Svenc
-
Posts
5,162 -
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
-
-
1 hour ago, DeanMon said:
Let me put it like this, SI does not intentionally cheat, but your imperfect ME (the black box whose operations are only revealed in anecdotes to the public) creates the favor to the AI.
AI makes all the PERFECT choices with the BEST effect: match preparation, TIs, pep talks, shouts,...
While to human, for example, with the same pep talks, sometimes it works to full team, sometimes it does not work to anyone.
There might be many more which I can't think of right now... But all the little details accumulate to an advantage for AI.
What do you say?
Let me rephrase this, and I'd probably agree.
The AI isn't as stupid as in older editions.
It won't leave its back doors uncovered on like every other attacking corner (check my first post in this thread and the little Leicester story).
It may not make stupid stuff like this anymore as oftenly.
It may also time its decision a bit more competitively (say when losing a match not merely finally start attacking too late into the match).
Therefore, its performance may have increased some (though players such as @Rashidi can still outperform it massively -- and do).
That's not rigging naturally, that's coding more competitive AI for some added challenge, in parts based on the feedback received throughout the years. Even if SI could (and they can't as that would require an AI that would be able to "read" the second by second match play just as a human can), Si would never code the perfect AI anyway. Coding the "perfect" AI would a) would alienate players b) make the purist main coder go all mad again due to the increase in cheating accusations.
0 -
31 minutes ago, Nacaw said:
Sorry, but no. I don't know exactly what arbitrary information is used to give a chance the label "CCC", but it's very clear that they are not all created equal. Sometimes my team will get what looks to me like a great chance, cross into the 6 yard box with no marking, and it's not counted as even a half chance.
In theory, over a fully Season and a breadth of different chances, this should even out some. There's a reason why the tactical forums (NOT the download section) has oft promoted the idea of "multi-dimensional" tactics, which is tactics that ideally see a variety of assists and finishers. This either way will be a case of the dynamic AI match management being more efficient as far as those flawed stats are concerned, and little more.
As argued, if the AI can do something the player cannot, the suspicious is immediately that there must be something fishy going on...
1 -
7 hours ago, Zemahh said:
Appreciate the shout-out!
I'd say part two was a clearer example of "FM'ing the FM" though (finished 6th as relegation favourites, with joint 1st place in Clean Sheets and 2nd fewest Goals Scored). I'm sure it'd be possible to create something even more extreme, but it does indeed show there's nothing "scripted" about defending well and scoring from the only shot you get.
The suspicious that there was is one of the reasons why SI are careful when improving their AI tactics and match management (the AI changes things around during a match according to scorelines, and may only attack for a couple minutes -- it REALLY doesn't care about who has the most shots come the end, and shouldn't). The main coder when he was still around on these boards is the most purist as you can get. And he took such accusations personal.
And yes, 9.9/10 FMers playing incredibly attacking tactics 24/7 oft not only exposing themselves but also compressing the space in the final third is the reason this never ever goes the other way (just take a look into the download sections to get a clue) . The rest is then being killed by the player overachievement and the AI reacting to changing Team reputations, e.g. the AI not playing exposing attacking football itself against the overachieving player anymore.
Interesting threads on playing defensive football, btw.0 -
Disclaimer: Reason for me posting this is that I want the AI of this game to become ever more intelligent. However, with more intelligent AI.... See, the reason this only ever goes in the AI's favor is nobody even bothers with similar ideas to this for long enough.
Back when the AI tactics barely protected ist own attacking Corners (good times, sigh!), with Leicester I had made half a season out of "rigging" the AI. One of the clues what was going on purely stats-wise is the oft much higher percentage of shots to shots on target for Leicester, but in terms of data the game offers, that's it.
Blame Match Of The Day and all that crap putting so much importance on shot Counts on a spread sheet (at least they have introduced xG, but too little too late).
That's all folks.1 -
24 minutes ago, Federico said:
Atalanta is the best attacking team in serie A. That was TRULY one game off.
1 -
5 hours ago, _mxrky said:
took me until the 95th minute to score and their goal was a penalty. Whilst their goal was a penalty (my player also missed a penalty in the game). I hate to say that the game is rigged, but sometimes the nonsense in this game is too e to be true. 6 clear cut chances and 3 half chances (one which was an open goal) only gives me 2 goals?
A part of the issue is what the game calls a "clear cut Chance". At least you scored not once but actually twice -- counting 9 finishes that would be classed "big Chance" in actual football analysis in this one.
