Svenc
-
Posts
5,156 -
Joined
-
Last visited
-
Days Won
1
Content Type
Forums
Articles
FAQs
Online Manual
Support: Blocks
Support: Games
Bug Tracker
SIGames Manual (beta)
Profiles
Posts posted by Svenc
-
-
20 hours ago, Whufc10000 said:
I haven't been past season 3 in FM21
Almost all of my saves are in the EPL and if I play as a "Big Club(not Liverpool)" then Liverpool usually reach 100 points or more in the first 3 seasons.
If I play as a mid table team or relegation fodder then whoever wins the EPL usually do so with a points tally around 79-92, until of course I'm able to challenge for the title and in that case usually Liverpool or Man Utd go back to picking up 100 points or more.
That said, there's nothing whatsoever like that in the game (as also mass recorded on many FM platforms). This is again not defending the game, but being concerned in terms of long-term development of the series. If "cheating" or "Unfairness" perceptions persist, the chance of AI improvements drops below zero (if the AI would improve, the game would be perceived as even more "unfair"). The game has actually not even ever giving a **** whether you're at all doing anything, outside of starting a save and pushing the space bar to progress the date.
2 -
20 hours ago, Whufc10000 said:
What does the Expected Points table look like by the end of the seasons in comparison to actually points taken? In general, I've seen AI managers underperforming that on occasion by as much as -15 points.
Liverpool in terms of data are very likely to be overrated based on their previous league campaign. As Statsbomb based on data now available in-game have analyzed, to break records like that in a competitive league such as the EPL takes more than being simply a great team (if anybody had managed City that term on FM, he'd have accused AI Klopp of "cheating" no doubt, as Man City comfortably lead in xG difference all throughout.
[quote]Being (lucky and good): Liverpool[/QUOTE]
Being (lucky and good): Liverpool | StatsBomb
However, LFC also took the lead an awfully lot as STatsbomb had pointed out in a later article. Which meant they could sit on it in an awful lot of matches, thus influencing their xG. (LFC and Statsbomb have just signed a deal, whilst we're at it, and in parts Klopp is at the club because their own analysis team had analyzed he was getting spectacularly unlucky in his last season at Dortmund -- sitting 18th place by February despite being equal or better almost every other week and finishing 7th when their own analysis had concluded they should have finished 2nd).
I'd place a bet on that in the summer and for the next FM this will swing the other way for LFC and the research will nerf their data a good deal (in particular if Liverpool indeed don't qualify for Europe). I've noted this with research before (for individual players also see Morata who was big time nerfed after a season perceived as lucklustre at CHelsea by the general public going from finishing ~16 to ~10 in the space of two FM releases).Whilst research shouldn't be based wholly on (data) analysis, I think it would benefit if it would still be at least considered, in particular as single football seasons might be influenced by chance quite a bit. FiveThirtyEight for instance still ranks LFC as the 2nd best team in the Prem, despite their struggles, and in the point tables too, things are reasonably close overall at the top with goal differences also being suspiciously similar (outside of City who dominate the point table as well as goal difference now as much as they did xG last term when LFC won the bloody thing).
Huge rerankings and/or significant boosts and nerfs in the space of but two releases should be a no-no in general.
2 -
4 hours ago, zyfon5 said:
I have argued before that the introduction of xG will not reduce the number of comments that complained about why they are not scoring and I am glad that I am not proven wrong. My argument is simple: players that have develop the ability to create and identify a good scoring chance will be able to do so regardless of xG available or not while players that did not have such ability will fail to do so regardless of xG or not. They only know that it is a poor scoring chance but they do not know how to improve it.
This is also a really old post I'm going to link to, but that's also a good point: Having (statistical) feedback that there may be an issue is not the same as being able to fix that issue. (In this case, answering the question of how you may increase the chance of scoring against a particular opposition)
In a sense, this goes straight back to this old post I was about to link to. One of the keys indeed may be making players asking themselves the right questions in the first place (and being able to answer those questions via the game's tactical UI).... Not a Rant, But.... - Football Manager General Discussion - Sports Interactive Community (sigames.com) (third page first post by wwfan back then).
[quote]Can you answer these questions, conceptually, in terms of real football?
1: How do teams break down stubborn defences? Think Man Utd, Arsenal, Chelsea. Do they all do it the same way?
2: How to teams set up their defences/tactics to guard against the quick break?
3: What changes do teams make to their style of play in different weather conditions?
