SI needs to fix the mid-season collapse

[Mitja]

6 hours ago, looping said:

Mentality means risk and a long ball is a risky pass so players under Defensive mentality try to retain possession. 

Last weeks I've found myself playing higher mentalities against attacking teams (to hit on the break with fast transitions) and defensive mentalities against defensive teams (to drag them out patiently). My results are somewhat better but I have an strange feeling that I'm playing the ME and not a football game. Who knows.

to me long ball is the safest pass otherwise all teams would play like Barcelona in real life. but in FM like you said it's the opposite and indeed 'Defensive' mentalities produce much more similar tiki taka football than Attacking mentalities (the willingness to pass back ) add to it poor defending (pressing and positioning), compared to attacking part of the ME. and I defenetly don't agree that Defensive football is about keeping posession, modern (defensive) football is all about counter attacks. that's also the point I'm trying to make, that football which is produced by Contain and Defensive Mentalities is almost non existant in real life (as first plan match strategy, used maybe in last minutes of games only). and that such football is regulary used in FM unfortunatly and it holds back the potential of this game because some games between Att vs Att Mentalitiy look like real match. especially those that end with not many shots. 

[Svenc]

51 minutes ago, Mitja said:

to me long ball is the safest pass otherwise all teams would play like Barcelona in real life.

It's all about context, and there's several ways of playing defensively (which is part of the problem too). Aside of that, stating that the "longest ball is always the safest" is naturally difficult. If sides would just give away possession even when not pressured (which is the issue currently), the attacking team would just mop it up and keep coming. The dynamics those matches in FM currently produce are partly down to limitations, partly also due to those defending weaknesses, which I don't like too. Luckily the AI doesn't ever do this, but playing endlessly keepsieball in deep areas of the pitch has been very easy on lower mentalities for quite a while now and effective for seeing out games (you would still need to tweak that passing  -- those passes played aren't merely due to mentality, there is passing instructions too, and by default it's mixed/direct for any defender). Which also isn't only ME related, but also AI. AI never reacts to anything such much, just doing it's "thing" when it could put those players under added pressure by a few switches, even despite the issues. The day the game will see a truly reactive AI is the day this forum will turn into a madhouse of terror, rage quit and infractions though [and that is just the tactics forums -- General Discussions will be something else].

What *is* "modern" football (aside of the more obvious, such as the ridiculous amounts of money, the annoying commercialism, Miles not shaving for months as Watford FC have fixed their mid-season collapse, Football Leaks)? Italy at the Euros kept the ball in between their "ball playing back three" "endlessly", patiently waiting for an opening up front, which you can't replicate on FM either, as the ball has always been played out of the back fairly quickly despite all sides apparently playing like Stokelona. Or as Rehhakles said it when his man-marking side went all the way a few years earlier: "modern is what wins you games".

[looping]

4 hours ago, Mitja said:

to me long ball is the safest pass

When I started to try to understand fm tactics, the first thing that shocked me is the concept of a risky pass. In my opinión, risky pass is a pass that can cause you defensive problems. Risk free=simplify procedures--->send long ball. That was my idea: safe pass=long ball (The worst is going to happen is the opposition winning the ball back in their half. Nothing very bad in terms of risk, understanding risk as something very bad can happen if you do something)

 

 

 

 

 

[SuperChaos]

The way I understand risky passes is, risky pass = through balls (more penetration). Less risky passes = pass to feet (less penetration).

Mentality should affect this with defensive mentalities encouraging pass to feet or if not available clearance from danger and attacking mentalities encouraging pass into space for more penetration.

 

[HUNT3R]

The PI/TI Risky passes/passes into space is about playing passes into space/through balls, so giving the option of not playing it to feet.

When we look at Mentality and the risk of passes, it has nothing to do with long or short balls, passes into space or to feet. None of that. It's simply about risk. "If I play this pass, what's the chance of it being successful?" A player in clear open space 30m away with no one in between is a safer pass than a player 10m away who is being marked.

Looking at that, it's still possible to be quite attacking with a Defend/Counter Mentality. It's also possible to just be defensive with such a Mentality. There are a lot of ways to set up

[Mitja]

2 hours ago, SuperChaos said:

The way I understand risky passes is, risky pass = through balls (more penetration). Less risky passes = pass to feet (less penetration).

Mentality should affect this with defensive mentalities encouraging pass to feet or if not available clearance from danger and attacking mentalities encouraging pass into space for more penetration.

 

 

yeah but isn't mentality already the sum of all other team and player instructions? so will a player try more risky passes with this instruction ticked on defensive mentality than without it on attacking? and creative freedom? isn't get stuck in already active on attacking mentality, just like dribbble more etc? isn't willingness to pass back already defined by your passing style, shape, creative freedom, tempo and player instructions, why the need for mentaility to define risk even further?

[Mitja]

13 hours ago, Svenc said:

It's all about context, and there's several ways of playing defensively (which is part of the problem too). Aside of that, stating that the "longest ball is always the safest" is naturally difficult. If sides would just give away possession even when not pressured (which is the issue currently), the attacking team would just mop it up and keep coming. The dynamics those matches in FM currently produce are partly down to limitations, partly also due to those defending weaknesses, which I don't like too. Luckily the AI doesn't ever do this, but playing endlessly keepsieball in deep areas of the pitch has been very easy on lower mentalities for quite a while now and effective for seeing out games (you would still need to tweak that passing  -- those passes played aren't merely due to mentality, there is passing instructions too, and by default it's mixed/direct for any defender). Which also isn't only ME related, but also AI. AI never reacts to anything such much, just doing it's "thing" when it could put those players under added pressure by a few switches, even despite the issues. The day the game will see a truly reactive AI is the day this forum will turn into a madhouse of terror, rage quit and infractions though [and that is just the tactics forums -- General Discussions will be something else].

What *is* "modern" football (aside of the more obvious, such as the ridiculous amounts of money, the annoying commercialism, Miles not shaving for months as Watford FC have fixed their mid-season collapse, Football Leaks)? Italy at the Euros kept the ball in between their "ball playing back three" "endlessly", patiently waiting for an opening up front, which you can't replicate on FM either, as the ball has always been played out of the back fairly quickly despite all sides apparently playing like Stokelona. Or as Rehhakles said it when his man-marking side went all the way a few years earlier: "modern is what wins you games".

what I would try to do to fix this loooong term issue is to improve defending. particullary a bug where players don't track/stay close enough to opponent when pressing but instead they simply cover/run back to the default position when his oppenent gets rid of the ball. that way, it's the defending team that creates space and time for the opposition and hence the posessional superority of more defensive setup.   marking should be much much more tight especially in own half.  the MCd being perfect example of such behaviour against teams that don't use AMC. completly defensivly useless in 442 vs 442. and all AI will employ MCd all the time. imo MCd and MCs/a defensive partership is really poorly implemented currently, the MCd droping deep breaks your shape and solid defense even further, he offers almost nothing in nowhere land if you ask me. maybe higher mentality tactics should get him in line with MCs when defending?! 

agree about centre backs being too uninvolved in build up too. especially when playing on more agressive mentality. willingness to pass back is awful on attacking, even on more fluid..it shouldn't be so rigid.

