actions on the field.

And what has this to do with anything? Maybe nothing. But I suspect that the lack of possibility to actively affect the score by changing tactics etc, could be a consequence of using a top-down simulator. It could be a bit of a mess to drastically change the already calculated match scenario, mid-game. With a bottom-up simulator it would not be a problem, because the future has not been written yet.

But with all that said, I do not blame SI if they went for a top-down simulator. Because a bottom-up simulator is probably a heck of a lot more difficult to get working in a realistic way. It would easily lead to chaos phenomenons and weird situations without careful and excessive restrictions/rules for each player. Someone ever heard of the "butterfly effect"?

On the other hand. With such a simulator working I think it could produce wonderful matches with much more obvious cause-effect on tactical orders and more visible effect of each player's attributes.

As I said, I don't have much ground to back up my guesses. It's mainly the lack of causality effects and the fact that the players' ratings are mostly influenced by the score rather then what they have been doing on field. This is by the way the one single issue that I think SI should focus most on (besides more tactical influence during a match). The only feedback we get about how well a player is doing, is the rating. This is the most important piece of information you have to base your tactical decisions upon. Still, the rating system feels mediocre at best. It's irratic and simple and this makes me wonder if SI actually understands how important it is. Maybe it is well thought through, but in that case SI needs to communicate how it really works. I get that you get a big minus if you make a mistake leading to a goal. You get a big plus if you score a goal or assist. You get a common plus if you are winning the match and common minus if you are losing. These events make up for the biggest part of the rating as I understand it. But it should be the other things, like how many correct passes you have, how many duels you won, how many intelligent runs you made, etc, that really defines how well you are playing. And those things tend to drown. Not to mention that the rating can pop +/-1 back and forth sometimes without no obvious reasons. All in all, I don't know if I can trust the ratings to give a fair view on how good a player is performing. But hey, now I got off track ranting about something completely different. Sorry about that!

I'm pretty sure you're completely wrong about the nature of the simulation. It absolutely does not determine score first and then work backwards from there.

I'm pretty sure you're completely wrong about the nature of the simulation. It absolutely does not determine score first and then work backwards from there.

Yup, he's actually got that completely wrong.

The whole match is calculated when the match begins, and each time a human player makes a change.

This is the only way they can make "highlights" happen in a sensible fashion. The ME is a probability machine, taking the input, calculating the output which is a 90 minute simulation that we experience "live". It is very obvious when a calculated event begins and when it ends; when the ball is put into play and when it leaves play - tactical changes only takes place after such an event is finished so we are watching something that is already decided. So far it is easy to see how the game is made.

The problem begins when one starts to wonder whether the whole sequence that ends when the ball leaves play is calculated cause-and-effect from the first second to the last in an open-ended exploratory journey, or if the whole thing is calculated at once like a complete module where the end result is a sort of guidance beam.

I think it is the latter, because how else can they make smooth transitions from one action to the other, given the seemingly advanced behaviour and decision-making each actor on the field displays from second to second. Programming a "tick-based" AI that does not fluctuate in action pacing and where every possible action's end aligns perfectly with where the next action begins seems like an unnecessarily complex way of doing this. Without having coding experience it nevertheless seems to me to be a much more efficient way of doing things to minimize the amount of graphical transitions by having larger "blocks" of actions with minor variations and only a few other such blocks which fit to their ends.

So as an example, a highlight starting with an attacking throw-in will have the first block lead to 1)a finish, 2)nothing, 3)a counter-attack. So 1-A will be a goal, 1-B no goal, 2-A corner cick, 2-B goal kick and 3-A finish, 3-B nothing, 3-C counter-counterattack and so on.

I just seriously doubt it is possible to have that Throw-in be an event with outcomes 1) player x - success, 2) player x - failure, 3) player y challenge, 4) player z challenge, opponent player u - success etc etc, then, one "tick" later 5) player x ball controll - success, 6) player x ball control - failure etc etc so that every statistics-influenced/ing action is calculated separately with an open-ended outcome until the ball leaves play again 50 000 such events later... all seamlessly blending with each other because the i.e 10 different outcomes of all 50 000 actions (or events) in a highlight/sequence have been coded together in 500 000 x 500 000 animations. Especially when each of those are influenced by millions of factors plus randomization.