4: How do various teams see out tight games? Again, think of the top four, but include sides like Stoke and Bolton, who are past masters at this.
5: What type of role does Henry play? Messi? Owen? Berbatov? Rooney? Noble? Ferdinand? Walcott? Ashley Cole? Carragher? Can you see equivalents in FM? Can you apply them to various WHU players?
Translate these ideas into tactical shapes and shout strategies and experiment with them. Don't expect immediate results or free flowing football. Work out how to win from the bottom up.
....
If you are unwilling to even think in these terms, you will never get to grips with the ME.
[/quote]As said, to me the solution would be tactical assistant manager proper (which would also benefit players who deem the game to be too "easy", as any assistant is AI all the same which would need improvements -- win win for both sides). That assistant manager would be optionally able to guide players through the thought process required, the UI, and also would be optionally able to take over match management himself naturally (but with added input of the player if wanted, such as choice in style of football as well as timing in-match decisions such as holding onto a lead, going for broke, trying keep damage to a minimum after an early 0-2 so as to not further damage team morale, etc.)
That's how I'd envision a "football simulation", no less as actual assistants in all areas have been part of manager job for long. However, I also suspect this would rub more traditionalist gamers actually the wrong way. I've already see traditionalist gaming press in Germany questioning that you can already be fairly successful in FM going on holiday... "What's the point in playing yourself when you could delegate EVERYTHING if you wanted to", they'd probably say.0 -
4 hours ago, Outrospective said:
If the game wants to show me that my tactics are bad and ineffective it shouldn't resort to illusion of my team making constant chances and failing to convert them (hence all "FMed" memes). That's why I'm inclined to believe that match engine doesn't show things related to my tactics and doesn't making it easier to improve it. Although it's a primary source of game feedback to the player.
What version are you on? That's why they'Ve introduced xG to FM21. (And from my experience, most FM players were pretty bad at judging the quality of chances -- probably not their fault, as TV analysis on this has always been pretty damn cack too, and that's what we'Ve all grown up on).
xCommentary | StatsBombTen years ago we've threads about this...... Lets Assume It's "MY" Fault (time to ask for help?) - Football Manager General Discussion - Sports Interactive Community (sigames.com) (on page 3/4 starts the then purely subjective discussion about the quality of chances -- the guys actually having no problem scoring were ignored in favor of AI cheating theories and more). Now we finally have somewhat reliable numbers to back that up (that will also hopefully get tuned in the future).
1 -
6 hours ago, Outrospective said:
Correct. That's what I was talking about. The game doesn't show why I had constant failures as well.
This could be improved by a few of things (already made a few suggestions --- the xG stat now in the game which SHOULD help on average and in the longer run of things I've also been advocating for years as the "clear-cut chances" have plain never worked -- they lacked definitions to begin with). On FM (outside of bugs) and in football, sides that create the better chances have always had higher chances to win football matches (which does not mean that the team with the better chances always wins). Whilst stats in isolation aren't everything (and many actually manage purely by final match stats): That's being finally expressed now more clearly and with more nuances.
That said, out of interest: Given your theory that the engine would show random highlights how goals are scored, and that the match play wasn't "genuine", have you ever tried to read what was going on? And what are your personal issues with it? The entire thread is criticism of the game (albeit from a specific part of its audience). It isn't called "Game's fine", its called "game's too easy".Therefore, people pointing out that you have a wrong idea how the game worked was being done because if you further insisted on that any highlight/ goal you get to see was arbitrary, you will always find the game to be too "hard", or at least harder than it can be.
Being able to read the second by second match play and making decisions based on that is actually where one of your biggest advantages over the game's AI is.
For the time being btw. there was a pretty neat FM myth busting thread a while back.
Greatest FM myths - Football Manager General Discussion - Sports Interactive Community (sigames.com)0 -
-
2 hours ago, Experienced Defender said:
If your aim is to win the title/trophy by playing a defensive style of football as your main (or only) tactical style, then it's highly unlikely to happen. After all, keep in mind that the Leicester side you mentioned was much more of an exception than the rule.
Haven't tried, but back then that worked well enough. At least until your reputation would rise and more and more AI would play defensively / shut the shop (even the bigger dogs... FM AI). I also went with an attacking mentality, as that is more suited to quick transitions from back to front (dropped naturally much deeper though than its default, plus encouraged a more direct game than its default). Reckon that's still one of the more common misconceptions to this day (UI issue if you ask me, plus "mentality" naturally isn't quite a football concept but a fairly abstract FM concept).