[MBarbaric]

Most tactical problems in the game (risky passes as well) are a result of universality that is dominant in the game. In real football, we have two fundamental splits that help organize, train... the team.

First, defensive/offensive phase. Players have different roles and duties when team has the ball and when the opposition has the ball. Even team shape in 90% of cases is different.

The second important difference between real football and the game is the pitch is split into zones. As a minimum, the pitch is split into defensive, middle and attacking third. Again, players have different tasks in each zone for each phase of play (with/without the ball). 

To get back to risky passes (and lots of other instructions in the game), risky pass in defensive third is a short pass towards a marked teammate. So, hoofing a ball forward is a much safer option. In attacking third, however, a risky pass is a vertical ball that splits the defensive unit and gets the striker in shooting chance.

It is much riskier to lose the ball close to your own goal than to lose the ball in attacking third. However, the game doesn't recognize this like it doesn't recognize that closing down is differently applied in attacking/middle/defensive third. It doesn't recognize (to an extent) the teams have different shapes in offensive/defensive phases and different roles...

A general problem with the game is that its instructions (and risky passes are representing this really well) are defaulting all over the pitch instead of being used in different zones of the pitch. 

[sporadicsmiles]

14 hours ago, Mitja said:

to me long ball is the safest pass otherwise all teams would play like Barcelona in real life. but in FM like you said it's the opposite and indeed 'Defensive' mentalities produce much more similar tiki taka football than Attacking mentalities (the willingness to pass back ) add to it poor defending (pressing and positioning), compared to attacking part of the ME. and I defenetly don't agree that Defensive football is about keeping posession, modern (defensive) football is all about counter attacks. that's also the point I'm trying to make, that football which is produced by Contain and Defensive Mentalities is almost non existant in real life (as first plan match strategy, used maybe in last minutes of games only). and that such football is regulary used in FM unfortunatly and it holds back the potential of this game because some games between Att vs Att Mentalitiy look like real match. especially those that end with not many shots. 

You have to remember that tiki taka was an inherently defensive strategy at it's core; the whole idea behind it was "if we have the ball, the opposition does not have the ball, and if they opposition does not have the ball, then the opposition cannot score". It worked so well for Barcelona and Spain because they had such technically proficient players that they could play this kind of keep ball at a frenetic pace, and with excellent quality of passing and incision when required. This is the main reason no other team has been able to copy this style of football so well. It is easy to keep the ball, but you have to be able to do something with it as well. Take Martinez, his teams kept the ball but their possession was sterile. It was tiki taka, but without the incision. Defensive football, which is about taking fewer risks, be that by being compact to deny space to the other team (at the expense of using the space the opposition gives you with as much ease), shorted passing (so you are not giving the ball away all the time) and cannot clear your lines). 

There are plenty of games where I see the AI parking the bus, clearly not having much intent than to draw my side in and try to hit us with long balls to their striker or strikers. I rarely see an AI team using a defensive short passing strategy, and if they try it then I press the holy hell out of them in their own third. That a human user can do this is perhaps the problem, but it still suffers from what I talked about above; you tend to lack incision unless you specifically look to give yourself that. If, for example, you play defensive but with a high D line and many players with creative freedom and attacking intent up front, then you are not aping defensive sides, you are more aping Barcelona. 

41 minutes ago, MBarbaric said:

First, defensive/offensive phase. Players have different roles and duties when team has the ball and when the opposition has the ball. Even team shape in 90% of cases is different.

 

This is somehow reproduced in FM. My offensive and defensive team shapes are not the same with the current tactic I use. Nor are they the same when in transition. The fact that I can change this simply by changing one role within my team further illustrates this point. 

 

44 minutes ago, MBarbaric said:

To get back to risky passes (and lots of other instructions in the game), risky pass in defensive third is a short pass towards a marked teammate. So, hoofing a ball forward is a much safer option. In attacking third, however, a risky pass is a vertical ball that splits the defensive unit and gets the striker in shooting chance.

 

I was under the impression that certain roles in the game use risky passes less than others by default. For example, CBs have a higher risky pass "threshold" than CMs, and will not try anything particularly fancy. I know for sure that a BWM(D) in central midfield will play short simple passes that are risk free almost always when in attack. To the point where this is detrimental if you actually want to try to unlock a defence. I could be wrong on this, but experience seems to tell me not. I usually find when my players are doing weird stuff with the ball in defence then it is my own fault. Perhaps my keeper is distributing the ball short against a team that is pressing hard. Perhaps I have "play out of defence" set under similar conditions. One thing I do see is players dawdling on the ball, especially the players who are not supposed to be creative, which I guess may be down to their options all being relatively risky and they do not know what to do. 

The one thing I definitely agree with from all of this thread is that the documentation should be much, much better. What I would really like are diagramatic explanations of what roles do. Show me arrows that indicate their likely movement (not on the tactic screen, but the information screen), show me arrows that indicate their likely pass length and risk (colour code it green for simple, red for risky). Show me how this changes when I change their duty. Show me how this changes when I change the shape and fluidity. The TI screen is so much nicer now it is visual and changes with respect your strategy, so extend this to the individual roles, because the explanations are not really adequate. Finally, give me the ability to compare to roles side by side so I can decide how they compare. This is not meant to replace watching your matches to see how the roles play and most importantly interact with other roles in the team, but visually it would provide so much more information and intuitive understanding, I think.

[Mitja]

MBarbaric

completly agree with you about real life but I'm not sure that the game needs such system which would further complicate tactical setup. on the other hand better and more realistic Defending would improve many aspects of the game automatically. compared to it's attacking part Defending still has plenty of room for improvement and tbh pressing particullary is not implemented really well. 

[Mitja]

12 minutes ago, sporadicsmiles said:

You have to remember that tiki taka was an inherently defensive strategy at it's core; the whole idea behind it was "if we have the ball, the opposition does not have the ball, and if they opposition does not have the ball, then the opposition cannot score". It worked so well for Barcelona and Spain because they had such technically proficient players that they could play this kind of keep ball at a frenetic pace, and with excellent quality of passing and incision when required. This is the main reason no other team has been able to copy this style of football so well. It is easy to keep the ball, but you have to be able to do something with it as well. Take Martinez, his teams kept the ball but their possession was sterile. It was tiki taka, but without the incision. Defensive football, which is about taking fewer risks, be that by being compact to deny space to the other team (at the expense of using the space the opposition gives you with as much ease), shorted passing (so you are not giving the ball away all the time) and cannot clear your lines). 

There are plenty of games where I see the AI parking the bus, clearly not having much intent than to draw my side in and try to hit us with long balls to their striker or strikers. I rarely see an AI team using a defensive short passing strategy, and if they try it then I press the holy hell out of them in their own third. That a human user can do this is perhaps the problem, but it still suffers from what I talked about above; you tend to lack incision unless you specifically look to give yourself that. If, for example, you play defensive but with a high D line and many players with creative freedom and attacking intent up front, then you are not aping defensive sides, you are more aping Barcelona. 