2 -
Where the perceived randomness comes in (assuming it wasn't one of those games where every other shot goes flying in, re: long shots) is the AI, though. In the real world of football, this result would have turned heads. Every other manager would be eyeing what Everton were doing and trying to copy that a little. Furthermore, somebody may have figured out something long before Everton. Something at least a little more deadly than the mighty threat of apparently scoring zero Goals on average. It's just one match sure, but it seems doable, after all.
In FM, that doesn't come to pass.
What's more, the Everton manager didn't specifically set up to exploit a found weakness likewise, as the AI cannot do that… this too is about random chance, the opposition squad and the manager's attributes. So in a sense, the game world invites occasionally random swings, plus it promotes complacency when managing too. The result can be heads getting scratched.
Just like here.0 -
There's only one way to find out...
0 -
1 hour ago, yolixeya said:
I don't think so. Yes, most would say that if you would just post possession stats and shot Count
Problem is, as soon as a team drops off against an attacking side (it may be yours), that's an insta domination of both of those by total default. Improved feedback may not tell players exactly where they are going mediocre to wrong, which would make the game too easy (at the end of the day, this is all about increasing the chances of scoring against a specific opponent) -- it may give them hints however that there may be something that could be improved at all.
The CCCs and half chances were meant to solve this, but SI's subjectivity and glitches aside, it's obvious nobody knows exactly what they even are meant to encompass. That leaves players with the body language type of feeback, flagging forwards as frustrated/nervous once the scatterbrain approach of throwing bodies forward doesn't work and / or the crucial early/opening goal doesn't come. Even for something as simple as figuring out how many shots would be from genuinelly created from space and open play -- it's a lot of work and clicks to do that.
0 -
On 06/07/2020 at 17:37, Nacaw said:
It's already happening, Pep is literally complaining about what we would call "getting FM'ed" https://www.bbc.com/sport/football/53302856
"We played more than good but it is not enough to win the games,"
"We cannot say we are not a team who score goals this season. We are leading in goals. We create a lot of chances. We are a team who concede less - no team conceded as few chances as us but we lost a lot of games."
(Manchester City have faced the fewest shots in the Premier League with 249 - 59 fewer than Liverpool, who have conceded nine goals fewer)
Yeah, there's a reason why FiveThirtyEight's ranking has them sitting on top. And make no mistake, this City season would be reported as "getting rigged" by quite a few FM Players -- in particular considering that at the same time, Liverpool win Matches in sequence whilst barely having their final shot advantage.
https://projects.fivethirtyeight.com/soccer-predictions/premier-league/
As do raw xG rankings have them sitting up top.
https://understat.com/league/EPL/2019
Nevertheless this may also be another case of what we've been arguing for so long: FM's data just isn't sufficient to cover it all (and that may not be the case even with something aping xG). Because whilst raw xG Rankings may look fine, there may still be (defensive) issues if you digged a little further:
https://statsbomb.com/2020/03/how-much-do-manchester-city-need-to-overhaul/Nevertheless, I'd personally bet on them getting back into a closer title race the next season -- I've actually seen a few bookies listing them as favorites. Also, FM is still in need of more intelligent stats than what it has had now for so long. Plus, the data it has needs to be interpreted smarter, e.g. no final match reports arguing a side was "superior" simply because it had more shots to show on a spreadsheet come the final whistle, and so on.
2 -
7 hours ago, Federico said:
Honestly, I can't see anything wrong with that from a team comfortably leading the table. It's not that he played against Real Madrid you know. Duties or not, his team shoot on target 17 times (39 in total!!!) and this should not happen whatever your tactic.
It's not the duties, as linked to in my previous post, it's what is effectively being encouraged by pushing every single Player up, in particular against an opposition obviously sitting deep. I don't disagree with the shots -- it's just that if you went all experimental on this, the game has always been a case of: garbage in, ocasionally garbage out. With knap's for instance, for all the success and general results (which nobody doubts), you could have gone 60 shots without scoring way before FM 20.
Mind you, this a game flaw. Arguably, it's actually a design flaw, what with the game pretending to simulate actual management. However experimental/illogical/unrealistic/overly aggressive tactics have always tended to make the game look even "funnier".
1 -
1 hour ago, Federico said:
Blame Knap for that
Just a guess (2nd half of post, in particular the set piece part). Plus another tactic that sees no defend duties outside the centre backs applied 24/7. That kinda stuff has always lead to further… curiosities. Arguably the bigger issue still is the winning run with a subpar Team, rather than a one-off match, but heh.