However, the AI on that release also helped back then in that if you scored the 1-0, not only went they more aggressive with ultra high lines to exploit for fast forwards such as Vardy.
nnwB8V5.jpg (1228×812) (imgur.com)
They also oft had but a single player kept back to protect their own attacking set pieces.
CGw92DK.jpg (1204×798) (imgur.com)
Was top of the league by December with the worst pass completion, the best shot conversion (closing in on the 20%, that is roughly every sixth shot average being a goal) and, had FM dealt in any such stat, the most shots conceded average of any team in the top of the table (Leicester likewise had won the league by conceding ~14-15 shots average). I think my role/duty setup for the defenders iin parts wasn't that different to the one in Jack's post (three defenders on defend plus a midfielder), however, I played off a 4-4-2 or 4-4-1-1 with both wingers afair on attack.XvbF1vS.jpg (935×496) (imgur.com)
coh2t5N.jpg (919×506) (imgur.com)
nbhpdCn.jpg (930×504) (imgur.com)
sOLwprt.jpg (880×333) (imgur.com)0 -
I've got a request (the bolded part at the end).
That said I've fielded players out of position before, albeit full teams such back then with Barcelona (didn't do them particularly good). Pic-Upload.de - barca.png (pic-upload.de)
As to scoring goals, I've always been arguing that on this game the chance is more important than the player/forward who has it (which to a degree is naturally true in actual football itself). Systematic finishing issues have always been tactical as to FM, as they would influence the amount and type of chances had. Additionally, the AI has many times never been able to replicate Messi-like scoring rates because many of their tactical approaches are not suited to making a single forward as much of the focal point of play (players like Messi have 5+ attempts average per match for a start, CR7 during his prime even 7).
Still, these were youth player GKs. Naturally, that's a tad different.
Now with xG in the game, I'd actually love to see a repeat / update of this with more indepth stats. Would the youth GKS actually underperform their xG in the lower run (as you'd expect them too)? Likewise, would an elite forward overperform it in the longer run. At a club such as Liverpool, you're going to get many chances, after all. Moreso with a tactic downloaded from FMBase, which is basically FMExploit Central. Despite scoring semi-regularly, you may still be actually underperforming.0 -
Yeah, actual ME faults naturally have to go away. What's become apparent to me in general is that there's also a HUUUUGE difference between having success on this game (many/most do) ---- and understanding WHY you have this success. (Would be interesting how many of the download "super tactics" are purely trial&error jobs, whilst we're at that). In a sense, that's a game failure too.
Until that gap closes there will always be bad theory and even accusations of AI cheating. This can only ever done by honing the ME, and at least optionally improving feedback (readability of the match play, assistant advice, understanding AI match management).
Then again, cheating theory will always be there (unless they tweaked things to actually be biased towards the player, which they hopefully won't).
Gamasutra - Analysis: Game AI & Our Cheatin’ HeartsI mean, take a look at this thread. Guy's winning everything there is, barely losing, but every time the **** hits the fan, it must be the game being rigged.
2 -
Well, the game seems to somewhat simulate xG underperformance (at least with AI managed sides, e.g. this Galatasaray). By in-game xG, sHould have had 15 points more, but were relegated first season (cheating AI, clearly ;)) .
image.png.07495f20fc0219dbb5e310dbe1aa160e.png (3440×1482) (invisioncic.com)
image.png.c275facad9c8537fb4eb913d1a3a2057.png (3494×1210) (invisioncic.com)
One of the long-term "uses" of xG and similar models is an additional help when trying to gauge performance and whether it was good or bad luck behind current results (data fluent clubs such as Brentford have not renewed manager contracts after seemingly successful league campaigns because their analysis had concluded their performance hadn't warranted the good results, e.g. they were being lucky).Ideally, in-game xG would be able to do the same. If it were real football, things only even out in the longer run.... which a single season is anything but.
Opta Expected Goals - YouTube (the part about Juventus at the end)
Just how unlucky are Brighton? Expected goals analysis of Graham Potter's side (sportinglife.com)
Borussia Dortmund's crisis isn't a crisis at all, and stats prove they will rebound (espn.com) (Klopp was actually in big parts signed by LFC because their own analysis team had concluded he was having one of those seasons)
However, data alone are still but data alone. A good example of that is in this Man City analysis from Statsbomb from last season (their xG difference was above all, however, they still had issues when delving a little deeper, plus it was suggested that LFC weren't fully onpar as they were really effective at taking leads, so never had to push for higher xG).