 

I don't think the Attacking football means anything for pro managers, it is just our or media perception of how teams play ''in general''. Barca/Spain being the perfect example of it. maybe it would be more precise to describe their footbal without defensive/attacking, like without the ball they play very high dline, with pressing employed all over the pitch. with the ball they play quick, intelligent, posessional football with plenty of player movement around box, etc etc.

your point how nobody was able to replicate Barca is far more important, because in FM any team can play like Barca. you (or AI) just need to setup your tactics to employ such style which irl wouldn't be possible. and that's the problem in FM. personally I believe the easiest and best way to improve the game further is to improve defending, closing down particullary, the 'run back to default defensive position when opponent gets rid of the ball, instead of chasing this player further and mark him more closer'. and that's where the suffician player attributes would kick in. currently all teams can play tiki taka. irl almost none..  

[Mitja]

run back to default defensive position when opponent gets rid of the ball, instead of chasing this player further and mark him closer.

this is also the major reason why Containment defending is much easier than Pressing. because in your own half players are more confident to mark opponent tightly and chase them further than to defend in such manner all over the pitch where it is of great importance to have players who know when to recover to default def. position and when to press further. and physical ability to employ such defence of course. 

[Svenc]

What all this "mentality" thing is ultimately about was part of my concerns mailed to SI. On a mechanically level, it all makes completely sense, and once you've toyed with it a bit, it's great, also from simulating a virtual football match. At the higher ends, the previous "instructions" on this were actually officially labeled from "defensive" to "all out attack", you'll also have loads more forward passes, as that whether you label that "risk", "mentality", or whatever, you would need something in there that governs the overall adventurous in player decision making.... atop of anything else. [It is quite curious that Beautiful Game Studios had pretty much aped this in their Champ Man Successors, then again they probably didn't want to alienate by breaking the franchises core values sources of potentially, initially confusion ].

However, on to translating that all into a coherent football whole, that had caused a few of grief prior already. If you dig through those, you'll find that even player who initially repeated the same matches a hundred times plus to be able to micro-tweak everything down to the T (not sure SI ever wanted a majority to be able to do this, as at such the AI is hopeless), were corrected some by main coders. What all of this boils down to is that from my view, and I may be wrong, the game seems to have taken steps to move back into that direction again. Even Rashidi previously for instance didn't give much of a hoot about "team shape", er, "Philosophy", er "Fluidity" before, and that is a player off the  "initially repeat the same match a 50 times" select kind who is able to micro the hell out of this, up to the point that AI wouldn't have a chance even if his side showed up with players not up to scratch, completely tired, drunk after a night-out with the Wags and Ferraris, and as cocky-confident as Donald "Here's King Of The World, You Pushover Suckers", Trump (to slightly exaggerate). Which may/or may not tell you something.

This goes straight to the feedback, including a re-introducing a "mentality bar" as a visual cue, and with it hints at an instruction/under the hood mechanic that a) newcomers can't possibly be aware of. And b) well, see the threads above. I have a few issues aside of reintroducing those, but that's why I mailed to SI myself. The thing this boils down to is this. The easier people have it to link everything directly to football, rather than mechanics, the easier they can get to grips with it. The easier they can go through this step by step and estimate, yeah... that's how that side my play, the less likely they may be confused. Whilst all of football is subject to subjectivity, the easier they can break it down to straight questions, the more likely you'll have a winner in the long-run. If SI can't do that themselves, and a few of the tactical niche were always a step ahead in some ways, the less likely they are also to code more pronounced AI management.

[fmFutbolManager]

What about those who don't have a mid-season slump?

[looping]

3 hours ago, fmFutbolManager said:

What about those who don't have a mid-season slump?

Show me how you do that!

[fmFutbolManager]

5 hours ago, looping said:

Show me how you do that!

Not really possible to show you how I did it. I just watch games in comprehensive mode and adapt to what is happening on the pitch. Moreover, you can use the analysis to view where teams are exploiting your system. I had a period over Christmas where we drew 2 games 0-0 and lost 1-0 and also started conceding silly chances. So, I looked at what was happening and for me, the DM(de) in front of my back four was getting too carried away due to the hard coded instruction to close down more. This was pulling him away from shielding the central defence so swapped him to an anchor man who is more disciplined. This provided a shield which saw my keeper break all clean sheet records. When looking at the attack which had stopped firing, we were taking too many long-shots as we were centrally orientated. Teams were dropping back and being compact. So I moved my wing-backs up a notch, instructed them to cross from the byline and instructed my outer two CMs (I play a 4-1-3-2) to run into the channels. This stretched the opposition and created more clear chances and started turning the draws and close results into comfortable wins.

Bearing in mind I'm now managing Barcelona, so I have an edge in most games anyway.

If I could offer advice I'd suggest using the analysis in the game, look at where the opposition focuses their attacks, where their assists are coming from and where their received passes are over the pitch and look to tighten up there, through role and duty changes. If needed, pull people deeper to offer more cover. As for going forwards, look at where your shots are, where the support is lacking and change the players around your strikers to provide better changes into the penalty area.

That said, this is all rather generic because I know nothing of your tactic or your predicament. 

[FatRonaldo]

9 hours ago, Analog said:

 

It says link is unavailable, does he changed it to patrons only?

[Svenc]

4 hours ago, Analog said:

I don't think so.  I can hit play directly from your quote and watch it.

Same, and everybody should! In particular the "Bad Shouts" part of the series for Rash essentially crapping all over the annoying "Morale / Team Talk / Body Language Manager" myths that pop up every single year. In a sense that's naturally also being guilty of approaching this from a "reductionist" angle of thought: If I keep this up, the other stuff doesn't matter yonk anyhow. Crucially though, in focusing primarily on match management (or "how to beat the AI" in this area), it is the factor that is by far the most important, rather than anything that is -- and better should -- be "tertiary factors". Matches aren't settled in the dressing room or anywhere. They are on the pitch, though SI have revealed that both is linked, as what happens beforehand funnels into player attributes some, and thus is a factor (some more: match preparation not being cosmetical either, e.g. set piece preparation boosting set piece related attributes slightly during such, morale rubbing slightly off on decision related attibutes etc).

Still, if your midfield has a hole in there you could drive a truck through, your primary playmaker is Makelele, what does it matter if the guys are happy and enjoying their crap football? Despite setting priorities straight: Still highlights why, despite this, such may never be recommended / endorsed as an overall guide how to approach Football Management in Football Manager, and the guys coding all that man management and press stuff may cast their veto simply by the virtue of them spending years on coding it all into the game.  This will never, ever happen, but if Rashidi were to ever face AI managers that were on his level in his area of forte, he'd likely find that whilst all this stuff doesn't settle games,  it may gain him just that small edge on occasion in what are much closer encounters of the FM kind, as his current understanding of match management on that micro level is that infinitely superior to AI. (SI may correct where it's me stating wrongs).