Focussing on the word "simulator", FM is not and never has been this and can never be until with ball / without ball tactical planning is possible. We all know that is how real football managers set up their tactics but since the CM0102 wib wob option was decreed to be just an "exploit" it has vanished without trace. I never found it a particularly exploitable device, arrows should be the same and maybe there is even an argument for sliders but instead of making these realistic "tools" work they have just been removed (I believe sliders have gone this year?). Chalkboards with runs, feeds and player positioning zones (areas players are encouraged to stick to but not at all with glass walls) are the way forward. Alas, over 10 years since these ideas were first trialled they have just been discarded. Shame.

I haven't read all the replies, but you simply can't ask a team to jump between balanced and fluid etc. Its asking for a completely different way to play the game. I don't know anyone who would change this in match or game by game. Its a season by season decision if that.

Secondly, a team's knowledge on playing overload and counter and control all come down to experience too. I've had my one formation loaded up in the three different ways and the familiarity on each would be different.

What you've got there are a huge number of games played in a way alien to your team.

You haven't said how long you've had this team, how familiar they are to your chosen tactic or how good in comparison the opposition are. At least with your usual control you scored three wins, a draw and a loss. 60% win rate.

The question is, does the game replicate the problems facing a rl manager?

If it does are the solutions (or lack of them) represented in a rl manner?

I've always felt the real missing factor in FM is the impossibility to converse and that's not going to change any time soon.

You say "converse", I say "communicate", I guess this much the same thing. Communicating tactical instructions has become frustrating, unrealistic and kind of boring. I appreciate that a myriad of tactical roles for players is one way but not nearly flexible enough and no doubt the roles will be interpreted, re-interpreted and then ultimately described as "misunderstood" by SI, the tactical gurus and the experienced player.

The tactics also weren't drastically changed, all was changed was the strategy, and philosophy. Those are but pieces of a puzzle.

I still don't get that statement. Strategy and style changes ALL pieces of the puzzle (almost). It's what they are there for!

If you don't recognize such, and keep on looking solely at results, you will never ever make much sense of this.

If a tactical decision is not reflected in the score, but merely as changed statistics of ball possession etc, then what is the point of making that decision? Football is only about scoring the most goals, remember .

Your explanations are mainly focused on the fact that football is a complex game with lots of intricate factors that counter each other, which makes it difficult to evaluate single parameters. Or something in that direction, right? Are you sure that you aren't a victim of the illusion that FM is more complex than it really is? When you feel that a certain tactic works for you, are you sure that a different tactic wouldn't have given the same result? Maybe FM is just letting you see what you want to see?

I still don't get that statement. Strategy and style changes ALL pieces of the puzzle (almost). It's what they are there for!

If a tactical decision is not reflected in the score, but merely as changed statistics of ball possession etc, then what is the point of making that decision? Football is only about scoring the most goals, remember .

Your explanations are mainly focused on the fact that football is a complex game with lots of intricate factors that counter each other, which makes it difficult to evaluate single parameters. Or something in that direction, right? Are you sure that you aren't a victim of the illusion that FM is more complex than it really is? When you feel that a certain tactic works for you, are you sure that a different tactic wouldn't have given the same result? Maybe FM is just letting you see what you want to see?

No. Which is why I don't change tactics

Strategy and style don't change all pieces of the puzzle because of the limitations with how the individual sliders work and interact with the team sliders (even if using the TC, they still control sliders).

They will in 2014 though, judging by what I've heard about it so far.

The following has been confirmed a number of times over the years on this forum by SI, so I am just paraphrasing without any assumptions of my own:

Every kick of the ball is simulated from start to finish. The simulation doesn't know the final score until the final whistle is blown.

Where confusion arises is that the whole match is simulated before you see anything so that highlights can be chosen. This leads some to conclude that the result is pre-determined. However, when you make a change everything is newly simulated from that point. Every time you interact with the match engine it will re-calculate every kick up until the final whistle.