How Much Do Manchester City Need to Overhaul? | StatsBomb
What does your own Expected Points table look like? And how do your Expected Goals total compare to your ACtual goals scored?0 -
4 hours ago, Mcfc1894 said:
I sure I seen an SI post saying the ME decides what it's going to do before it's shows you on the graphics engine and it only re calculates it self when you make a Substitute, Tactical change or a shout,
Yes, that's something different to what Outrospective claimed though. Basically, things are actually simulated before you get to see anything, so that the highlights picker knows where a highlight is. (If things weren't simulated in advance, the highlights wouldN't work). Naturally, tactical changes cause recalculations / resimulations from that minute on (because else those tactical changes would have no effect). That does not change that the ME is a kick by kick second by second simulation. Outrospective claimed the ME would decide that a goal was going to be scored beforehand, so fabricates completely arbitrary highlights to fit that (e.g. the player with a low rating in long shots arbitrarily scoring a screamer because the ME has decided a goal was now going to be scored).
It's completely the other way 'round. Goals are scored because players in the second by second sim have eventually done it. As that is still a semi-prominent FM Myth, no wonder some feel everything to be random, whilst others less so. You can't ever make logical decisions if you go with bad (and evidently) wrong theory that the match play you get to see would be merely random windows dressing. Such as some of the guides on these side suggest, e.g. Creating Tactics - The Book of Roles - Tactics, Training & Strategies Discussion - Sports Interactive Community (sigames.com)7 -
8 hours ago, Outrospective said:
What makes you think so?
So, how do instant result skins function?
My Critique of FM10 (long post) - Football Manager General Discussion - Sports Interactive Community (sigames.com) (Ov Collyer's post on page three).
[quote]The argument seems to be that the ME generates some arbitrary action to fit the fact a goal isn't going to be scored.
It simply doesn't work that way. The match engine doesn't know during the course of a move whether or not a goal is going to be scored until it happens or the chance is missed.
The only clause in the above is that sections of the match are run twice, once to determine if a goal gets scored and thus to select it as a highlight. But the second run through has no knowledge of the first run-through, generating the exact same logical and animation sequence as the first play through with exactly the same outcome. How do we do this? By what we call 'seeding the random number generator'. This means we can effectively 'rewind time' from the match engine's perspective by ensuring the same random number sequence used to influence the outcome during the first play through occur in the second play through, with the exact same outcome.
But the bottom line is when two players are running side by side in either play through, the ME doesn't know whether it's going to lead to a goal or not. The only entity that 'knows' this is the higher level highlights picker but that has no impact on in-match calculations, indeed the ME doesn't even know it exists.[/quote]
Plus own analysis / match reading. The ME were a complete bust too if it worked the way you suggested. I'd never been able to barely concede a goal past the 75th minute mark (as I made the team recycle the ball / sit on a lead, visibly) on a previous release if it were like that too. Plus, on the occasion, reporting bugs such as this marking bug: A player had instructed his wide backs (team in dark kits in possession at the start of the vid) to man mark the opposition wingers.
Sevilla FC vs Club Atl�tico de Madrid - 32 minutes - YouTubeYet, when his team drops the ball and the counter attack starts, the wide backs run BEHIND the wingers they are supposed to man mark, gifting them tons of acceleration space. On both left and right flank, they remain chasing the wingers all throughout the sequence. At the end of the sequence thus, they arrive completely unmarked in the box, both combine for the goal to be scored, and naturally, the opposition shot conversion went through the roof. What you see, including all the second by second on and off the ball positioning, is what's being simulated. A goal is scored when one of the sequences of play eventually leads to a goal. "Fixing" ME bugs thus likewise involves addressing this second by second behavior (in this case the bugged marking), otherwise, the opposition team would score easy goals over and over again.
I've never used instant result skins, but I reckon they work the same as if you'd holiday through a match: The assistant manager then takes over match management (which he can already do, and you can actaully have success that way, but to me it's not a tactical assistant proper).0 -
36 minutes ago, Outrospective said:
The game just calculates that one of the teams should score, so a LB with 3 in long shots pulls off Roberto Carlos all of a sudden.
This is not how the game has ever worked, neither back then nor now. The game has never calculated that a goal was scored beforehand, so fabricates arbitrary highlights of how a goal is being scored (the engine doesn't actually "know" a goal was going to be scored until the sequence of play is actually finished with a goal).