Speaking about the WDL runs that all lead to the creation of this thread (the OP still hasn't brought much further info to the table). A few years back I read an article about how Maths guys broke football down to a simple Maths model, and simulated what a league table would look like when all teams were of equally strength (which naturally on FM levels also includes managers that see eye to eye, rather than a mismatch....... perfectly possible be worse and better than AI managers on this, see above). I then set out to test this on FM. Went into the editor back then and cloned Bayern Munich. Player by player. Their managers. Further reduced the injury proneness to zero to not further randomize, and let it run on full details. Beforehand: What those Maths guys found was replicated, even though FM's ME is neither that same model nor football.
 

 

The virtually only discernible big difference between this and a normal league table, where teams of much different ability compete, was to be found in the goal differences, as there wouldn't be much of any thrashings, naturally. Other than that? For as long as the game, same as football will on average have low scorelines, i.e. matches tended to be settled in select fey key moments during the 90 minutes, WDL runs will transpire in and on itself. Even if all sides were level, and all matches close, there will be serial losers and winners.

 

 

 

 

Part of the OPs job is to a) judge each match in isolation as it is individual matches making runs. B)  Don'T fall for the scapegoat excuses that such runs were inevitable because of morale/team/talk/What I just had for dinner/Reload caused the game to punish me (check Rashidi out!).  C) Be able to tell what individual result was lucky/lucky/to be expected roughly, to get a better overview, and check to adjust. It is unlikely he would do such at this rate, and admittedly, imo a few of the game's feedback is quite flawed on that, including that feedback you get on individual matches. Plus the game were better off it if introduced a few further rankings than just points and goal differences (and how to read them).

[Goosewinkle]

15 hours ago, Svenc said:

Part of the OPs job is to a) judge each match in isolation as it is individual matches making runs. B)  Don'T fall for the scapegoat excuses that such runs were inevitable because of morale/team/talk/What I just had for dinner/Reload caused the game to punish me (check Rashidi out!).  C) Be able to tell what individual result was lucky/lucky/to be expected roughly, to get a better overview, and check to adjust.

Well this thread got a bit off my original topic, which is fine.  But I wasn't looking for help on getting out of the slump.  I was posting about how routine this is:  AI plays silly at season start and loses constantly -> I get a false sense of my own superiority -> AI realizes it's being silly and switches -> I struggle for several games as I figure this out -> I pull back and have a normal rest of the season.  For New England, I eventually put in a DM and played Defensive mentality on almost every match.  I watched Rashidi's vids too.  They were useful.  The screenshot shows how the season turned out.

I have learned that what I see is not as widespread among people here as I expected.  @fmFutbolManager doesn't see it (though maybe that's because the AI isn't going to underestimate Barcelona).  I dunno.  Maybe my season was just weird, and my tantrum is not be justified.  But New England is not a great team, and those first 7 games still look suspicious to me .  From about May onwards is probably a more normal New England performance.

 

[HUNT3R]

4 hours ago, Goosewinkle said:

 I was posting about how routine this is:  AI plays silly at season start and loses constantly -> I get a false sense of my own superiority -> AI realizes it's being silly and switches -> I struggle for several games as I figure this out -> I pull back and have a normal rest of the season.  

It's just as routine for users. They create either a poor tactic (I'm referring specifically to movement and use of space), relying on just having that space given to them by teams who feel superior or it's an overly attacking team which catches teams bu surprise. When that space they rely on so much disappears - struggle. Well, it doesn't dissappear... it's in a different area and because more players are kept back.

But you are right, the AI should set up better attacking tactics in general too. They are reacting more in matches and think a lot more than they used to, so this can only get better. Remember the days when every manager used a 442 with long arrowed fullbacks? That doesn't happen anymore. There's a lot more thinking going on now. Different formations, roles, mentalities etc.

[Svenc]

@HUNT3R Re: Movement, which is also if you wanted to keep it simple good advice to never mirror both flanks in their roles and duties to start with. If say you'd have a fb/s and a wm/a on both sides, your side would create the same patterns of movement and play on either. If stuff doesn't work, it doesn't work on either flank anymore. Same as employing the same roles, wide midfielders sit far narrower by default than a winger, and also aren't encouraged to run wide with the ball to stretch themselves, with lack of width oft being a constant in all of such tactical struggles in some way or other.

 

8 hours ago, Goosewinkle said:

Well this thread got a bit off my original topic, which is fine.  But I wasn't looking for help on getting out of the slump.  I was posting about how routine this is:  AI plays silly at season start and loses constantly -> I get a false sense of my own superiority -> AI realizes it's being silly and switches -> I struggle for several games as I figure this out -> I pull back and have a normal rest of the season.  For New England, I eventually put in a DM and played Defensive mentality on almost every match.  I watched Rashidi's vids too.  They were useful.  The screenshot shows how the season turned out.

The AI doesn't "Realize it's being silly". It targets difference outcomes against difference sides, which happens every week. It is based on a team's dynamic reputation (how big a team it is in the game), whether the match is home or away, a few on the team's more recent results (imo perhaps a tad too much), etc. Therefore, even if you won all those matches (which is a real short run anyway), not every team is now just going to park the bus. It depends on how big and successful they are themselves. 

A decent indication is in the match odds prior to match-day. It you're on long odds, expect the AI manager to target no less than a convincing victory against you. If it considers itself BIG FAVORITE, it expects no less than a win. However, as it also switches stuff DURING matches, how things turn out also depends on the RUN OF GOALS. If you take the lead in such a match, in particular away, you may find that they become even more aggressive. If vice versa happens, it may switch to something slightly more cautious to consolidate and not risk a counter or not give the ball as quickly away. Conversely, if you are the big favorite, expect an opponent to become increasingly defensive the longer the odds for it. If they still open the scoring, they may become even more cautious, and vice versa.  Which is also part of what usually impacts WDL runs... opponents either aim for WINs and DRAWs themselves, via "match managemen" and sometimes, may get them, and that may also include to throw all cautious to the wind in the remaining few minutes of a match, which can too either backfire or get that result. A bit like football.

You are currently assuming this is all caused because all those sides, played really high flying football. Unless you can't estimate whether thats true, you are guessing. The game has included more "in your face" feedback on that... If you keep the opposition formation widget open, it is quite explicit. Also even newbies look at the match stats to assess which results may be justified or lucky. (I recommend a different method, which is going into the analysis, click in team->shots, and check a) how many shots are from within reasonable range, and b) actually created from decent play, rather than from a free kick, corner, etc. lumped into a packed box). At such low scorelines, this is guessing. You may not have lost all, but a few of those likely could have gone totally either way. That's one way to go about it. The other was to not worry and accepting it as being a result of your lack of match and probably man management ability until you have build a superior squad and can overcome the streaks that-a-ways. After all, in some ways, not every manager in real football excels at all areas of his job too!

[MHovel]

On 7/12/2017 at 08:42, Analog said:

I don't think so.  I can hit play directly from your quote and watch it.

It doesn't work for me either.

I cannot find his video in his channel.

