Jump to content
Sports Interactive Community
Totalfootballfan

Team Statistics in real life vs Team Statistic in FM

Recommended Posts

Yes, here, AI has always been limited on that front. The argument is often brought about that some statistics in FM work differently, and indeed, they often do. Many stats sides (not all, not sure about Whoscored) count possession by the number of passes played. FM goes by a chess clock method whereby the time a team is in possession is measured, rather than the passes. Squawka does collect similar AFAIK. Either case, if FM would collect the number of passes, in some cases, statistically the bias for possession-based managers would arguably show even less strongly. There was also an defending issue prior imo that allowed weak teams to dwell on the ball too easily, or any team really, not sure if that's been fixed.

That said, this looks like the competition weren't set to full match details at a quick glance, as no team seems to average more than 400 passes per match (could be wrong). The numbers on the quick sim, which isn't actually fully played out matches, have been in need of upgrading for quite some time. The number of goals is also typically lower, despite the AI also not being as good as a more experienced human player here, naturally.

Edited by Svenc

Share this post


Link to post
Share on other sites

I'd say that only 69 goals and 50.78% possession for Barca is quite unrealistic... I hope that it'll be looked at for the release

Share this post


Link to post
Share on other sites
Guest El Payaso
1 hour ago, Totalfootballfan said:

So I'm curious what you guys think about that… does it bother you or you don’t pay any attention to that difference? :ackter:

Been the case for several iterations so it doesn't really surprise me anymore. Also for me a really big deal in terms of the gaming experience as I usually play just in La Liga and would hope Barcelona and Real Madrid to be the fortresses that I will never break through during my career and usually I do it already in the second season with basically any team I select. It's also really odd considering how easy it is for human users to really get those world class players going. While with the AI you don't see Real Madrid or Barcelona being any better than Liverpool and Tottenham while IRL this gap is major. 

Strong tactical identity doesn't only apply to these major teams but also to smaller ones. For example with Las Palmas they were one of the best in terms of possession last term and that is also required as you take over as their manager while I can see in your SCs that they either aren't dominating possession in the game. The downside though with the smaller teams is that for example Las Palmas don't even necessarily benefit from large amount of possession like Barcelona as they do not necessarily possess the quality to use the possession and it ends up just being a stat instead of clear benefit. I think on FM with possession you benefit a bit too much as the AMs and STs are not set to defend when not in possession.

But yeah, summarized it is super important to me and makes the gaming experience and the ME a bit of boring. 

Share this post


Link to post
Share on other sites

Here's Messi's statistic from the save

He made 35 Appearances in LaLiga and scored 13 goals which is about 2-4 times less than his average goals per season in the league in real life

Messi.png

 

 

 

Here's Ronaldo's statistic from the save

He made 27 Appearances in LaLiga and scored 8 goals which is about 2-4 times less than his average goals per season in the league in real life

Ronaldo.png

 

Share this post


Link to post
Share on other sites
Guest El Payaso

@Totalfootballfan just by that I think that I wait until full release before I start anything serious. Quite game breaking if you manage in Spain. 

Share this post


Link to post
Share on other sites

So to clarify did you run the league in full match or quick match engine? As that'll have a substantial bearing on the result? 

And if you run a test like this surely you've come to a conclusion? If you think it's a problem then please raise it in the bugs forum. But as said we'd need the above information. 

Share this post


Link to post
Share on other sites
14 minutes ago, Neil Brock said:

So to clarify did you run the league in full match or quick match engine? As that'll have a substantial bearing on the result? 

And if you run a test like this surely you've come to a conclusion? If you think it's a problem then please raise it in the bugs forum. But as said we'd need the above information. 

 

Neil,

I loaded 2 Leagues: English Premier League and Spanish League BBVA

I picked Girona to manage in Spanish League BBVA and I picked Huddersfield to manage in English Premier League  then I went “On Holiday” until the end of the season in both leagues

I guess in that case the leagues should run in “Full Match Mode”… Am I wrong?

