[Breaking News: Football Manager 2018 Headline Features]

German engineering, a suspiciously Feeder-soundalike tune plus a confirmation that the guys are some more involved responsible for one of the most criminally underappreciated and boldest games in the AAA space in recent years (despite its frustrating bits). Decent tease. The team dynamics looks like something that Fifa Manager had, though internally FM's internally always had it anyway in various shapes and forms.

[Football Manager 2018 Announcement]

17 minutes ago, El Payaso said:

Nowadays it feels that the point of view from SI is that they are counting the average sources of goals in football in general and the view is that for example every team should score one third from crosses, one third from through balls and so on. While this should be more down to style of plays of the teams and the attributes of the players that you have in your disposal. 

Aye, at the same time, stereotypes. The truth is that in football characteristics between players isn't that hugely pronounced, plus things aren't as clear cut. Free kicks are perhaps the best example as even the best teams score perhaps 5 of them at best. This is going to fluctuate on a season by season basis simply by random chance. You have fairly average guys outscoring specialists as Ronaldo too can go almost a season without converting, as the chance of conversion is rather low -- and even for that specialist the likelyhood increases a few percent. Likely, whilst Messi et all may be more consistent at a) getting into position b) working out chances for themselves and c) putting a few additional "difficult" finishes away, they also score consistently because they take like 5 attempts per match, whereas the average forward only 2-3.

 

In real football, teams spend millions for a few extra percent. And oft fail. Generally agree though. I expand it further to team playing styles. Which is partly engine limitation, partly AI managers. I think fairly everybody is able to get possession numbers of 60% going, whilst Pep and Co. ingame never do (even if you consider that FM's possession stat works by chessclock methods rather than number of passes -- the number would often be lower actually if FM worked by passes). PES and Fifa are more "fun" in that sense, however they also exaggerate hugely. That you control players yourself adds to that naturally. If you control Luca Toni, he feels like one of those tanks in Armed Assault. If you control Aubameyang it's like switching from tank to jet fighter. There's a dichotomy at work. It goes like this:

simulation-----------game----------arcade

The closer you are to the left, the less immediately pronounced things may going to be. I agree though FM "generalizes" much. For instance, individual finishing techniques are rare and limited, and it's been made a bit "worse" ever since "tries tricks" PPMs etc. were taken out (which seemed slightly glitched though).
 

 

[Football Manager 2018 Announcement]

^ A cookie to Miles if he's going to dress as Indy for his performance to announce he'd be selling these fine leather jackets. :-P

 

1 hour ago, El Payaso said:

Phew, was starting to panic a little bit as my FM playing career would surely have ended there.  Or I would just be using the commentary. 

Every professional analysis tools is build around 2d dots, which makes a lot of sense as tactical analysis is primarily about patterns and movement, rather than spotting whether a forward picked his nose during a corner. I don't think they'll ever exclude this, ever. The initial 2D view was also clearly influenced by any such tool used by professionals in analysis, plus they actually have a deal with Prozone.

[Football Manager 2018 Announcement]

2 hours ago, El Payaso said:

The only things I would ask for the ME is that there were more scenarios on how the chances are created (if you have a team like Crystal Palace then you're strong through crosses, if you have prime years' Lampard then you're threat from long range, if you're midget team like Barcelona you won't expect too many goals from floated crosses) and it would actually be hard for example for wingers and strikers to perform.

I agree. If it's targeting a sim though, it's easy to fall in stereo-types too though. For instance, if you go through the numbers here, you will find that even "midget" teams don't differ hugely in their amount of crosses (the pass type section), or long shots from range (the shot location section of that side), or anything. Purely approaching this from a "game" perspective things would be naturally be much more rewarding if there was a huge, immediately obvious difference between teams. Still was able to build quite a few specialst sides on past iterations.

By the way, a more acccessible screen like that would also be nice, just outlining a few strengths and weaknesses of teams in-game (the characteristics if you scroll down). Plus, as you highlighted in your thread, a few improvements to AI squad development. :-P

[Football Manager 2018 Announcement]

6 hours ago, forameuss said:

This forum will, on the other hand, be a carnival of idiocy with people all over the place.  People will get accused of being fanboys and others will get banned for "just speaking their mind".  Calls of SI rigging matches, suggesting they haven't tested the game.  Questioning their capabilities as developers.  The whole routine.

I suggest ignoring both for some improvements. I know it's always tempting to comment on classics like this (the comment on wingers being useless was targeted at FM 2016, by the way ).. (in general: as soon as something starts with: "I've been playing this since Cloughie ruled Europe , which means I am totally not a newbie to this...", run. But don't comment. The thing is, I actually think a game of FM's kind attracts and nurtures quite a few interesting mental traits itself. Look at the in-game clock, how many hours do you put into a save even a more casual one? It's a few. The game is this cruel it makes sure it reminds you of it. This is the kind of product where you tend to find: The more persistant somebody tends to be at directing all kinds of bile towards people, the more hours he puts into playing this regardless, and feeling miserable as a result (maybe that's the aim). So there's no point in arguing, I've tried it. I literally showed one guy in tiniest detail what the issue was, which was partly the game engine to blame, but also his input accentuating it (often the case), and I've been trolled by: "No; you must be on the pay-roll to whitewash!" since. [Whish I was, how much, SI?]