 

[MHovel]

I manage to get it work using proxi.

I guess the video is blocked in specific countries might be due to copyright issues.

[looping]

This is the AI:

 

Again the AI

 

 

And finally me:

 

Ok. Then the conclusión is collapsing happens not only to humans. I have no understanding about coding and I totally suck as a fm, but.. I suspect there is something that triggers this and is out of control. Those snowballs of results (positive or negative) are not normal, even if they can happen IRL, not that usual (every single save in my case) and so extreme. This is not how I expect results to be (nor my team nor any team).

I can totally suck and be a complete disaster as a manager, but it also happens to the AI. Don't you think is a bit strange?

 

 

 

[Cougar2010]

7 minutes ago, looping said:

I can totally suck and be a complete disaster as a manager, but it also happens to the AI. Don't you think is a bit strange?

No because its normal to have spells of good form & bad form.  Every football team in the world at any level has them.

Its the manager's job to minimise the length of bad form and lengthen the spells of good form.

[looping]

13 minutes ago, Cougar2010 said:

No because its normal to have spells of good form & bad form.  Every football team in the world at any level has them.

Its the manager's job to minimise the length of bad form and lengthen the spells of good form.

I agree.

What I'm trying to say is, don't you think are too large (good or bad)? Don't you think something triggers a good/bad run and a snowball is created being very difficult to turn things out?

Whem I'm on the winning streak I can play almost any tactic and any players and I'll win. I really think there is something wrong with this, I feel losing/winning streaks last too long and are too easy to mantain/difficult to turn around. It's not normal a promotion candidate loses 5 consecutives matches or wins 1 in 4 months and later becomes unbeatable. Neither is a relegation candidate to become unbeatable. I feel it's all too long, too often... I see teams too attached to the trend of results.

I've watched matchesof the AI when it is on a losing streak and it experiences the same tan me (human). Strange goals, weird decisión by players... I feel it out of control too influenced by the trend of results.

[Guest]

Your collapses have been discussed at length in the tactics forum, unfortunately you are unable to cope when things don't go your way & then start digging yourself into a tactical hole.

As for the AI manager they are also quite limited in their capabilities & can suffer because of that.

[FootballTrainer]

It´s true that often you start seasons by winning a couple of matches to get build of full of confidence just for midway through midseason start crashing. And since most forum users are certainly known for telling us beginners that the AI doesn´t learn your tactics, then what is going wrong midway through the season then? Of course you always have the random results where for some very unexplainable reasons you are leaking in goals or not scoring goals even though you have 18 shots and they have 2.

And the basic argument that it happens IRL too, well yes of course it does. But IRL there is a big maintenance for managers to sit and watch video analysis of former matches which leads us to the introduction of analysts in their roles. Wouldn´t it be cool if you could analyze opposition players on video conferences just in the same way as you could analyze previous matches. Therefore in the middle of all that you would be given tools to combat why your side are dropping in form. But that of course would also presume that people actually want to actually watch the matches on the below bar match engine that is now available (and most obviously always the biggest feature to be promoted for changes for each new release) whilst game play just stands still. 

[forameuss]

1 hour ago, looping said:

I agree.

What I'm trying to say is, don't you think are too large (good or bad)? Don't you think something triggers a good/bad run and a snowball is created being very difficult to turn things out?

Whem I'm on the winning streak I can play almost any tactic and any players and I'll win. I really think there is something wrong with this, I feel losing/winning streaks last too long and are too easy to mantain/difficult to turn around. It's not normal a promotion candidate loses 5 consecutives matches or wins 1 in 4 months and later becomes unbeatable. Neither is a relegation candidate to become unbeatable. I feel it's all too long, too often... I see teams too attached to the trend of results.

I've watched matchesof the AI when it is on a losing streak and it experiences the same tan me (human). Strange goals, weird decisión by players... I feel it out of control too influenced by the trend of results.

Go and screenshot every team in a nation.  If the majority are doing it, then maybe (I stress that word) you have a point.  Giving two examples is hardly enough.

[looping]

1 hour ago, Barside said:

Your collapses have been discussed at length in the tactics forum, unfortunately you are unable to cope when things don't go your way & then start digging yourself into a tactical hole.

As for the AI manager they are also quite limited in their capabilities & can suffer because of that.

No idea about the AI, sure I'm quite limited in my capabilities, but don't you think it's a bit exaggerated?

I mean, I even got promoted so the final picture is ok (and I say ok because my team underpeformed overall and showed a terrible inconsistency) but, don't you see this a bit exaggerated? Even if my capabilities are severely limited... I have never seen a team that gets promoted after losing 10 of the last 14 matches (Girona last season 11 of last 33 points is the closest thing I remember, but that's the exception, not the rule). 

Sure are reasons for that, reasons that due to my quite/severely limited capabilities I don't understand but the overall result... is... a bit weird, don't you think? 

I may add in playoffs my team again played the same was doing before the collapse. I'm not asking for help, just stating this results are shocking assuming that the way I play the game is wrong because my capabilities are limited.

On top of that, the same happens to the AI. Well, perhaps the AI is also limited, but to that extreme? If that happens regularly (and certainly happens in my experience with the game), isn't showing that there is something wrong? Teams don't behave like that (rare cases), but this consistenly happens in my experience with the game not only to me but to the AI.

Runs are too good/bad, last too long and results are too much influenced by the trend of results. That's the point, even if my and AI's capabilities are limited.

 

[Mitja]

it's just bad representation of normal  events and of course it's weird if you score 3 from 100 shots and conceed 3 from 15. too many shots which can be achieved by human managers is a long term issue wit FM. poor defending imo.

[Svenc]

3 hours ago, Mitja said:

too many shots which can be achieved by human managers is a long term issue wit FM. poor defending imo.

If you add a poor in between he "many" and "shots", I agree -- and I have had access to the matches of he most notorious "ranters" on this, who where explicitly told what their issue was ten years ago (and didn't listen). E.g. solely stats focus, when none of the stats the game has reliably shows if a defense is actually consistently carved wide open (and on the other end, whether the defensive is set up this ****** that every of the few time an opposition goes forward it is in reasonably space). Recently went a bit into it all in this thread plus following posts.   I am concerned in particular as for set pieces. As far as I know, SI solely look at the percentages of how many goals are scored from set pieces, and if the percentages at up. What I suspect they don't look into is for instance how many set pieces on averages it takes to get a shot off, and how that compares to football. I too want that to stop. Actually, this is straight from the introduction of Tactical Theorems 09, the first bigger community effort by players who didn't merely look at numbers, but judged the play (many of which producing detailed second to second bug reports on match play).
 

 

By FM08, while still being congratulated by contributors, it had lost direction and was being overwhelmed by threads promising super-tactical solutions to the Match Engine. These ‘super-tactics’ were not just in fundamental opposition to TT&F, but often generated frustration and anger for users, as they tended to produce unrealistic matches in which possession and chances were dominated by the losing team. Despite this seeming domination, the numerous chances created were often very poor and easily defended by the AI team. Many users failed to recognise the half-chance nature of the attacks and some resorted to accusations of a random Match Engine, super-keepers and a cheating AI.