Share this post


Link to post
Share on other sites

Did you get sacked by either team? Why don't you raise an issue in the bugs forum with relevant save games if you think there's an issue? Thanks. 

Share this post


Link to post
Share on other sites

The likes of Messi, Ronaldo, Madrid and Barcelona are quite extreme. Sure, it would be great for EVERYTHING to better mirror real life, but the reality is that things like reputation mean that "lesser" clubs will play lower tempo Mentalities against these sides, and the chess-clock method of calculating FM possession has a corresponding effect.

This also means that the output of these sides is reduced (deeper, more compact defensive shape faced), so you see fewer goals / assists. If, on the other hand, people want Barca and Real to be spanking teams, then they'll possibly end up with an imbalanced game world in which Barca and Real look fine, but everyone else is below par.

Of course everyone wants things to be perfect, but this has always been one of those trickier areas to "get right".

Share this post


Link to post
Share on other sites
1 minute ago, RTHerringbone said:

The likes of Messi, Ronaldo, Madrid and Barcelona are quite extreme. Sure, it would be great for EVERYTHING to better mirror real life, but the reality is that things like reputation mean that "lesser" clubs will play lower tempo Mentalities against these sides, and the chess-clock method of calculating FM possession has a corresponding effect.

This also means that the output of these sides is reduced, so you see fewer goals / assists. If, on the other hand, people want Barca and Real to be spanking teams, then they'll possibly end up with an imbalanced game world in which Barca and Real look fine, but everyone else is below par.

Of course everyone wants things to be perfect, but this has always been one of those trickier areas to "get right".

I agree with this. In fact I'd go as far as to say Ronaldo and Messi's real life statistics are miles more unrealistic than what they are in game! Even though it's happening in real life. In the modern game, their goals to games ratio is hugely freakish. If FM isn't simulating this to save unbalancing the game worlds, then this wouldn't be an issue for me. 

Share this post


Link to post
Share on other sites
28 minutes ago, Neil Brock said:

Did you get sacked by either team? Why don't you raise an issue in the bugs forum with relevant save games if you think there's an issue? Thanks. 

 

Neil… sure, I checked it… when I went back from “On Holiday” I found myself still managing Huddersfield

Girona sacked me but it happened just 3 matches before the end of the season so it won’t have any significant impact on the results.

I didn’t post it in the bugs forum because it might be that I’m the only weirdo :kriss: who considers something like that as an issue so I wanted to hear what other people think about it because if nobody cares about it then there’s no point in raising in the bug forum :) 

Share this post


Link to post
Share on other sites

Bugs aren't really decided by democracy. It's either an issue or it isn't. Best leave it to our QA team and Devs to decide after you've raised it in the place where they can see it. Many thanks. 

Share this post


Link to post
Share on other sites
5 hours ago, RTHerringbone said:

The likes of Messi, Ronaldo, Madrid and Barcelona are quite extreme. Sure, it would be great for EVERYTHING to better mirror real life, but the reality is that things like reputation mean that "lesser" clubs will play lower tempo Mentalities against these sides, and the chess-clock method of calculating FM possession has a corresponding effect.

THe numbers often would be lower in FM if FM didn't use a chess clock, but the numbes of passes getting played. THat said, in prior editions it's been very possible to get Messi/Ronaldo to score a goal+ per match, however, similar as in real football, you had them to average a decent 5/6 attempts per match also, which AI was often incapable of too (Messi/Ronaldo average 5 / 7 attempts a match each in real football, which is thrice the amount of an average forward). Those numbers look pretty interesting in general, no least because the "dribblings" seem to have reached an all time record low at 3 per match (Ronaldo below 2 in La Liga!), whereas on prior releases the "super dribblers" averaged easily 5/6+. Consider me intrigued on the topic either way, no least because I'm curious as to what the AI may be doing here if things didn't go "quick sim". ;)  On FM 2016 Bayern oft struggled to even approach 2 goals per match with one of the most domestically dominant squads in that iteration, and that has nought to do with the players being incapable of that, but all with AI completely [no ME busting tactics required to get better numbers too]