If the game match engine is really that complex and in depth which I sincerely hope it is then

A. Why doesn't SI provide detailed documentation and allow detailed feedback from assistant. Why instead maintain a fog of information so noone knows what is conjecture and what is truth

B. How do they test if any changes they made to code translated into real differences since even by playing the same match 10 times with the same initial conditions we get different results.

If a tactical decision is not reflected in the score, but merely as changed statistics of ball possession etc, then what is the point of making that decision? Football is only about scoring the most goals, remember .

The most basic rule of football is simple Maths: Ye who scores the most wins. The game itself however, isn't. I'm not sure if Heynckes had the single goal in mind going into that match against Chelsea On That Night In May, considering he encouraged 30+ attempts, a good chunk of them on target.

Your explanations are mainly focused on the fact that football is a complex game with lots of intricate factors that counter each other, which makes it difficult to evaluate single parameters. Or something in that direction, right? Are you sure that you aren't a victim of the illusion that FM is more complex than it really is? When you feel that a certain tactic works for you, are you sure that a different tactic wouldn't have given the same result? Maybe FM is just letting you see what you want to see?

I don't know how complex FM is. What I do know is through experimenting myself is that for instance I can build and force a team to attack mainly through the wings, encourage the FB's to additionally overlap, provide plenty of crosses into the box to tall strikers and see that reflected on the pitch. This isn't imagination, if you go to your squad screen, you can insert columns with statistics that otherwise would never pop up anywhere - and compare them to league rivals. As well as I can try to close out a Barcelona side that by team report evidently creates most opportunity bang through the middle by employing an additional anchor man in the centre. Equally, as another simple example, I can increase the likelyhood of the opposition tall striker getting a bang on my goal by fielding my smallest most fragile defenders if I want to lose my job. The calculations may be simple or complex, but they reflect your as well as AI decisions well enough, which works for me. I also know that players attributes are reasonably reflected in a season's statistics across the league, which you can try yourself: When he gets enough playing time, Scholes is still amongst the best passers in the Prem, Crouchie (as he is modelled via FM data anyways) wins a ******** of headers, Götze makes a lot of assists, Vidal (pet peeve of some) sees frequent bookings. All of that can be taken into account when transfering and managing in general. I don't know how complex it all really is at the core, no. What I absolutely do know is that approaching things the way you are doing only leads to frustration rather than enjoyment, however.

I still don't get that statement. Strategy and style changes ALL pieces of the puzzle (almost). It's what they are there for!

By changing the strategy you change how (generally) forward pushing your team (passing), how high up the pitch they are encouraged to win back the ball, how risky their decision making is going to be, tempo of play. Something like that anyway, no more, no less. Depending on how your FBs are set (auto duty or not) and your roles in midfield (auto? covering? attacking?), you won't necessarily drastically expose your team when the ball is lost and defending begins or underman it when going forward. To explain: auto duties means runs from deep are set to maximum on "overload", and minimum on "contain". If you set those manually, the change in strategy will be much smaller in terms of positioning.

What you do not fundamentally change thus by definition is a very important aspect of football, and that is how you utilize space. how players position themselves when going forward, that is the team's attacking shape. That is what roles and duties are there for. Getting them wrong can cause a world difference, not changing them at all might not do much harm/benefit at all. For instance: During the early days of FM 2013, someone on the many fan sites published a tactics that was called "FullOnAttack", claiming it to the best thing ever. As people like their download tactics, it spread and was soon caught out to only really work with totally dominating sides, as all players were on attack duty (runs=maxed out), which led to lesser sides not making up for the lack of any midfield cover whatsoever when going forward. Additionally, even with superior sides, it for some proved problematic, as it dramatically compresses attacking space. Unsurprisingly, which you might agree when looking at this: http://s7.directupload.net/file/d/3126/a2e9hcr3_png.htm