There's no point. It's an obsession and a mind game fairly quickly. However, the same goes the other way around. And if you think the nature of the game itself through (hours logged in, no direct control but only influence over everything), it makes a lot of sense. That the game is targeted at generally football fans, and that every football fan has deemed himself an expert since men kicked ball is a whole different topic. Don't comment, just post what you think is good and bad. Hopefully more of the former, less of the latter. Hopefully an improvement over previous too. I also whish that people would hold actual football as their yard stick though, and on that topic of "getting rigged" I am hopeful BBC's introduction of Expected Goals into their stats zone may make a bit of a difference. It's not a perfect measure, by any means. But clearly.... not every sitter is in any kind of way a sitter. And pundits don't get away with it as easily anymore for making it sound as such.

[Football Manager 2018 Announcement]

16 hours ago, forameuss said:

This place is going to be an absolute zoo come the features video as people invariably get their hopes far too high for them to ever be met.

And they haven't even teased the revolution yet. Any reasonable mind may roughly know what to expect from an annual release, it's been like how many years now? Since we're at predicting player feedback: Totally expecting a billion vitriolic rants about "getting FM'ed/SI'd/rigged" etcetera once the actual game hits too (in particular if a few defensive leaks are tweaked). All the while some actual issues (wide midfielder positioning ) aren't picked up by that same audience, plus the internet doing its magic trick of turning perfectly decent guys into something else completely. I've been around here long enough to remember when things used to be a tad different. But I can understand why staff would limit their communications to a minimum. And this place is pretty tidy, all things considering, no doubt because most of the guys on here seem top blokes even when sat in front of a keyboard. Most of the actual Steam™ seems to be vented elsewhere.

The aforementioned type always distracting from genuinelly stuff gets my goat a bit every time too though as a player, which is why I've opted to stay away of the usual threads this year when I get to try the new stuff (likely after some point the public demo hits). Which also means I may jump into this as a "virgin" untouched by "general forum consensus" going around. Can only imagine what this must be like on a developer's end. This twitter going around these days records some of it nicely though. I mean, criticism is one thing, I tend to think I rubbed into some wounds the game has myself -- as a typical costumer I always want more and better and speedyfastier for my cash, I'll let you know soon enough. But dang. Play nice folks.

[Football Manager 2018 Announcement]

Speedyfastier 2.0 (not that you will be able to second-guess what the heck that is supposed to be until the product page goes live).

[Football Manager 2017 17.3.0 Official Feedback Thread]

Better statistics, final match reports that make sense.

Seen Miles more recent "dismissing" the idea of there being more refined stats, such as Expected Goals. Guys suggested to him assessing shots individually was mainly done to predict future performance, which is one part but not all of it.
 

Quote

When a soccer player takes a shot, there are only two possible results. The ball goes in, and the team gets a goal, or it does not go in, and the team gets nothing. But some shots are better than others. The xG statistic awards credit for how good a shot was, regardless of whether it winds up in the net. For example, a shot from very near the goal might go in about 50 percent of the time, so it would be worth 0.5 of an xG point. A speculative shot from far out might be worth only 0.1 of a point.

FM on a statistical level is purely based on volumes. The same goes for match reports. If a side has lots of shots for no goals, it's pointed out as being "wasteful" in the post match feedback, even if those shots are meh throughout. Whilst this may successfully mimic all the bad football punditry in the world and Charles Reep would approve of it,  the engine clearly doesn't work like that. Finishing research seems be taken into account. How factors such as the shot location influence the likelihood of a goal. The pressure on the player. The pace. Or how central through balls typically see the keeper in a significant edge over the forward as the angle of the assist doesn't handicap him. All the while a crossed ball from wider position does exactly that. And yet the game stats act in such a way, with the subjective stats in there border lining on the fairly useless. You've never needed a single CCC on this to score 4, as the gap between your average CCC and your average SOT isn't that big. Unlike CCCs, ExpG are actually used in football, by actual managers. There clearly is some hierarchy targeted in-game, and it's more refined than what the stats show, which draw everything equal.

If this or similar technically isn't doable, the very least that absolutely needs introducing is some benchmarks as to the stats that are there, and some added basic numbers. In-match display the counters/breaks conceded by a side. Make a distinction between shots from play (e.g. a defense actually successfully stretched) and set piece. As much as FM may have helped to pioneer football data ten, fifteen years ago, it's behind the curve significantly, and it's limited use what's in there. In particular as on FM you traditionally can "dominate" every kind of stat you can whilst barely ever hugely challenging a defense. Cue in the 15 pressured headers on target from 20 corners and 40 throw-ins for "but" one goal. Or the 5 "CCC"s, all of which from a ball which sees the keeper edging it significantly, and thus but one at best goal -- as may be expected. And that's not much changed.

Read more about those here (and take on the quiz!): http://www.bbc.com/sport/football/40699431

[Football Manager 2017 17.3.0 Official Feedback Thread]

Breaking the records / finishing on FM.

 

Something different, though connected to the last couple pages of this thread in a sense. Recently was having a chat with somebody who argued that if "finishing" on this were more efficient, he'd have his forwards (a slight cut below the world class) scoring at least 1 goal per match on average. These stats are always going to fluctuate, but my argument was that this was always going to be more difficult, as he played a multiple forward set-up, which meant the bulk of finishes will be pretty much shared in between the forwards. Reason for doing is thus: In football almost every player who consistently averages a goal plus per match or thereabouts also averages a high amount of attempts. Ronaldo consistently averages 6-7, Messi 5+, Higuian a few seasons back when he was at Napoli and scored his 36 out of 38 FM  likewise with averaging 5 attempts per match (to put that into perspective: The most consistent Bundesliga forwards Lewandowski/Aubameyang averaged about 4 in the more recent season, none of which falling in that bracket.