Whether the game is more streaky than real football, could well be on the level of general AI management (something I wasn't aware of: SI recently announced that managers under pressure can become prone to switching their "Prefered formations", to me that's a completely panic crap shoot if it wasn't for the fact that in their change they would look at the currently most successful teams -- then again, are they successful because of that formation or because they have better players, furthermore, what time frames are we talking given that runs happen all by themselves and always will be unless SI code average win margins of +3 +4 without any side completely making its defenses implode, which is doable mind). However I'd then compare this to actually football (first and second half season tables oft look completely different, as that is a completely arbitrary time frame already), plus Betting Outlets make good money on people betting on WDL streaks (there is a reason they "advertise" such on their sites). Whoscored.com is really good for that, it shows any such under the "progress" tab.   I had some of the same concerns on FM 2012 initially except in reverse -- due to a couple things once you kicked on a winning streak, it could last "forever"

Which was connected in some parts to how "morale" worked back then, plus imo also AI tactics [any AI that was a relegation contender on FM 2011/2012 was oft pretty hopeless, and the remnants of this arguably still remain in the fact that due to tactical decisions in particular on the more defensive AI end, relegation fodder sometimes takes an average of 15-20 shots to finally score for a long-term shot conversion of 5% ish, which is tactical. However, it's not only relegation fodder, any AI manager can become a bit more cautious after a bad run until that run is over. You can get results despite bad performances, and at FM's current level of stats, they are a questionably gauge for general performance. However positive results breed some confidence and vice versa. On lower levels, I would expect things to be more streaky than in most top leagues, as sides tend to be more level [which is connected to the relative spread of money, you're less likely to find a side that is able to spend multiple times the league average and vice versa]. Whether that transpires in-game on AI levels of management would be an interesting study.

[lemeuresnew]

10 hours ago, forameuss said:

Whem I'm on the winning streak I can play almost any tactic and any players and I'll win. I really think there is something wrong with this, I feel losing/winning streaks last too long and are too easy to mantain/difficult to turn around. It's not normal a promotion candidate loses 5 consecutives matches or wins 1 in 4 months and later becomes unbeatable. Neither is a relegation candidate to become unbeatable. I feel it's all too long, too often... I see teams too attached to the trend of results.

i agree. teams like Leicester City need to stop going on these sorts of runs all the time

what type of team are they? champions or relegation candidates? they just shouldnt be aloud to keep changing from one to the other every season, it just isnt realistic...

[Mitja]

7 hours ago, Svenc said:

If you add a poor in between he "many" and "shots", I agree -- and I have had access to the matches of he most notorious "ranters" on this, who where explicitly told what their issue was ten years ago (and didn't listen). E.g. solely stats focus, when none of the stats the game has reliably shows if a defense is actually consistently carved wide open (and on the other end, whether the defensive is set up this ****** that every of the few time an opposition goes forward it is in reasonably space). Recently went a bit into it all in this thread plus following posts.   I am concerned in particular as for set pieces. As far as I know, SI solely look at the percentages of how many goals are scored from set pieces, and if the percentages at up. What I suspect they don't look into is for instance how many set pieces on averages it takes to get a shot off, and how that compares to football. I too want that to stop. Actually, this is straight from the introduction of Tactical Theorems 09, the first bigger community effort by players who didn't merely look at numbers, but judged the play (many of which producing detailed second to second bug reports on match play).

not sure if I agree. there are just too many 5-2, 4-0, 5-0, 3-2 results and too few of 0-0 or 0-1s. both in AI vs AI and AI vs human games. my point was more about the fact it feels weird of not scoring after 20 or 30 shots. and I'm not saying this is impossible in real life but I can understnad the frustration and confusion of people. 

on the other hand in some rare cases where I guess everything fits together tactically , the ME is able to produce most amazing football, true midfield battle with plently of  misplaced passes, small errors, good defending and realistic score and shot counts..   

[HUNT3R]

15 hours ago, looping said:

I agree.

What I'm trying to say is, don't you think are too large (good or bad)? Don't you think something triggers a good/bad run and a snowball is created being very difficult to turn things out?

Whem I'm on the winning streak I can play almost any tactic and any players and I'll win. I really think there is something wrong with this, I feel losing/winning streaks last too long and are too easy to mantain/difficult to turn around. It's not normal a promotion candidate loses 5 consecutives matches or wins 1 in 4 months and later becomes unbeatable. Neither is a relegation candidate to become unbeatable. I feel it's all too long, too often... I see teams too attached to the trend of results.

I've watched matchesof the AI when it is on a losing streak and it experiences the same tan me (human). Strange goals, weird decisión by players... I feel it out of control too influenced by the trend of results.

You picked two random (or maybe not so random) teams.

There are still plenty of AI teams that do not suffer from a mid-season collapse. Some don't collapse at all. Some start poorly, others end poorly. There are plenty of human managers who also don't experience it.

[Svenc]

2 hours ago, Mitja said:

on the other hand in some rare cases where I guess everything fits together tactically , the ME is able to produce most amazing football, true midfield battle with plently of  misplaced passes, small errors, good defending and realistic score and shot counts..   

Some inherently niggles are always there... and they influence matches both own and AI (which is also a strong case for looking beyond the numbers). However I agree, whatever issue is there is oft exacerbated  by tactics. The problem is usually when somebody doesn't subscribe to the logics inherently meant to be rewarded by the game, which has happened back then, and still happens now. Whilst I think the "shot counts" should go down in such cases I also think that on a stats level it were worrysome if things would 100% replicate football with such approaches either. Which is related to defending -- it's then good enough to only allow for mainly really below average to poor attempts in no space, however not good enough to keep the shot count down, which is preventing those shots from hpapening. The fact that players in FM aren't a "physical presence" as such yet may play some part in this -- the "collision detection" you often hear talked about is really a "collision avoidance", that is unlike in earlier iterations the engine detects whether a player has another player in front of his in his off the ball runs, and then makes him go around him. You could of course make the case that it should be hard-coded so that you couldn't play experimental/crazy tactics, but I  think that would be a poor route as part of the fun of any game is being able to ask what would happen if.... question then is how the game responds to it.

That said, another unfortunate part in this is that AI opponents are limited in their ability to truly determine where they are currently try to be hit. In real football, management is more intelligent across the shop, whereas in-game in particular an AI manager given really aggressive traits would never "really" switch to something more cautious even if his aggressive d-line was exploited to hella back in a particular match (obviously bound to fluctuate some both depending on the Opposition's tactics, but also players, and the run of goals, as should such an aggressive AI concede, in tendency it goes even more aggressive in an attempt to get back into the game)... What this does also mean thus is that how "well" they are able to force an opposition to poor shots depends a good deal on simply research, rather than an AI man being able to spot: "Okay, this dude tries to overload the central areas of the pitch by cramming all his players there, let's switch to a back five involving a spare man to cover, plus a few additional CMs DMs he'll have more trouble squeezing his guys through).