http://s14.directupload.net/file/d/3126/rw7rjfbh_png.htm

Back to the strategy, most you do is telling your team if they roughly should focus on attacking (more risk, urgency to win the ball back and forward passes) or defending (less of each). However even with the least risk, they will attack somewhat and play forward balls, if far less urgent, as has been said. And even with the highest risk, they will defend, if your team is good physically, maybe even for the better of it, as overload drastically increases closing down, d-line and tackling (which team stops defending altogether?). With a superior team, this might actually be of benefit, at least sometimes, as the inferior side could get totally dominated. With an inferior side, however, this might drastically backfire. Still, in that you are correct, the most drastic strategies indeed are not as drastic as they seem. Not in isolation anyways. But again, they're but a part of a bigger puzzle. Here's a couple scenes, not all of which played well by the counter team all made possible by going overload on auto duties, that is one of the CM's as well as the full backs encouraged to rush forward ASAP in this case (on defend again a side won't expose itself): http://www.youtube.com/watch?v=9ZxmV8niQWI Naturally having an opposition that has players of calibre "helps" in seeing the fallout of this. You haven't said what your setup was like, and what team you actually picked for your short tests. Reloading this match a couple times, Zenith created multiple opportunities that FM classes as CCC in similar ways each match, one match ending in a wild 2-7 defeat.

Take this as a quiz, if you want. :-) Can you assign the TC concept (philosophy, strategy, shout, formation, etc.) to the question it addresses? Thanks to wwfan, I think this should come with the manual, btw., perhaps a bit more elaborate.

1: Do you think a formation should be a) heavier at the back, b) balanced, or c) heavier up front?

2: Do you think a team should be more focused on a) scoring, b) not conceding or c) mixing risk and reward?

3: Should footballers should be a) specialists (i.e. focusing on a very specific task), b) generalists (i.e. doing a bit of everything) or c) a mix or both?

4: Which positions do you want moving between the lines in a) defence, b) midfield and c) attack?

5: How will you cover for this movement?

6: Who will pick out these players when they move?

7: What is your team good and bad at when compared to the rest of the teams in the league (look at the team report)?

8: What shouts might make a difference to a) probing attacks, b) wing-focused attacks, c) keeping the ball, d) frustrating the opposition, e) playing on the break, f) seeing out a tight match, g) going for broke, h) playing on a quagmire in a tempest, i) playing on an icerink, j) playing on a dry pitch in a scorcher, k) playing on a good pitch in wet weather?

What I absolutely do know is that approaching things the way you are doing only leads to frustration rather than enjoyment, however.

There is a reason people are doing all this experimentation with ME. FM13 has forced us to watch and micromanage matches without giving us the tools for figuring out how to do it. Some of you have figured it out long back but many players haven't.

Here's my last few results in current Blackburn save. I drew against Norwich and Wigan (matches i was supposed to win) but won against Liverpool and Tottenham (matches i was supposed to lose). How else are we supposed to figure out the why ? The wins are equally frustrating as the draws since i have no idea if it was a strategy change or something that led to it.

Simple.

You have a tactic that's good at attacking teams that are attacking you, but you can't break down teams that set out to defend against you.

Well it is as much a simulation as an illusion, just like everything else in life.

Simple.

You have a tactic that's good at attacking teams that are attacking you, but you can't break down teams that set out to defend against you.

Team talks and motivation could also be influencing this. If he is using the "no pressure + you have faith" approach that is the likely outcome.

It's such a flawed test that any results from it will be meaningless. The assumption of the OP is that strategy is the central component of a tactic. This ignores player roles, player duties, shouts, OIs, and the AI managers dynamic changes, just within the tactical element of the game. It further ignores all the motivation, morale and training modules, although they'll be less relevant to a repeat play test.

The best tacticians employ a holistic approach to tactics. They focus on the overall shape, the interrelationship between roles, and the method of attacking/transitioning/defending (employed via the shouts). Strategy can be a key adaption once this is achieved. Cleon, for example, employs strategy shifts to manage a match dynamically, choosing to go up and down strategies depending on the match situation. In contrast, I barely touch strategy at all, preferring to manage the match situation through employing shout combos.

The one thing that neither of us does is change between strategies without being happy with our overall shape and tactical approach, and we certainly never do it randomly. Not surprisingly, random switching produces a random set of results. Which the OP has pretty much proved, although, as playmaker points out, he has also shown the influence strategy has on goal spreads.