0 -
9 hours ago, herne79 said:
But if we want to use some fancy system with a zillion attacking players and even more TIs, the AI can get overwhelmed because only one team on the pitch (the AI team) is trying to play football in a realistic manner.
And yet, a lot (not all) frustrations expressed can be oft traced back to just that kind of thing. The download sections for instance are usually filled with tactics that basically a) see no defenders outside the centre backs and b) have nobody occupying the centre of the pitch, which usually has an effect both on defending (speaks to reason) as well as attacking (no much space created if everybody sits atop of each other for lotsa additional set pieces / crap shots).
And depending how the AI lines up, (and how powerful the set piece exploits are) this can get quite nasty indeed.
This has been going on ever since the "collision avoidance" which forces players to run around markers, rather than ghosting through them. Prior to that the schtick was it to have a fistful players always staying behind the ball, and launching balls towards isolated attackers, which was too efficient due to that "ghosting". As it is: garbage in, occasionally garbage out.Not my problem though.
0 -
6 hours ago, Mandy42 said:
When you are playing in game against bottom 3 teams with massively negative goal differences, who somehow can keep your team out for 90 minutes and find a way to score low % goals to take points off you, it seems both unrealistic and unfair compared to the AI vs AI games who have no trouble scoring and winning.
1) Surely the first port of call was Looking at what those Ai was doing differently (SPOILER: It won't be rocket science though SI have received suggestions in the past.) And connected to 2) This isn't true as an universal experience as the AI drops a good amount of points, usually -- as recorded and complained about by many (in this very thread, by the way too, highlighting how even top Team shot conversions can be quite subpar). In particular if it doesn't have worldy squads to begin with (e.g. not PSG style squad Quality dominations in Ligue 1)...
0 -
57 minutes ago, Mandy42 said:
It doesn't seem logical though, why do all my errors come in games where I play against teams being more defensive?
It's logical insofar as the AI in these games can be this defensive, they only ever work shots from Errors, Long shots, set piecees, etc. as they are too defensive. Your Team makes Errors in These other Matches as well. Some of them may cost you Goals too. However, if AT ALL you concede a Goal against overly defensive Opposition, it will in tendency Always be from Errors, Long shots, set pieces and so on. Emphasis on AT ALL.
edit: sorry again for the random caps, Edge is a ****** browser with the worst Auto correction I've seen ever.
2 -
2 hours ago, themadsheep2001 said:
I do think the actual highlights need changing. Long maintained that showing a free kick tells you half the story, you need to see the foul too. If you're seeing that you're giving away cheap fouls on the same fast dribbler, that actually gives you information as opposed to cutting straight to the free kick swinging in your box. Even on comprehensive I don't see every yellow card. Better context means more informed decision making
Also how the AI (tactics) operate still, probably.
Take the perception of "Only ever losing due to bad luck, a worldy strike, an error etc. instead of the player being outplayed "
If the AI Plays at ist most defensive, that's all it's ever going to win by. That's just... logical.
That said, if there's anybody in internal testing who goes with the "game is unfair/cheating/rigged" theory, I'd demand him to be kicked out the door promptly. He's wasting coder's time and is holding the game back.
All SI MAY do is invite some of the better posters who insist on such theories to come in. And check why they feel that way (see above) and perhaps do something about it where reasonable -- obviously disallowing the AI from ever using "Let's park the bus, house and SI Offices in front of our Goal" tactics may not be realistic.
0 -
5 hours ago, themadsheep2001 said:
This, its all well documented. The AI isn't even that good tbh
And for as long as that suspicous is there the game wasn't playing fair, it never (overly) will be.
https://www.gamasutra.com/view/news/116071/Analysis_Game_AI__Our_Cheatin_Hearts.php0 -
9 hours ago, MBarbaric said:
This ME is as exploitable as any but who cares? it is single player game if someone wants to exploit the game, so what?
It also has multiplayer components though, plus the game is marketed as the "most realistic football sim game on the market". Now if Johnny McFootballfan consistently gets Huddersfield into Europe first Season or smashes the Ladbroke Premiership scoring 6.x Goals average, that'd probably make the game look worse than Fifa Manager. No, wait…
1 -
If you've always relied on download previous you're in for a rough ride. Simply put, with (the more popular) downloads, Player quality has never mattered much. Or anything else, for that matter. YOu can still test this: Don't do anything but putting the tactic in (not even make subs!), pick a sub-par Team, belt continue. Odds are big you'll be in Europe come the end of the season. If you go Cold Turkey here, you're playing an altogether different game from the Ground up. Stuff suddenly…… matters.