This is FM 101, really. Otherwise, players as well as AI managers would have never been able to actually influence the course of matches and results. Actually, it's a semi-common FM Myth from way back that all you get to see were arbitrary highlights. Whatever players you ever followed looking for advice given your struggles -- if any -- it were the wrong ones. I personally actually think it's quite worrying that these still exist. But then, back on early FM 2015 a significant portion of FM's userbase attributed the semi-regular hockey scorelines also to keepers being rubbish, rather than advanced players given an attack duty not at all tracking back to defend until this was patched, and many more (apparently a significant amount of players thus doesn't ever even watch at least a couple sequences of play from a half decent angle).
That's not a criticism, mind. I just find it worrying and I'm clueless outside of ever improving feedback how to better deal with this in the future. That feedback needs to be balanced though, as it may spoon-feed and make the game easier for players who prefer to connect the dots themselves. Maybe by making such feedback optional.
Additionally, a 3 in long shots has never meant a player could never score a screamer. In particular as these 1-20 ratings are 1-20 ratings on a "footballer" scale, not on a pub team player (1) to world class (20) scale. This goes some more so if the shooter finds himself in time and space. Naturally, in the longer run of a couple season/s, there should be a difference between the 1 and 20 player. (Accuracy of ranged shots shifts some in between releases though).
6 -
2 hours ago, Outrospective said:
I'vequit Torquay in the end, it's unbearable. When you watch the matches in such situations, it becomes painfully obvious, that ME shows just some caluclations result and nothing close to the real-time simulation. You either guess the required parameters or not. At least other strategy games are responding to your actual inputs and not show you the predetermined video based on previously pushed buttons.
Dunno about the disallowed goals (my sympathies). But you've been registered for ten years and still don't know how the game engine works (nothing like this at all. You're not alone, btw). I think this is a major issue in terms of (long-term) development. At least insofar: How could there ever be AI growth if after so many releases players get the game so fundamentally wrong? And how would you personally address this?
The game isn't documented very well, plus the tactical UI has pitfalls and plenty of "FM speech" yet not properly translated to football speech. It also remains questionable that the game as a sim of semi-pro football allows nonsensical input that has nothing to do with semi-pro football. But nowadays you've got at least channels such as Bustthenet (in prior years also a couple decent threads on how to read a match, analysis play etc). Plus more in-game match feedback and analysis tools than ever before (when ten years ago, there was pretty much zilch).Would you personally hire (tactical or otherwise) assistants if they were available and decent? The way I'd personally implement those is players still optionally having a say. Tactical for instance over the a) style of football played and b) match plan/management (e.g. ok, we've now scored the 2-0, assman, it may be time to put the foot off the gas / decrease our risk and be happy with what we have -- please put that to action now etc.) That way, as a player you can still judge or misjudge match situations and be burnt/rewarded for it rather than just leaving it all to an assistant so that all the match fun is gone.
That said, SI are never gonna make the game deliberately much easier/harder. It's not their thing (and they've always been happy with how things generally are.)1 -
Just now, duesouth said:
The big problem for SI is trying to cater to people playing the game very differently - from complaining when they lose 2 games in a season - to people who really struggle with tactics and keep getting fired, leading to giving up the game and therefore not buying future editions.
I've previously argued before that they are actually simulating the perfect job. Real managers delegate all kinds of stuff to assistants if they want to, either due to time restrictions, a lack of commitment to a certain area of their job or even knowledge (including tactics, see Klinsmann and Löw, ten Cate under Rijkaard at Barceclona, etc.).
Assistant managers are AI all the same. Therefore ever improving AI would benefit everybody.0 -
The indeed limited AI is still far more dynamic / reactive to match events than the average FM user from online/forum experience, which is why SI sadly seem very careful when improving it (or listening to criticism regarding that AI from the more tactically astute). The story is always the same, if the AI is capable of doing something (like challenging superior sides, scoring off fewer shots, staging comebacks, holding leads, which it is all very well capable of), you can actually do that better.
However, the game is sort of stuck in an AI rut, in that every time the AI can do something the average user can't, it must be "predetermined", and virtually nothing you can do about it. Additionally, as most players are fairly successful, there's the perception the AI must be this limited that the game can only provide prolonged challenges if there was a sort of hard coded balancing in that you are destined to lose matches in the most comical ways, etc. I think this is kind of dangerous and bad for possible future developments.