[sporadicsmiles]

The more I read this thread, the more I understand why people get frustrated that they cannot be successful with the game. It seems people are more willing to blame everything except their own shortcomings for anything negative. You never see the reverse, success is always down to the user and never some weird behaviour of the AI or ME. 

[looping]

13 minutes ago, sporadicsmiles said:

The more I read this thread, the more I understand why people get frustrated that they cannot be successful with the game. It seems people are more willing to blame everything except their own shortcomings for anything negative. You never see the reverse, success is always down to the user and never some weird behaviour of the AI or ME. 

I agree success is always down to the user (even if there was a rubberband, which I deny). I didn't say my results are terrible because the AI or the ME are wrong. I just felt the game is too streaky but obviously, even if I'm right (and I admit I can certainly be mistaken) this is not the cause why anyone/me isn't successful

[Goosewinkle]

6 hours ago, sporadicsmiles said:

The more I read this thread, the more I understand why people get frustrated that they cannot be successful with the game. It seems people are more willing to blame everything except their own shortcomings for anything negative. You never see the reverse, success is always down to the user and never some weird behaviour of the AI or ME. 

Seems like this is the reverse right here:

On 7/22/2017 at 12:44, looping said:

Whem I'm on the winning streak I can play almost any tactic and any players and I'll win. I really think there is something wrong with this

 

[forameuss]

On 7/23/2017 at 06:30, lemeuresnew said:

i agree. teams like Leicester City need to stop going on these sorts of runs all the time

what type of team are they? champions or relegation candidates? they just shouldnt be aloud to keep changing from one to the other every season, it just isnt realistic...

How come that post you're quoting is referencing me?  Odd, definitely wasn't me.

[lemeuresnew]

5 minutes ago, forameuss said:

How come that post you're quoting is referencing me?  Odd, definitely wasn't me.

you can always tell a @looping post

no idea why it has you as the poster of it. does it make you feel used...

no idea how that happened

[Svenc]

1 hour ago, Goosewinkle said:

Seems like this is the reverse right here:

 

Yes, but looping has an extremely long and unprecedent history of closed threads in the tactics forums, which included wrongly corectting some of the better tactical mods, so any tactical observation has the be taken with a gigantic pinch of salt. Not taking a jab, just saying.

Whilst this is GD, any specific issue anybody has won't be revealed at such rates, which can also include AI management impacting. When googling for "set piece conversion",  I found this thread on the aforementioned FM 2012 for instance, where additionally an initially faulty AI set piece tactical routine plus a bug as to their match preparation caused them to be a bit over the shop themselves. Additionally I think only post that iteration it was discovered that AI central midfield setups were oft quite suspect, influencing further. This highlights why looking at purely results won't reveal much, and whilst it probably wasn't anybody's intent to look for solutions, it's easy to see how the guys at SI towers would end up filing this the same as any "FM is too random" thread in the past, of which there were loads, even with worse AI competition and engines that had different issues. Outside of improved AI assistants taking over anything, same as intelligent data assistants taking context in account (and occasionally probably pointing out: that's a bit like football) I don't see how anybody who suffers here ever since FM 2009 (!) would suddenly face completely different all by itself, personally.

In the meantime I'd be interested into FM studies like this, in particular numbers like the "distribution of shots" graphics. What also could be interesting how many matches on FM are settled by basic mistakes (or, as this is a computer code, bugs), as this could also be an indicator whether "luck" plays a larger role in the game world on the general level of AI management, but then you'd also need to find studies on football. Knowing the other "subjective" stats (HCs, CCCs...), the one for "mistakes" probably wouldn't be the best gauge for this though.

[lemeuresnew]

15 minutes ago, Svenc said:

In the meantime I'd be interested into FM studies like this, in particular numbers like the "distribution of shots" graphics. What also could be interesting how many matches on FM are settled by basic mistakes (or, as this is a computer code, bugs), as this could also be an indicator whether "luck" plays a larger role in the game world on the general level of AI management, but then you'd also need to find studies on football. Knowing the other "subjective" stats (HCs, CCCs...), the one for "mistakes" probably wouldn't be the best gauge for this though.

i obviously agree mistakes or 'bugs' are annoying. but i like the game, so try to think of it as a bit of luck. after all, they dont seem very often in my game (had the keeper drop the ball behind him twice in about 20 seasons maybe). and they can go both ways. you can see it as the game not being able to do things like stop a ball dead in a puddle or have the ball stay in off the corner flag

the more i know about different stats, the more frustrated i get about mistakes in the game. think a lot of people are the same. so i would probably avoid any 'official' posts like this

[Svenc]

33 minutes ago, lemeuresnew said:

the more i know about different stats, the more frustrated i get about mistakes in the game. think a lot of people are the same. so i would probably avoid any 'official' posts like this

Probably not useful in the game, but for balancing things out. Of course there's stuff in the game that is some closer to football, whilst others isn't.  In terms of development you'd probably need to build a hierarchy, and balance from the top as there is stuff that is more important and fundamental to the sports (say typically close winning margins and matches settled in key seconds), others perhaps not as much (average number of throw-ins per game). Which is of course purely on a stats level, which isn't really much the actually engine on a play level, which is the problem. So whilst "happens in football" responses can be a bit of a cop-out, outside of watching a couple poss and shot counts very few much delve that deep into football numbers to begin with too. Crazy streaks happen in football, plenty of, Malaga last season went with but one win in 16 matches, upon which they almost managed to win the next 7 in sequence -- Inter blew their solid season by almost losing 7 on the trot towards the end whilst hapless Crotone, 3 wins at that point, managed a streak of 9 matches with but one loss to avoid relegation, Braga being on course for CL qualification topping Sporting, only to blow it winning no more than 4 of the remaining 17 matches.

Bremen's run in spring may have made them look like Bayern in terms of results, however dissecting their matches you may have found that may not have been the case much (and banking on them to compete for European spots the next season may be a tad premature). There may or may be more of that in the game, but there's plenty stuff in football. The question is how and how often all of it comes about, the influence and impact of soggy pitches in spring on a short passing game, of key players missing, of players adapting and managers changing clubs are all for SI to model, same as how big it influences whether a side is full of confidence and vice versa. Another question is whether and how it all can be influenced. As history shows, some are a tad better than this, others not as much. And as for AI managers here... Bayern on prior releases on occasion losing up to 8 matches in the Bundesliga and sometimes struggling to approach 2 goals per league match, would never happen in football, but was easily topped by any newbie to this, so yes, AI management is always a factor on this, but oft also to be judged on a case by case thing as different managers are edited different traits, complicating things further.

[knap]

On ‎23‎/‎07‎/‎2017 at 11:16, Svenc said:

"Okay, this dude tries to overload the central areas of the pitch by cramming all his players there, let's switch to a back five involving a spare man to cover, plus a few additional CMs DMs he'll have more trouble squeezing his guys through).