WIth that out of the way:10 hours ago, DamianJMcGrath said:I've read loads of things in the Tactics forum, such as the Pairs and Combinations guide, and I seem to have followed all the principles. But even with a good Tottenham side, I end up losing 3 or 4-0 to Brighton, Huddersfield, Leicester, and getting totally smashed by the bigger teams.
Do you have any Player, in particular the CBs on individual marking orders in their instructions? This pulls them all over the place. Virtually the only Thing that could cause this from my end. Except for really Basic mistakes, but if you've been to the tactics Forums, you may have those covered already.0 -
6 hours ago, MBarbaric said:
i wonder what is the reasoning behind it as that is mostly how tactics in real work. Is it exploitability?
IIRC a part of the reason why it was removed was also a debate regarding how positioning works in football. Is it that static that Teams basically employ two "formations" with/without the ball or is it more dynamic than that.
People will findexploitswhat works best
(even if that may not make an ounce of sense and on the occasion will lead to added frustrations) by sheer trial&error each edition anyways. However, it's all a part of the reason why it's gone too, for sure. Same reason why SI have never been in favor of a truly set piece editor (unlike former competitors).
0 -
1 hour ago, Federico said:
Do you also have a stat of shots from teams that played against Leicester in that season?
Slightly OT, but for totals (Season 2015/2016): 523 shots for created vs 517 against http://www.footstats.co.uk/index.cfm?task=league_shots
For individual Matches, you can go through all of their matches here: https://www.whoscored.com/Regions/252/Tournaments/2/Seasons/5826/England-Premier-League1 -
20 hours ago, XaW said:
Oh, and it's very possible to FM the AI. I just did it...
ThatÄS then not "FMing" though (e.g. losing to a Buggy Goal, or whatever) -- that's just counter attacking.
When Leicester won the league, they barely ever had much more shots. Actually, they won it with an average 13 shots vs 13!
Fm just doesn't have the data to Show it, e.g. this.0 -
1 hour ago, damicus said:
Currently Liverpool have the best conversion rate in the EPL at 27.8%. City are much lower at 18.8% and Norwich sit on a lowly 9.5%. ManU have the most shots on goal (317) with a conversion rate of 15%.
Agreed about there being too many shots (and one-sided Matches). But a shot conversion of 27.8% means that more than every fourth attempt total goes in. This is so wildly out of anything going on in Football that it could only be taken from Transfermarkt. I've Always wondered what it is they counted, because shots it isn't, and shots on target neither. Liverpool have a shot conversion of 14.3%, City 13.5, Norwich 7.3%. The EPL averages are just about 10%.
It is worth keeping in mind that all of these stats are the result of schooled personell managing. FM has also Always for better or worse allowed too much, bluntly put, "crap" you'd never see on a Football pitch, certainly not 24/7. We're talking totally beginner's mistakes in anything when it comes to even team sports basics. I mean, the entire download section so popular is FILLED with tactics that basically see no defenders outside the centre backs.
Speaking of which:On 23/06/2020 at 20:18, Sunstrikuuu said:It was a while ago that I found it, but I suspect you're right. Their shot numbers are quite low, but they're very explicit that they're not counting just shots on target.
Have you found a Definition of their "shots at Goal"? stat somewhere? At one Point I even considered Mailing them.
0 -
And speaking of FM 2015/2016...
if that doesn't make sense to you, and how this couldN't be "undeserved" the game never will (I was Leicester).
0
(even if that may not make an ounce of sense and on the occasion will lead to added frustrations) by sheer trial&error each edition anyways. However, it's all a part of the reason why it's gone too, for sure. Same reason why SI have never been in favor of a truly set piece editor (unlike former competitors). 
Call for Discussion:Does Home Advantage Really Exist in FM?
in Football Manager General Discussion
Posted
I think the biggest contributing factors are about decently understood by people actually earning money off of this, e.g. Matthew Benham and Company. He had argued that travel distance can also be a thing IIRC, hence why odds look a little different in the MLS, plus it isn't super rare for even the winner of the conference to have a negative away record (Portland Timbers WDL 4 4 9 in 2017, for instance).
That's something that doesn't appear modelled into the game, as the MLS (Brasileira too) unlike actual Football doesn't appear much different from the rest.