As others have pointed out, the OP's theory that FM's ME is a top down simulator is also wrong. The ME simulates play in 1/8th of a second calculations, from the first to the last kick of the game. The match calculates before you see it so the ME "knows" where the highlights are. However, that doesn't mean you can't change things. As in real football, during breaks in play, managerial instructions can be transmitted to the players, from which point the match is recalculated (illustrated by the text "making tactical changes").

All the OP has illustrated is that if you only focus on one or two aspects of tactics, in his case strategy and philosophy, you won't dramatically influence results. If you are holistic and make logical and dynamic decisions pertaining to strategy, philosophy, adjustments, roles, duties, OIs and shouts you will overachieve tactically. If you pay attention to morale, motivation, training and logical player acquisition, you will overachieve across the board.

I just seriously doubt it is possible to have that Throw-in be an event with outcomes 1) player x - success, 2) player x - failure, 3) player y challenge, 4) player z challenge, opponent player u - success etc etc, then, one "tick" later 5) player x ball controll - success, 6) player x ball control - failure etc etc so that every statistics-influenced/ing action is calculated separately with an open-ended outcome until the ball leaves play again 50 000 such events later

Quite remarkably, that's exactly what happens. People rarely realise quite what a mind-boggingly complex and sophisticated piece of coding FM's ME actually is.

To say that tactics, etc, has nothing to do with it is wrong. You can tell just from basic scenarios. Two examples:

Your technically-inferior opponants are 1-0 up at half-time. You've been playing 4-4-2 while they have shut you out with 4-5-1. There are a number of options without even changing your formation. In my typical 4-4-2, I would have two wingers, a defensive-minded MC and an AP, strikers as a DLF and Poacher. Without altering formation, I'd make my AP a DLP, make my wingers into wide midfielders, and make one of the strikers a target man, the other DLF (if he isn't already). This might require a sub, of course. This all effectively takes out their game-stifling midfield and moves your former wingers into the channels, while the tgt man holds things up for them and the DLF.

If this doesn't garner results (it will usually mean more shots on target) switch formations but keep the same basic setup, maybe use an anchor man and make the wingers outright inside forwards. However it's done, I usually get a result.

Another scenario: you're 1-0 up with 10 mins to go and the opposition are getting a lot of shots off. Clearly, they've gone all out attack. I will usually drop to a 4-5-1 with defensive wingers, a supporting target man and an anchor man in the DMC position. I have found that being super-defensive means a nail-biting finish, often a 1-1 result, but keeping it balanced and standard, with "counter-attack" ticked and "retain possession" shouted, more often than not I'll ride out the last 10-12 minutes comfortably, around one in four times coming in 2-0 up.

These are very clear tactical changes, manouevres that work more often than they do not (I'm talking about 90% effective), mostly attained through my own trial and error and relate to on-field football in a rather simplistic sense. So while it isn't perfect, it is wrong to suggest tactics have no effect on the game or that it is COMPLETELY random.

Thanks Svenc and wwfan for your elaborate answers!

I'm impressed if it really is true that the matches are simulated from scratch, pass to pass. That is a great accomplishment! Although if they simulate an entire match in 1/8th of a second then I question the complexity of their algorithms a bit. Also somewhat disappointed though, that I still don't see the characteristics of each player shine through in their playing style that well. Yes, it is often reflected in their statistics over a season, but to me the players seem quite anonymous on field (except for guys with high flair doing some showing off). But of course the difference in playing style isn't that obvious in real life either...

But I still don't agree that my test in meaningless. The talk about having to consider so much else I don't buy. It is possible that the tactical changes do have effect, but that effect is in any case a lot less than what is to be expected, according to how the strategies are explained in the game. The only outside factor that I can agree could cause the results I got, assuming that overload and contain really are as extreme as they are explained, would be if the opposing team counters my overload with contain and my contain with going overload. And perhaps that is what happens? EDIT: But on the other hand I remember that the stats (shots, possession etc) were quite even in the games I played myself, so I'm not sure I would buy that scenario either.

Does any of you guys have an idea how to experimentally figure out how much impact tactical decisions really have?