To anybody following stats outside of FM that may not be hugely surprising, as long-term the finishing rates of better forwards are pretty much in between converting 1 in 5 to 1 in 4 shots or thereabouts, with only a few standout seasons (The "Hunter" Huntelaar at Schalke in a one-off season never reached prior and thereafter converting almost at 1 in 3), Kane looks similar the last season), whereas popular perception oft is that the difference between forwards must be like huge here (the same goes for scoring "streaks", which can be fairly random for just about anybody, unless of course you are Ronaldo and average 7 attempts of which it is expected you score a few. In FM you can flick that kind of feedback on manually, and currently, unsurprisingly at anything below 3.x shots a match you wouldn't even need to start from my experience (outside of exploiting a few ME defensive holes perhaps which mean every shot will be in loads of space). This also brought back memories of previous community experiments of "showing the AI how to make Messi Messiyish) on prior editions. Unlike AI, we have all the capabilities to completely tune the play in a way that makes most finishes fall to one guy, and I wasn't surprised to find out if you do the Maths here, Messi too averaged 6 attempts per game.



Would be pretty interesting what such stats look like for anybody who expects his forwards to consistently break the records, and how SI try to tune this -- the total numbers from my experience of fluctuate in between iterations.

[Let's have a game of guess the score...]

2 hours ago, Cougar2010 said:

@Svenc

I'm not convinced the ME plays anything more than a minor role tbh.

AI/opposition tactics are a different matter but users should be recognising the issues and dealing with them during the match. 

As for user tactics likewise, if they are causing an issue users should be recognising that and dealing with it.  Different tactics employ different ideas and different styles.  Its only natural that some choices will lead more shots through natural means.  Its not the MEs or SIs job to then go and crowbar in some coding so the user doesn't become frustrated.  Poor choices should be punished in the same way that good choices are rewarded with a little luck/random added along the way.

The long-term issue is two-fold:
1) Fm on a stats level not providing that decent a feedback. How much can you trust a stats page that never shows how many counters a side concedes or how much opening it creates (from actually play, given that boxes are packed on anything else)
2) There is always some approaches in there that make it very easy to dominate possession and shot counts, even against the best teams. Imo FM 2017 has made it worse as all you would need is an advantage in central positions. You can clock Barcelona down to 30% of possession in their own back yard with average sides that ways. However, as they are pushed back then, that regularly creates lots of set pieces typically, in particular when the measures taken to dominate the possession run fundamentally opposed to stretching an opponent (often the case). This gives off the illusion that you would be "all over an opponent" if all you watch is stats, when in fact something else is going on. With some more adaptive AI, it wouldn't be as big a deal naturally.

However I agree it's nothing you can't deal with! The problem is typically when somebody isn't suscribing to team sport logics (opening/denying space), and/or confuses having three four five times as many shots plus dominating possession (:rolleyes:) as "playing well". Which is naturally what most stats shown on Television encourage. It oft isn't. Not a slight bit. It's having more of the ball and having more "attempts", no more no less. Unfortunately though, the post-match "reports" players get also draw heavily from such "logics", reassuring continuously frustrated players that they were just "wasteful" or "unlucky", never making them question that their approach to that match/opponent may have been fantastic at keeping the ball from them and pegging them all throughout back -- but at carving their defenses wide open it were pitifully poor.

@sporadicsmiles Good post!

[Let's have a game of guess the score...]

25 minutes ago, Cougar2010 said:

Not sure why you have quoted me there Sven, MDW implied it happened more often on FM than RL, I simply replied with how often you should expect it to happen compared to RL.You seem to be suggesting that I don't experience it or that it doesn't get me frustrated which is far from the truth.  You can see that simply by reading any of the career threads I've posted over the last ten years.

  Reveal hidden contents

For those that want to know I won the Kotwica match 4-1 and lost the Sosnowiec match 1-0

 

It's totally worth to point that out, though it was off-topic, hence my response. The difference between you and quite a few players is that for them it IS happening regularly, in parts as so many download tactics suscribe to limited "football logics" whatsoever, the universally stuff that's been rewarded by this ever since (which is about creating and denying space), both when going forward or when the ball is dropped. Both in combination oft a toxic combination, in particular against opponents dropping deep. When somebody posts these "stats" to prove something, check for the corner counts for instance. In retropsect, going through a few FM 2014/2013 discussions, where the immediate reply was "happens in football", it is remarkable that nobody had ever picked up on corners piling up into the 30s whenever somebody posted such "stats", and that was just the corners, basically one "created" every 2nd minute the ball is in play ME or otherwise. Some average 15-20 corners on this in particular against opponents dropping deep, whilst in football teams as Barcelona average just 7! It's partly ME, but it's also related to tactics. In real football managers would adapt, well manage. On FM, that can be a very one-dimensional affair, with only the AI using some of the tools at the disposal. Cue in that short AI switch to 3 central forwards mid-match never reacted to by anybody.

Additionally, another thing worth pointing out is that in football (and to me, in the ME likewise despite the flaws) it's not the side with more shots more likely winning this, but those with the better ones. Not always, as that's not football, but FM's missing all of the latter parts of this.  It's easy to big time dominate the stats whilst largely creating nothing but a load of crap, and some of the most "obnoxious" posters on the "loads ofshots to no return" kind are still suffering on it despite they were explicitly told what their issue was.

@Maw74 3-2? Stopsley taking something from this is kinda given away by the match ratings (better to hide those). :-)

[Let's have a game of guess the score...]

48 minutes ago, Cougar2010 said:

Go look at the Spanish League table when Barcelona were at the top of their game, despite steamrollering pretty much all the opposition they still failed to win approx 1 in 4 matches.