AI Tactic

[Svenc]

1 hour ago, knap said:

AI Tactic

That's one knap (also fixing some of the centrally defensive issues in this version to a degree). The "issue" is, depending on how you look at it, that this is employed by random chance, e.g. the respective AI manager having it edited as a preferred formation. If you're lucky and your tactics was geared towards purely overloading the centre (in rather one-dimensional ways), you wouldn't face it once. If you're not, you may still not face it that oftenly except in that Champions League return leg which may contribute to your knock-out. What is also interesting in such cases if you arrange a string of friendlies against your reserves, and don't change anything in the staff setting, the reserve side will play the same starting tactic against you down to the duties given trying the same stuff you do. Considering that those are the reserves, that's some interesting matches... I think the "tactic testing" leagues you linked would be a lot more conclusive (and avoid individual frustration) if they didn't only take team reputation into account, but would also test for the more popular AI formations used, forcing every AI side to prefer the same formation. However each of those can be interpreted both defensively and more attackingly, out of a running save you're more likely to face this with a defensive AI, e.g. in tendency a team that considers you the big favorite (the duty setup in this one gives that away, not a single attack duty).

Re: Serie B may make for interesting case study in all of this not only because it is a lower league were teams oft tend to be some more balanced so that everybody can beat everybody anyway (depends on the research on this too), but because ever since a few seasons basically half he league is given a chance at promotion, with even the 8th placed having a shot at the end of the season to compete in Seria A in the follwing. Pescara were promoted the season prior despite a winless run of 9 (5 losses in row) followed up by 5 wins in a row. However this is looking purely at individual clubs, there may be better ways to determine whether too streaky or no. The Championship is oft a bit similar too, but the huge amounts of parachute payments after relegation from the cash cow that is the Premier League may oft skew that some [the final promotion play-off at the end of a Championship season is oft tagged as the most lucrative single match in club football for a reason].

[looping]

8 hours ago, Goosewinkle said:

Seems like this is the reverse right here:

 

All I said is:

1. I see WWWWWWWWWWW LLLLLLLLLLLL streaks.

2. This happens in real life too.

3. I felt in fm happens too often and too long (too easy to win when in a winning streak, too easy to lose when in a losing streak)

4. if that's the case, there is certainly something wrong

5. Even if there is something wrong, this is not the root cause of my bad results.

6. I even said the final picture was ok.

If you want to understand, nice. Otherwise, feel free to carry on if it makes your day better.

 

[knap]

Not sure why TFF set it up that way, unless it was to specifically highlight the effect of reputation and defensive tactics having the upper hand. This test league also uses a morale freeze so dips in form are more likely in reality than with this league.

This test league is set up to use managers preferred formation of the top clubs, I am not sure why you would want to test specific AI 1 formations, as a tactic has to be robust against many AI formations, and TFF has very specific and detailed reasons for setting up the test league in this manner. .

As all is working again you are welcome to take the test with your own tactic playing every game.

At the moment I'm playing LL with a flat 442 and keep playing narrow AI exploit formations!

 

[sporadicsmiles]

11 hours ago, Goosewinkle said:

Seems like this is the reverse right here:

 

As someone said, this is Looping, and I am also not too shocked he see's things this way. I have posted on a number of his threads over on the tactical forum, and that he does not know what he is doing wrong or right tactically really makes sense. This is also not a dig, because he tries to understand. In general on these forums though, you see people complaining about long losing streaks, but never long unbeaten runs. The former is the fault of the AI/game, the latter is down to good management. 

As always, it is much more complicated than that. I go through periods where my team plays superb and demolish sides, and then this can be followed by playing like a bunch of amateurs who have never seen each other. I assume it is complacency, or nerves, or just a team I come up against being in good form. Hell, I will always check to see what kind of a streak the opposition is on, because it gives me an indication of whether or not I expect to ease to a win or have more of a fight. Either way, winning or losing, the fault is mine. Be it that I set up something stupid tactically that the AI exploits, be it that I did not stamp out complacency, or that I put pressure on my team when I did not need to. Maybe I was in good form but then changed 10 starters to rotate in a busy period, screwing up their momentum (I am sure I read changing too many players at once can give you weird disjointed performances). 

Another thing to do is to pay a lot more attention to what is going on when your team is playing badly, or your form stinks. I can happily watch 5 minutes of a match and know I will be fine when in good form. Stick it on only commentary, and wait for the win. I never, ever would do that when I am playing badly. I want to watch and to know what is going wrong, where are we struggling, what is the opposition doing that we cannot handle. Then I can start to alter things slightly to minimize these things. I do not want to generalise everyone here, but I get the impression that a lot of people do not do that; for example they watch key highlights and only ever watch key highlights. Doing that, it is damned hard to understand what is going wrong. I also will spend time looking at post match analysis and even rewatching bits of matches to further understand (it is so much easier to analyse something when the result has happened, you are not involved emotionally anymore). I therefore tend not to have horrible losses of form very often, and when they do I put in a lot of effort to snap us out of them. This is the advantage we have over the AI; it is a computer, it runs on an algorithm and can extract information and respond in a set number of limited ways (no offence to the Devs, you do a great job). Humans can adapt in unexpected ways, spend much more time in analysis, and have so much more flexibility tactically.

[Svenc]

As for the streaks, if you have a bit of time, you could go here and try to compare, no source makes them as visually obvious as this.

 

17 hours ago, knap said:

This test league is set up to use managers preferred formation of the top clubs, I am not sure why you would want to test specific AI 1 formations, as a tactic has to be robust against many AI formations, and TFF has very specific and detailed reasons for setting up the test league in this manner.

Not to drag this OT: The overall point was the AI has limited capability. If it would be able to "think", it would intelligently adapt, rather than "going through the motions" of applying "preferred formations", with the general level of AI management possibly making AI sides slightly less consistent than they could be in general. As the general level of competition is AI, that should level some out probably. Plus, if the AI ever did, there would be a lot more of such threads, as it would make the game naturally far more challenging and streaks, in particular winless ones, some more likely for any player. But to answer this:

It's because tactics oft tested have a borderline "rock paper scissors" effect, in particular of the TFF kind as they lack in balance [attacking in very one-dimensional ways, e.g. overloading an extremely narrowed centre which some AI formations prevent better like the one you linked] and/or are leaving the defense exposed [the narrow strikerless narrow wing back type down the flanks, which is suspect in particular against aggressive opposition wingers]. Additionally, it's curious if you do a couple of friendlies against your own reserves with a TFF type tactic, in particular if your first eleven aren't world class, as the reserve staff keep throwing the same "super tactic" back at you. Don't think that will ever happen, AI managers copying your a "winning formula", but who knows.

 The averages are not in doubt. The tactics work great. Merely an observation to the "random" frustrations of people playing these, in particular in knock-out competitions, where unlike the league, every result counts. The error isn't so much the tactics or AI though, but the engine, as it always has holes (this year centrally, FM16 the flanks). Ideally AI wouldn't need to play such formations in defensive ways to not get steamrollered as often, even by significantly worse sides. Then again, without such holes, TFF wouldn't get those results despite significantly weak sides...