Edited by Svenc

Share this post


Link to post
Share on other sites

Apart of how players are individually being utilized -- really old patch releases revealed that older releases at least had AI working on variables such as "expected scorelines". These would big time influence what each respective AI manager in each match would target as far as the result could go, could be a 0-0, a big win, anything. As of the aforementioned Bayern example (with them on occasionally coming out with no more than 50 Bundesliga goals on the odd occasion): Whilst personally I thought there were some dubious duty allocations involved in then Guardiola's 4-1-4-1 ---  was it ever checked if such top sides would be too "content" with just wins, and never much aim for the "thrashing"? Managing a league dominating side, if you wanted to yourself, you could spam 1-0, 2-1- etc. victories in sequency pretty easily, after all. All you would need would be to "put the foot off the gas" after going in front (obviously, a simple switch in "mentality" only goes so far, which is still a common newbie misconception, but that's another topic). Such things hugely influence scorelines coming about, no least because AI visibly also don't mererly change mentality in their match management, they change the number of players they throw forward -- or indeed, don't. Primarily role/duty area this, has always been.

In prior releases I edited every AI manager to prefer top heavy formations for instance, and leveled the club's reputations. What this meant initially anyway was that even the poorer teams in unbalanced leagues would have a go at the bigger guys. Bayern immediately scored like 30 goals more, and naturally, the amount of goals increased in general accross the league. Part of the enquiring on this is that from my experience, it's never much taken tactical genius to outperform the big AI boys in this regard most of the time. The totally basic stuff would do. By FM 2016 I used to joke that any new player to Football Manager could outperform Guardiola easily, as far as that it concerned. To be blunt, simply fielding those far superior players was plenty. However, if AI mangers themselves wouldn't even much target for big wins prior to kick-off, then that would be fairly understandable. Naturally, every AI manager also has different traits which may/should influence this. Some may be more concerned just getting the wins, whilst others may be more of the more adventurous/attacking kind. Question is, what are Barcelona / Real in-game aiming for these days, actually?

Hopefully TotalFootballFan may provide the stuff to SI, rather than just recording videos and then posting his "findings" on FMBase either way.

Edited by Svenc

Share this post


Link to post
Share on other sites
On 10/27/2017 at 19:38, RTHerringbone said:

but the reality is that things like reputation mean that "lesser" clubs will play lower tempo Mentalities against these sides, and the chess-clock method of calculating FM possession has a corresponding effect.

Is the issue possibly the fact that the players are apparently good enough to keep possession for so long in lower tempo mentalities? Shouldn't these players be turning over possession more? I've occasionally seen the AI in 17 play a stacked midfield and just pass it about, racking up 1,000 passes, and it seemed near enough impossible to take the ball off them. This includes like, teams from very low down the tier system, and they're *very very* good at keeping that ball, obnoxiously so, which feels a bit odd. (Tactics aside, obviously.)

Share this post


Link to post
Share on other sites

I just finished the 1st season with Getafe and in my save Barcelona scored 65 goals in the season… LOL :eek:

Share this post


Link to post
Share on other sites
Guest El Payaso
On 27.10.2017 at 20:05, Svenc said:

THe numbers often would be lower in FM if FM didn't use a chess clock, but the numbes of passes getting played. THat said, in prior editions it's been very possible to get Messi/Ronaldo to score a goal+ per match, however, similar as in real football, you had them to average a decent 5/6 attempts per match also, which AI was often incapable of too (Messi/Ronaldo average 5 / 7 attempts a match each in real football, which is thrice the amount of an average forward). Those numbers look pretty interesting in general, no least because the "dribblings" seem to have reached an all time record low at 3 per match (Ronaldo below 2 in La Liga!), whereas on prior releases the "super dribblers" averaged easily 5/6+. Consider me intrigued on the topic either way, no least because I'm curious as to what the AI may be doing here if things didn't go "quick sim". ;)  On FM 2016 Bayern oft struggled to even approach 2 goals per match with one of the most domestically dominant squads in that iteration, and that has nought to do with the players being incapable of that, but all with AI completely [no ME busting tactics required to get better numbers too]

Yeah I think that still in FM 2015 Messi and Ronaldo possibly could get close to the amount of goals they score in real life but if the amounts are like that on FM 2018 it's really worrying and at least for me it will prevent starting a save. 