 

The topic of this thread was a light-hearted "guess the score" play of all kinds rather than "I have loads of shots but can't score" frustration dumps, which indeed, are a classic. But if we're here again... everybody who's ever consistently dropped points in extremely one-sided matches (on FMs simple stats level) has an interesting idea about "steamrollering" sides, long-term experience. Some of it may be his fault on the occasion (or at the very least, the game's stubborness to never teach his team sports basics 101). However, some of it is also ME related, and AI tactics. Consistently having huge stats and not scoring is typically down to barely creating space -- (the problem then is oft that it is still easy to get shots.. just mostly not very good ones). You should try some of the stuff that is shared, (and so should SI, as some of it could help to improve the ME/AI or at least understand // avoid player frustration). However, that's not the point of this thread, and any such is a flawed assumption, not only due to ME/AI. But also due to top class management typically managing space and play, rather than stats.

[Let's have a game of guess the score...]

34 minutes ago, Marathon Dress Week said:

I love how people point out the once per season something like think happens IRL as a justification for it happening weekly in FM. 

1) FM isn't RL (for better and worse)
2) The managers typically competing in FM aren't real managers either (they wish they would be)
3) They seem however, too missing the point of threads completely

[Let's have a game of guess the score...]

1-0 Southampton of course, as that wide ball not only results in a shot that changes direction instantly to go on target, making it an even bigger challenge if it's a first time shot rather than taking from the 2nd, 3rd, fourth touch (possibly involves bad defending for the winger to be unmarked, of course). Unlike the Drogba scenario it's also a ball played straight into the running path of the winger, whereas Drogba is facing the ball / player rather than the goal and either has to immediately turn and finish, or if he stops and shoots, has immediately all those defenders breathing down his neck plus the keeper possibly coming out to worsen the (centrally) angles too. Ok make that 2-0 from an own goal, as else that isn't a particular curious match from the stats, unless the scoreline would be 3-0. Naturally all of those could have resulted in a goal as far as you can say from still shots, but why else show those when the match went for Bournemouth? Perhaps it is 3-0?

 

As for this one, huge possession counts coupled with an above average corner count always looks suspicious (this is experience), and smells of a side pinned back all match but rarely stretched for set-piece-o-ramas (3/4 of those on target shots so-so headers etc. from a forward marked and under pressure, which I personally consider an ME weakness that it is so easy to amass those so "easily", btw), so probably no goals for Unterhaching, with the possibilty of Nürnberg winning it from an own goal probably helped a [tiny] bit by the frustration that would typically settle in upon wasting so many "quality" shots.

]

[Let's have a game of guess the score...]

 

 

On 26.6.2017 at 07:27, KlaaZ said:

 

Probably because the away team had an average rating of 8 which indicates a very high scoring line.

Good spot.

 

I'd have a pack of three numbered in case anybody just wants to predict merely one... there was a purpose for posting multiple matches. Edited and added the scorelines and reasons into the spoiler tags below so to not leave anybody hanging.

Match 1:

 

 

Unfortunately, AI managers don't deal in makeshift backlines consisting of a wide full back and no protection in front of his. So yes, Concord was me. However, should SI ever want to stress test CCC / one on one conversion... or generally decision making in front of goals... everybody who vents about 40 shots no scoring should test this to get a feel for what decent chances really are ... and which aren't.

 

Match 2:

 

 

 

Leicester was me and City was an AI manager that went more stupidly aggressive which each goal conceded.

 

Match 3:


 

 

Yes, those are full-match stats. We should be glad that AI isn't able to take that entire Tikinaccio crap to the next level after opening the scoring early on.

The general purpose:

 

This thread is good fun! However, be aware that the ME isn't quite 100% real football, and that we, to some extent even AI, can easily make sides play like no football team does, so real-world stat comparisons can be a bit tricky. Additionally, a few of those "key" stats are REALLY simplistic compared to what football analysis and actual managers use these days. Basic rule of the thumb, don't trust a stast page that shows little stats that give an indication of how ridiculously open anybody's back doors may be right there.

[Let's have a game of guess the score...]

3-1.

2nd minute: 1-0 off a counter attack due to a hole in there that big no team in competitive football actually commits to outside of schoolkid's levels. Lynn's manager applies some fixes quickly, hopes nobody was looking
3rd minute: Grimsby retreats into its defensive shell and lets their opponent happily dominate their stats
10th minute: 1-1 King's Lynn lucky scramble off an indirect free kick. Grimsby opens up again
12th minute: 2-1 scored off an assist of the jumping reach/header/strength 16 Target Man each, who as predicted proved troublesome for the Lynn's dwarven defense before
13th minute onwards: Grimsby building the Walls Of Jericho for realz, doesn't even try much for the remainder of the match
90+3th minute : 3-1 scored after the badly managed King's Lynn team went all frustrated/nervous/pretty pissed, leading to an error and an OG

Aftermath: Doing its very best to pay tribute to all the completely clueless football punditry in the world -- post match reports goes "wasteful King's Lynn!". What it failed to pick up on was any of the above. Or that King's Lynn tactics compressed the visible attacking space to the size of a tuna fish can even without Grimsby needing to "contribute". In turn leading to 3/4 of the 9 "on target shots" from Lynn  being from set pieces exclusively, which meant the forwards were under pressure on all, making life easier for the keeper (still a MOM award from those clueless amateurs, mind). 3 of the supposedly "clear cuts" were from yards out / immediately blocked, and two of them each fall in that range where you would expect to convert 1 in 5 at best, as whilst the forward was in reasonable space, the GK had it reasonably easy to worsen the angle by rushing out of goal.  Lynn Manager pretty pissed, rage-quit, logging onto community.sigames.com to post how he's never going to buy this again.

tldr; Not actually quite serious, but then it's tough to tell without knowing any context.