There was no suggestion of "cheating" mind; but this thread at some point might come close to entering familiar territory of kinds in particular once the right type of user reads it. A classic (mind, not actually written by me!).
(Conspiracy?) theory about this game - Football Manager General Discussion - Sports Interactive Community (sigames.com)
Point drops against a run of play and shock performances are meant to be a part of the game. There might be even bugs involved in the above match aside of the suspicious CB ratings (which is impossible to tell from stats, so a match maybe worth reporting as pkm). However, if "it" keeps happening at a high frequency, there's usually still something you can do about it. That aside, any mechanism such as complacency off-days as well as reputation shifts and thus opposition tactical approaches etc. affecting the human manager is also affecting the AI, so there's no edge for the AI to be had there, unless it likewise was better at managing those (indeed, AI top teams not seldom drop plenty enough points throughout a season).
3 -
9 hours ago, sthptngomad76 said:
Let's say that you were right though, and say the Programmers couldn't develop an engine that could counter all tactics- I would take an additional difficulty with 'rubber band logic' (overpowered AI etc) which, when balanced enough, could at least provide the illusion of a challenge. Surely that's attainable? It's been happening in other games for years.
I personally think that's a very bad suggestion. One of the strenghts of FM is indeed that it never had done such rubberbanding. Of course, this also means that there's a "one size for all" kind of AI, and with FM players you still have many who confuse having more shots/possession/(and now xG) was all there ever was to match management (the AI has always had an edge over any such player, as it actually reacts to events/scorelines/time left on the clock during a match with subs and tactical shifts).
Besides, there are numerous theories about AI cheating/rubberbanding as is. Funnily in parts because it's always that attainable/easy to just waltz the leagues (download a trial&error tested wildly popular and available AI/ME busting super tactic from someplace, push continue, consistently every season far outperform any player ability, climb the tables).
In fact, it's often the players who have continued success beyond all ability who go for such theory. The logic being: "My tactic / management is obviously this superior to the AI, and it's so easy to do too, all it can do to stop me winning the league from the get go with West Ham is my forwards having a couple additional scripted shockers, and the AI scoring with every few shots".
The irony is, of course, should the AI itself actually improve, the theory of scripting/rubberbanding/AI cheating would still increase alongside to it, as it would win additionally points from all players.edit: And speaking about super tactics, if the AI were to one day use knap tactics (/as suggested here), the game would be dead. Awkward to defend super tactics exploiting ME defensive issues and as such outperforming all player ability from the get go had been one of the reasons why FM Live is dead now (imagine you had built a decent squad over years, only for a couple newbies to the game to download an ME buster and lol over it right from the bat.
1 -
Arguably, "Mug player" says it all, probably. There are no "mug" players, they are competing at the same level as your team.
On a more serious note, as always, had you provided the save, somebody would have won this "unwinnable" match immediately. If anything, on FM you can have more control over match results than a real manager has (in parts as the game isn't quite as fluid as real football, and there are, of course, patterns, as the ME is a piece of code -- same for the opposing AI managers' decisions).
Nonetheless, my favorite "freak result" was losing to 3 DFK's on ca. FM 2015ish a couple years back. Can't do much about that, on the occasion, **** may happen.
Borussia Neunkirchen 2-3 SG Betzdorf - Match Highlights - YouTube1 -
3 hours ago, alian62 said:
Im hoping that after a new soccer manager game is released in June that things will change because its being touted as a rival to the FM series . Think it is call WE ARE FOOTBALL by Winning Games .... Made by some people who had connections in earlier FM games . FM needs good competition so hopefully this might do something
Slightly OT: It's made by former lead designer/s of Fifa Manager / On The Ball in Germany. There are no FM connctions. It will also be more similar to a director of football experience. The German market for which they are primarily developing has always prefered such experiences over tactical match simulations. As such, they also won't do a full match sim the way FM does it (text sim similar to Fifa Manager / On The Ball plus fairly abstract 2D highlights for the match day experience). It's also meant to be a more streamlined experience, where rather than managing individual matches, seasons are to be completed within a few hours at max. As a football sim, this won't be a competitor for FM.
Might still turn out to be a fun game, mind (and as the German market prefers such games, might lose them newfound customers in Germany again).1 -
As an aside, a mentioning of FM in this xg-Brighton piece.