[Football Manager 2017 17.3.0 Official Feedback Thread]

14 hours ago, El Payaso said:

Also this tells a sad story about the tactics system when you see the team selection. A player like Tiote should not perform well in a playmaker role as he has no qualities whatsoever to play it but the game doesn't seem to produce these "this role is too much to handle for this player" kind of situations.

Not a defensive of this, but partly it's a conceptional thing, it would be likely easy to introduce something such as a genuine "role" gelling itself and a harsh penalty if a player would play here (under the hood, historically at least, every role is a player given sets of instructions). The game simplifies much, I'd argue it starts right with the transfers, but tactical gelling naturally too... basically you put a training slider on tactic preparation for a couple weaks, play a couple matches. What all of this comes down to if how deeply a sim game you want this to be, and you may get different answers on that -- the sheer existence and popularity of download tactics tells its own story on that. But if you watch, Tiotes mostly not much under pressure anyway, which is due to the compactness or lack thereof the ME produces, this would have been little different before FM 2017 (personally I oft man marked the deepest opposition midfielder due to this, in particular playing 4-4-2ish). Can also be seen on the video clip I posted two, three pages pack, the deepest guy tends to be free to recycle. Plus with a set-up like that where a marauding centre five midfield wreaks havoc via overloading (easily) (6:50, 18:30 -- CB steps up -- d-line is no more etc.), he doesn't really need to provide much of anything tbf.  Which naturally, is what makes those things so effective as of now. Too effective, as you said. Fully agreed.

Not that Tiote's much involved here apparently thus. You're bang on, but it's not something I'd personally call partucilar slow build-ups (by FM standards).  The guy's instructions don't seem to much allow for it, the obviously d'l tactics on this (seems to provide those himself) tends to push the CM strata up rather quickly via his exploit the middle TIs etc., opening a gigantic amount of space pretty early in the moves behind it and the defenders, which is also how he concedes the space so easily against Celtic in the last match.  Early interception, and they can break forward. If anybody is ever wondering from where that "game is rigged" posts/reviews so popular originate from, oft based on the genuine experience opponents regularly find it easier to score, here it is. In particular that there are spells where the AI conversely neither pushes its backs up an inch, nor any of its central midfielders, plus doesn't universally play such highly aggressive lines like that, where every through ball to a forward can see him goal-side, but sit deep. 9/10 downloads tend to be as positionally superaggressive if not moreso than any AI opponent on this game at best is when being totally desperate to score.  Naturally they're being advertised "plug&play", that is they're not being used for the last ten desperate minutes, but every single minute of a season with any team no matter how poor. On that part at least this were a quite silly ME if that wouldn't lead to easier ways goal-side, but others tend to disagree.

[Football Manager 2017 17.3.0 Official Feedback Thread]

@ElPayaso

Yup, that's feeding into the weak areas covered two pages back, and pretty much the shape how the most efficient exploit tactics look on this. This shouldn't be that easy, but the central areas have been severely weakened on FM2017. Flood the centre, and defenders are dragged everywhere. As to why that was covered, you have detailed it too. They don't at all have an issue with scoring or general conversion though, unless the AI drops deep and stuffs that area with players usually, breaking any kind of scoring records in the process... even mid-table to bad teams seem to average 3 goals per match with this, cue in the additional miracle results (8-0, 11-1, etc.). It seems even more pronounced if the AI plays a back three and oft doesn't have a proper protection in front of it. It's the biggest bug bear I have too, but outside of avoiding to play like that, no choice atm. Highly doubt that this isn't acknowledged though. Which can be both good and bad. Good for anybody who wants this to play out more and more like footie. Bad for all the guys reporting in a couple months how FM 2018 may have broken their "perfectly valid tactics".

The most interesting in that thread is naturally somebody downloading "exploit" tactics on a forum reknown for publishing engine breakers -- only to report them!! Must be fairly new to all of this. It's a good example of how that kind always highlights issues that may be always there lingering, affecting anybody in one way or other anyhow, but by specifically targeting them making them stick out like the guy dressed pretty in pink at your familiy's funeral, rather than them getting overlooked or flagged minor issue in the initial testing.

[Football Manager 2017 17.3.0 Official Feedback Thread]

5 hours ago, El Payaso said:

One thing also to note in the engine is that it should be balanced on small details too: defenders at the moment seem to be really error prone (clearances straight to opposition, missed headers etc.) so it should be balanced by making attacking players make bad first touches, runs to wrong direction and misskicking the ball when a chance is presented. I've seen none of those happen to them and while it's been made quite easy for attacking players to perform in the engine, I would say that they are too precise at controlling the ball in many situations.

I'd agree with that too... technical stuff isn't modeled precisely/varying enough. I'd even tested that on previous by going into the editor with extreme data experiments. That goes back to the defending stuff some. At the moment FM is limited in how it models the more physical aspects of defending. There is no real shoulder charging modeled as such when players compete for a ball, they can't really shield it either (either to waste time at the corner flag or in general), etc. I think that is also in parts where that perception of chances is coming from... If the more physical side is somewhat limited, it can appear as if players where unchallenged. However, as FM obviously tracks the movement of players (or 2d dots, whichever tickles your fancy), it's not hard to code into that the ME recognizes when there are defenders near, and the guy is being flagged "challenged". Hopefully will see improvements. It's interesting that you should bring up 2013. AFAIK was a version where due to lacking support movement around the ball carrier the shot counts could rack up by itself, as they would then be forced into additional ranged shots.

As for defending the game via the CCCs, my experience is thus. Often an accusation directed at the mods. Take a few pages back at what themadsheep posted. He'd want the thing to be taken out as he knows it's bad idea and is never going to work. He doesn't say that because he defends anything. He says that he knows it's tosh. Firstly, whilst the argument against Expected Goals or similar seems in parts that this wouldn't be standard and used anywhere much  -- tell me where exactly a wholly subjective clear cut / half chance category is ever used on  Match Of The Day (possession and shots, that's all there is). Secondly, even against better judgement, such a category tends to be treated as if every attempt was equal, which is ludicrous. Thirdly, it's being misused thus as objectively "proof" how the game were broken, which no number will ever show how and why, as it's the match play making those numbers. Realistically, all of this would only really go away if SI could introduce a red flashing bar that pops in every time, like "NOW THIS IS A GOOD CHANCE! (what we have programmed one to be anyhow". However, none of that were of any use if it wasn't acknowledged that even a good chance may see the keeper in the advantage still. Unless, it is a penalty, tap-in, etc. outright.

I would always upload any such match, it was acknowledged there to be an issue with certain kind of one ones on an older iteration before. This harks back to how SI should invite the more popular "engine breaker tacticians". Their stuff will be copy/pasted at the core all year anyhow, there is few variation on that. They shouldn't care about that. What they should care about is that this always highlights issues that are missed on testing. It should be obvious that tactics influence who gets a chance in what positions from what kind of assists, i.e. keeping wide guys back in a narrow formation or playing ultra narrow in general will tend to never involve much angled balls from wide positions and create exclusively narrow one on ones (may favor the GK some more). Which still also happens to be where an issue was with before, as contrary to "exploits" likely no balanced "tester" tactics nor any AI tactics visibly ever is that one dimensional in the type of shots it creates. They may all try to let their team play "football" within the confines of it all, rather than aim defensive holes. If that were the case, this would be always picked up on testing.

[Football Manager 2017 17.3.0 Official Feedback Thread]

4 hours ago, El Payaso said:

The CCC in real life is a bit of mystery. I would say that IRL it doesn't necessarily need to be an easy chance but instead it might fill some certain requirements. So it might even be a shot from a narrow angle and under pressure but it fills certain requirements and because of that it's a CCC but still not necessarily even close to an easy finish.

What should be a mystery is the FM CCC! Unlike the Opta one, it doesn't have a definition at all. Which is why on some releases the header from point blank range into an empty net was considered clear-cut, and some not so much, and you're left guessing whether this season it's header season or isn't (there seems to be an on-going debate behind the scenes whether header finishes should be inluded or not, and several more, it's always changing). Opta doesn't do that. A CCC (or Big CHance, as they call it nowadays is always a resonably high probability shot, typically a one on one. Yes, the keeper may edge those, but in the grand scheme of things, that is a high probability shot. Even for a shot to have a goal expectancy over 1 in 10, a forward would need to be in reasonably range and space. What's oft forgotten that he still has to trap and control the ball at oft high pace, and that shows. Defending is easier than attacking.

http://optasports.com/news-area/blog-optas-event-definitions.aspx

In-game I ignore it entirelly -- and even that Opta stat is barely used anywhere. Looking at the shot spread in the analysis (ranges, positions) and watching whether a side just spammed stuff from set pieces (easy to do in-game) tells me more about an FM match than any other stat in the game. Which is the pity. There's a whole load of fake domination in this game, some of which relates to the defensive issues you also touched some (i.e. rather easy to keep the ball and/or drag defenders around the pitch). Not least of which because there also isn't a stat displaying how much that all conquering super tactic collecting 80% possession every match plus getting off 30 attempts (none of which in any space) has its back doors open, i.e. counter attacks, through balls in behind aggressive unprotected lines. The only one of which, naturally, would be constantly conceding a stupid amount of goals out of a lowly amount of opposition attempts, at which point it's all too little too late...

Or dare I say too little to not ragequit.

[Football Manager 2017 17.3.0 Official Feedback Thread]

6 hours ago, El Payaso said:

For me the match engine is quite much unwatchable because of the reasons I gave above. The variation of goals scored is really narrow and I would claim that maybe one out of four of the goals we see at the moment are 'valid' and could happen in real life. I don't prefer watching games where slow tempo attacks with two players against six lead to something. Leigh Griffiths might well again be the best striker in the game because one of the most common scoring chances is the "good" old through on goal which once again is happening far too regularly and from slow build ups. In just two games against Barcelona and Dortmund I've had about ten through on goal type of chances. And with realistic finishing quality we would have scored 15-20 goals in those two games but had to settle only for 6 goals.

You should upload those in the bug forums. I have never seen such matches outside of FM 2015 before the first patch, where there was space all over the pitch as some AI could have its entire AML/AMC/AMR strata not getting behind the ball, turning matches into counter attack / through ball / 6 man defenses easily overloaded fests. They had hockey scorelines every other week, same as human players who did similar (this isn't possible to do anymore). Class forwards convert no more than about 1 in 3 of their one one ones  though. I agree with quite a few of the defending stuff you've brought up (also seen on the last page). But there is the theory that the engine would nerf finishing down to comical levels, which isn't the case. If the CCCs were legit, you should expect to score no more than in between 1 in 3 to 1 in 5. Or any shot in space. The equivalent of that would be Opta data CCC / Big Chances. If you google, you will find EPL starters who missed 80% of their totals. Whilst I don't think that FM is accurate, not least because it is a limited computer sim, it must take any such into account or it would be a fantasy, such as Fifa, Sensi Soccer, Nintendo World Cup. Outside of tap-ins and penalties, the keeper is supposed to edge this.