Brighton & Hove Albion: The Enemy of Expected Goals (lastwordonsports.com)If you asked a random fan what they thought of Brighton this season, they would likely say ‘looks bright, usually fails to win’. This is of course a fair assessment of what has happened. What may surprise many, is quite how much they have lost by defying statistical probability this season.
Based on xG results, Brighton would be fifth in the league. Fifth. Where are they in actuality? 16th.
The thing about Brighton is that their problems are not particularly hard to identify. They embody every annoying Football Manager game ever, where chances are spurned and opposition goals are gifted.Some distant memories of Klopp's final season at Dortmund, sitting 18th place in February 2015 (when according to xG/xPoints, they should have been ~4th!).
Whilst in-game the stat may likely still need some tuning, what's the biggest xPoints difference you've seen so far (either AI or own?) This AI managed(and relegated) Galatasaray actually seemed to come quite close, trailing their xPoints by ~-15 points.
image.png.07495f20fc0219dbb5e310dbe1aa160e.png (3440×1482) (invisioncic.com)
image.png.c275facad9c8537fb4eb913d1a3a2057.png (3494×1210) (invisioncic.com)2 -
4 hours ago, sverige91 said:
AI having "few Chances" and taking them and "normal" human player still not scoring because of amount of chances "super tactics" create? - How do you even Come up with this logic?
It's very easy. Overall, these tactics DO SCORE more goals than any AI (or they wouldN't be this popular). The "bane" are the individual matches and management of those.
"Super tactics" are 100% attacking ones full of attack/support duty players. Additionally, due to the general success they bring, an AI manager most oftently in the longer term sees itself as the match underdog, so applies defensive tactics. Therefore, the AI always almost will have fewer shots.
The AI then ever wins matches with fewer shots because it is actually the team with the fewer shots. (Most of the time it doesn't). You can only ever score off few chances if you aim to have fewer chances (go through Man Cities matches in the last three seasons on Whoscored -- all of their opponents equally always score off far fewer shots and/ or lower xG). Sometimes that is luckily (or even off a bugged goal), sometimes not so much (as super tactics are this aggressive). Everything the AI can do, you could do better, that's factual -- however, super tactics are not set up to be that way.
An introduction and a question about (contextual) feedback and stats in the game - Tactics, Training & Strategies Discussion - Sports Interactive Community (sigames.com)As said, that's not defending the game whatsoever. This AI conspiracy stuff is nonsense. It's hindering (AI) development. Additionally, it distracts from the game's actual issues, which have always applied for both AI and player. The AI has never actually cared about the amount of chances in a match -- but rather the (current) scorelines and time left on the clock, and makes proactive changes throughout matches accordingly. That's why it on the occasion can actually score off fewer shots / lower xG. It however can't "read" a match as a better play would, which is where his / her edge is. I'd argue one of the keys of getting rid of these conspiracies is further improving match feedback, without alienating those for which this could make the game "too easy". Perhaps by optionally further assistant individual match advice, I dunno.
3 hours ago, sporadicsmiles said:None, but if you are seeing every single 1v1 being missed, you have some serious personnel issues.
I'd argue even in the game engine not every 1vs1 is 1vs1. This goes back as far as FM 2009, which had an issue with specific types.
For all its issues, even FM2020 had 1vs1s that were converted pretty regularly (above 50/50 rates). SI staff are on record as saying that their 1vs1 in 2021 used to be generally converted at above 50%, which is likely why there were so few/er complaints (the average xG for a 1vs1 in actuall football is ca. ~0.33, e.g. a 33% chance). Tactics have always influenced the angle of the 1vs1, the running distance, and more. Speaking about match feedback, xG should ideally help with creating / reading the better ones one day. (Also, please get rid of the CCCs).
3 -
4 hours ago, HUNT3R said:
My xG is 18.77 and I've scored 32 goals. That's a million times better than any other team in the league. Other teams are scoring around their xG totals.
This nonsense that the AI was this efficient (or even cheating) or that there was a balancing mechanism needs to stop either way, not for the sake of "defending" the game, but for the sake of hopefully future AI improvements.
The thing is, this Sverige guy is that he's getting riled up over individual matches on FM 2020 just the same -- the "super tactics" he downloads overall convert chances at better rates than any AI manager.The reason that in such individual matches it's always the AI on the "winning end" (e.g. ever scoring off few chances) is that the AI actually manages to have few chances. It is that simple.
As with with the WHUFC guy, these guys will never see this the other way around as their match management is fundamentally different to the AI's.