Imo it isn't that there is a gigantic amount of shots in space, outside of exploiting the stuff from the last page, and that is a major defensive issue (simply google for DEMOLISHER tactics for 10-1 miracle results) -- unless a team jam packs that middle with players. What is rather easy creating all those other ones. This goes years back, with human controlled matches uploaded showing up to 40 corners a match (unwatchable). It's not limited to those illogical tactics, it's just highlighting it to make for terribly viewing experience (and little goals). Aside of the more obvious (ranged attempts, easy blocks) as argued, if teams dismantle an opponent 7-0, it isn't at all unusual for them to have 90%+ of their attempts not merely within the box, but also from open play. On FM that is very rare if at all possible. A high amount of shots not being from play would typically be the telltale sign of a defense not getting stretched, FM or otherwise, but clearing every other move. As there isn't a stat for that, you have to trawl through every shots post match manually, which  is quite a bit of busy work. Hopefully SI are able to collect those internally.

Just seen that FrazT logged something that may be connected, though it relates purely to amount of free kicks (which are converted at about 1 in 25 on average no more, and for all the highlights one could upload, Ronaldo et all did go fully seasons with minimal return). My first experience with that dates back to FM2012/3ish, where for fun I made ten players just sitting back and not advancing, and still it was impossible to go without a shot, as eventually you would win a free kick, a corner, a throw in, and get shots off those. Which just happens to be how those impressively defensive AI teams also collected their sneaky 1-0 wins (off a corner, the only shot all match) once in every SI logo shaped moons. In other words, shot stats on FM are always inflated by attempts that you aren't much expected to score from, no matter what you do.
 

 

[Football Manager 2017 17.3.0 Official Feedback Thread]

On 24.4.2017 at 10:25, El Payaso said:

Statistically for example the amount of goals is alright and I think that the results in general are in realistic way which makes it okay to play with only commentary. But looking more close I would say that there are far too many good scoring chances. Even with a decent club you can basically in every game rely that you get more than enough chances to actually score. Even with big teams like Real Madrid it's okay to win games by 3-7 goals but getting that huge amount of CCCs and HCs is telling that something is wrong. In real life what creates the goals for them is not that they constantly create huge amount of chances and shots, it's the quality of the players.

Touching on some good points. The shots and CCC issue though is a long-term ME (and definition) issue, but I don't think in the way you make it out to be in parts. Firstly those chance categories are and have always been as reliable as Arsenal are at missing those (up to 5 a match in football). Most of those CCCs in-game really aren't all that when clicking on the dots. Some of that was on the last page, and Dave1990 will likely link his CCC bug thread again. It has always been easy to amass shots however, this isn't just human players but also AI. If you want to provoke it some, simply drop deep. Good shots, that is some different. What SI should be looking at is those low-percentage shots inflating the numbers, leading to fake domination depending on tactics of both sides. The matches where up to 3/4+ of the stuff is from set pieces, range, forwards immediately closed down, a mix. Oft highlighted further by one-dimensional human player tactics, of course.

That's a domination in all but a very simple stats. If this had anything like the Expected Goals model off analysis, you could have 20+ attempts, and the expected amount of goals to show still well under 1. At its most extreme, matches can be borderline unwatchable, with every single attack deflected or stopped by a foul or the ball kicked into the touch and the forward becoming increasingly frustrated/nervous blasting the few half decent over and over again. I too wish for a ME however were bad tactical approaches / massed defenses don't necessarily lead to a big amount of low-percentage shots, but no to few shots at all. Crazy tactics will likely always be prone to funny stats. Still think Si should invite somebody like that into their Beta testing, if they haven't. This can highlight areas that aren't picked up, as the testers likely mostly by "logical" setups, which then is also little different to AI managers soak tested.

You can win/dominate leagues with averaging but 14-15 shots per match and two plus goals averaged off those, which is on the lowly side, provided those are decent shots and there is space created. So it depends on tactics at least some. It seems a fairly subjective area. Whilst for you it is perceived there's always enough chances to score at least one for your side, somebody struggling to score that one goal despite masses of statistically attempts typically perceives that a bit differently. Thumbs up  to any encouragements made for AI manager improvements.
 

[Football Manager 2017 17.3.0 Official Feedback Thread]

20 hours ago, Mr U Rosler said:

Whilst I've ended up back at a 4141 (+ defensive variant) I actually used 11 different tactics during my last promotion season seeking consistency.

Every system has strengths and weaknesses as there is no perfect set up, i'm over achieving and happy with this............ for now!  

Also a fantastic statement!! I think I confused it though. I meant the 4-4-1-1 with two b2bs that lead to wild results as you had recorded with 6-4s outright, and as you found, streaks could already be influenced by the ups and downs in reputation. That one looks fun, but as you said, wild!! https://postimg.cc/image/mzxg9inhx/

 

Quote

I think half of all FM discussion is people being fooled by randomness.