As of the "super tactics", in recent years I especially loved the crapfests that endured when they all tried to exploit the open centre of FM17 -- but an AI manager happened to plug just that zone of the pitch (by the virtue of research)... now imagine that AI were actually to one day become smart.
I really don't know what to do next - Page 8 - Tactics, Training & Strategies Discussion - Sports Interactive Community (sigames.com)1 -
On 17/01/2021 at 10:49, Viking said:
And no, I can obviously not learn it. I have played the game for some 25 years including CM, I have been on these forums for well over 20 years, and I've read everything I can read about tactics. Still confusing, ungraspable and scary. Must be that brain damage.
Back to CK III.
Whilst I sympathize with your struggles (with the results you've always claimed to be getting you must field your youth players into the 1st eleven or something equally obvious!): I lol'd.
Crusader Kings 3, Are We Dumb or is This Too Complex? | Listen Notes
This game is WAY too much :: Crusader Kings III Allgemeine Diskussionen (steamcommunity.com)
Help! Paradox games are too complex but I still buy them! :: Crusader Kings III Allgemeine Diskussionen (steamcommunity.com)
Crusader Kings 3 Made Me Feel Stupid But I Loved It Anyway - YouTube0 -
As my last comment on this, I also love it how downloading a super tactic, plugging it in, belting conintue and immediately challenging for the top with crappy West Brom is no possible issue at all to investigate -- whilst the AI managers every once in a blue moon scoring off fewer chances/shots immediately is such an issue to report.
This stuff needs to be challenged, no less as everything the AI can do, you could have always done better. However, every time the AI does something the user can't it must be a game flaw, e.g. no chance of AI improvements further, as with AI improvements SI risk alienating players (and if the AI WERE improved, it would become MORE efficient). Key is, you've got to try and play the game in a similar fashion the AI does, e.g. rather than managing overyl simplistic stats/ shot/chance volumes, managing scorelines and match play.
edit: This (type of) user btw. has been around for very long, and has been often times explicitly told what his issue is many times already. And as he says, it's not exclusive to FM 2021 at all, but that it has been like that since "forever".Additionally, for as long as nobody puts up a tactic that concedes zero shots during a season, one will concede goals always. Some of those will cost points. Like Guardiola, who's City team in both the 2017 and 2018 averaged barely 5 shots against, and still conceded over 20 goals, all the time off fewer shots, naturally. Which ones conceded off bugs may be worth reporting.
5
.png.9a7f1925db5404b03cf23655dbef96e7.png)
Ah the utter pointlessness of playing deep defensive football in FM
in Football Manager General Discussion
Posted · Edited by Svenc
It's completely the other way around. THe ME is a kick by kick simulation, with second by second on and off the ball movement. Whenever a sequence of play leads to a goal is a goal. The totals are whatever the toals are at the end of that.
Otherwise you'd never be able to influencne matches and the run of play, as better players clearly do.
Now as to deep defending, and it being hardish to do (may be true): If I'd be into conspiracy theories, I'd go more alongside the line of the following:
- human players hugely love attacking tactics with loads of goals, so they tweak the game in accordance to that.
- human players absolutely detest having loads of shots without scoring against an AI playing park the bus tactics, ditto (see also "getting FM'd") Still laods of complaints about that mind, curiously.
Don't know if true, but both of that has been the case ever since FM 1799.
If you want to do a quality analysis as to WHY deep defending was flawed, you had to analyze and You'd then have a chance of getting listened to. If you don't believe me, as an example from a previoius edition: This wide midfielder defensive positioning thread from ca. FM 17ish. SI had tweaked wide midfielders to almost always cover out the flanks, rarely coming inside to defend central areas (and help the centre mids). This influenced play -- and results.
Concerning tactical issue - Football Manager General Discussion - Sports Interactive Community (sigames.com)
Another example, FM 15 advanced players in top heavy formations given an attack duty not at all tracking back to defend so that there respective defenses were overrun every other attack, and if the attack was intercepted, a counter attack would start (hockey scorelines incoming) -- actually a good deal of player attributed this to keepers being "rubbisH" back then, revealing that apparently they barely ever watch the match play, maybe because of such wrong theories as yours.
Official Football Manager 2015 Feedback Thread 15.2.1 - Football Manager General Discussion - Sports Interactive Community (sigames.com)
Page 42 scrolling down as the link doesn't seem to lead to my own posts back then.
Each of those were luckily fixed in subsequent patches / editions, though.