A bit like in football, where good runs can be wholly down to the fixture list plus nothing but luck. I love the current Bundesliga season, where the difference between getting embroiled in a relegation battle and competing for European spots is basically a couple lucky strikes and the top performers legs finally lasting a fully season. Yet  despite it all there's still a club for every Euro / World Cup that overspends on the guy performing above his ability for five matches and when arriving at the club, soon degressing to the mean (or bench) thereafter. Translated to FM, that's why rather than looking at trends, it may be worth looking at the longer-term. The game out of the box don't provide much on this outside of the player and team reports (overall shot conversion). This thread is a classic as it demonstrates how we can perceive those streaks/saves, and sometimes things aren't as dire as they are perceived at all:  https://community.sigames.com/topic/243260-why-is-the-ai-so-much-better-at-finishing/?page=10#comment-7831263

I think part of the problem of this game is that you can be incredibly successful overall without having the slightest clue as to why, up to the point that outside of transfer windows you would just hit continue. In particular as the game seems considered a "learning sim" as per SI, a thing where you never really stop learning (in particular after AI upgrades and the like... which hopefully would be slightly more clever by now than on the no holding midfielder set-ups as of FM 2012ish, making midfields a walk-in every time, but then there is still enough weird stuff going on, as argued previous). If one's doing fine overall, why bother? Just carry on, and blame every hiccup on the FM gods, the weather, the ball being too bouncy on the night and the broken match engine (which is clearly not without faults, see also last page). But that's a different topic.

 

[Football Manager 2017 17.3.0 Official Feedback Thread]

17 minutes ago, Mr U Rosler said:

The upshot of this is I only use a lone Striker now (4141), don't focus play through the middle and spread the goalscoring burden across the team (especially midfield) so you are less vulnerable to Strikers hitting a lean streak.

    

 

This is very very good advice! If most of the shots fall to one guy, or you create only one type of chance in general by channeling play through one two players, you are more likely to be defended when those guys don't switch on.

The last Rainmaker you posted though is also vulnerable to those streaks imo, as it is at least rather risky and doesn't suscribe to how teams play football. I mean it looks fun and nothing should ever prevent us from doing anything like that (part of the attraction of the game surely is all these what if questions...). But it's a 4-1-4-1 with both of the central mids which would/could protect the d-line on sup duty / box2box, plus the exploit the middle TI enabled on top. What that does is making both push up ASAP (exploit the middle pretty much tells every supp duty player in the middle of the park to "push further forward), compressing the attacking space too -- if nobody is staying slightly deeper there is no back pass on to ever rebuild attacks and back passes also force defenses sitting ducks to push up again rather than sitting ducks until the move is intercepted / the shot is blasted. Depending on how the AI plays (pack the middle, drop deep, lure out, etc.) this can become more/less effective overnight. One of the reasons to have a reasonably balanced core, though as said, in an ideal world AI would actually adapt some to this, rather than creating streaks simply because different managers have different bias and preferred formations.

[Football Manager 2017 17.3.0 Official Feedback Thread]

The game may actually be too streaky! The question is always, why does it happen to your team then, but not the opponent. One-off matches in which forwards don't hit a cow's arse? Possible. Can also a feel streaky. Continued issues? Impossible without own contribution. And I'm saying that as somebody who currently also goes through a bit of a rough patch in some matches I haven't had in seasons. My suspicious is that it is due me compressing the space too much in parts, but I haven't yet had the time to analyze that further. I.e. playing a very aggressive narrow style, which usually isn't my wont. Please don't go by Half Chances and CCCs. They are borked still, there's multiple threads in the bugs forum about it, and long-term, the CCC conversion isn't much higher than for SOT (and shouldn't be). Anybody arguing by those shows that he hasn't been much around the good advice usually imo, since there is like zero decent players who argue to go by this (I know quite a few who have argued for them to be taken out for years, I'm strongly for it for a host of reasons). That aforementioned shot breakdown would be tons more useful and telling. Fairly easy to rack up attempts like crazy, which in parts may be a stats, but no doubt also is an ME issue, the quality is another matter.

Even if the game were accurate on this and that you could (man-)manage this some further, I'd expect this to be quite some streaky if this is to be a sim. This is because outside of tap-ins and penalties and the like, the keeper should pretty much always edge this (i.e. shots that see a conversion chance above 50% should be very rare... and one one ones aren't converted above 1 in 3 to 1 in 5 rates in general. Defending is easier than attacking, which is one of the reasons what makes football that unpredictable to begin with -- not the most ideal sports perhaps for a management sim, let alone your blood pressure, but so it is). I view each shot as a dice roll with weighted dice (general type of chance the biggest factor). As most of that should be favored for the defense, that is what can and should of happen. I have never had general long-term issues with that though, and the stats in the game in generally back that up. I don't for a second believe the game to be fully accurate here (it's a computer code, and one that models defending differently to a lot in football for a start, see above), which is why individual uploads are always worthwhile, but I know that the teams I compete against are all affected as well, as their matches go through the exact same thing. Some AI is actually still that borked that they need 20 attempts on average long-term to convert. It's mostly the aforementioned bottom type who sits half their team for prolonged minutes of a season behind the ball and thus never generates much from actually play, exaggerated by a couple of other AI tactical issues imo. Still, Arsense can't blame a bloody game.

http://www.skysports.com/football/news/15118/10189870/wasteful-arsenals-premier-league-title-hopes-hit-by-poor-finishing

Assessing counter attacks is an interesting topic though I'm personally also interested in. Traditionally, they've been rather toothless overall. That doesn't mean there may be no bigger issues though... Talking about defensive cover for attacking corners, the "good" thing is that the cover for AI on this is also fairly threadbare. It seems the default instruction is "stay back if needed" pretty much. That can be pretty easily overrun by now on the occasion depending on which. Not sure how about the reaction time, but I'd be interested in that as well as it is the play making the numbers, not the other way round. Anyhow, happy Easter to everyone including SI Towers Of London!