Even though on FM 2016 I found Bayern as the team that actually was hard to beat, you are correct. Something similar still might be happening in the game also in 2018. In FM 2016 Guardiola used a 4-1-4-1 formation which was defensively really solid but like you said, lacked scored goals. I noticed something same when I played FM 2018 and faced Atletico Madrid. They use 4-4-2 setup and with that setup their problem was that the midfield wholly was really slow to shift forward as they basically always played the ball high up to their strikers who ended up having really slow support and even when a winger was dribbling the ball to danger area they basically always attacked with three players against my 7 players grouped to defend. In real life you wouldn't create anything with that but on FM they ended up scoring couple but in general I feel that their attacking was totally rubbish and looked also really bad in the ME. I think that on FM 2016 Guardiola might have struggled with this too and probably it was even worse as he was only selecting one striker. 

This once again backs my (and many others') opinion that the strikers should drop deep when the team is defending and by that they would not be that isolated from the rest of the team and the football produced would look lot different. And this possibly could make teams like Barcelona and Real Madrid better as they could be more devastating in quick breaks as they would have more players rushing forward when the opportunity rises. 

In general Real Madrid and Barcelona are the two teams that certainly should be in category one and clearly better than for example Liverpool and Tottenham or even Chelsea. I'm not saying it's easy but there simply needs to be bigger difference between Barcelona and Chelsea and Messi compared to for example Hazard. 

Share this post


Link to post
Share on other sites

My guess is this is due to world class striker missing too many chances. The conversion rate between a good striker and a top striker is not significant enough.

Share this post


Link to post
Share on other sites

@ElPayaso AI Guardiola also tended to isolate that lone striker fairly frequently too... a thing to watch that any newbie guide of this had covered for half a decade too (basically, sticking to such guides would make it almost impossible). This is no more when he moved over to England, as the City research had given him another formation entirelly. (Another curious , but vry natural FM "specialty" -- Players and managers may shift traits based on who's assessing them). :D

 

38 minutes ago, russell9 said:

My guess is this is due to world class striker missing too many chances. The conversion rate between a good striker and a top striker is not significant enough.


Purely stats-wise, the gap in finishing the end product isn't that big. Forwards convert about every fifth attempt they take. The difference in between players isn't merely a huge gap in how clinical they are, it is how often they get into decent position or being able to carve out additionally chances for themselves via their superior movement and on the ball skills. Therefore, to get Messi to break the records, on prior releases you would need to get him up to averaging 6 attempts per game (his real life: 5+, Ronaldo 7). Perfectlyl possible. Actually, that's even better rates than he typically has, but an older release.


av5rmo.jpg

 

You can actually insert columns into squad screens how scoring attempts are spread in between players. On FM 2017, stats wise this was pretty bang-on, btw.

Also, as AI managed Bayern could be outperformed so "easily" prior too, I'm primarily eyeying the dynamic match management of AI -- do they lower their gears too readily after taking leads rather encouraging their sides to push on? However, multiple players are already reporting that top forwards (AI managed, full sim too) would just score 15 goals on the occasion (Kane, etc.). Question is, how many attempts did they take? In the above Messi profiles posted by TotalFootballFan, it seems he took an average of 3.5 shots per match, which is a fraction of his real life numbers, and that Ronaldo page doesn't make sense based on any prior experience with the game. If he was primarily used as a central forward, his goal numbers are much too low. If he was used on the wings --- less than 2 dribbles per match !!!, also doesn't add up based on previous releases.

What I do know is that better players, same as the old Messi example, seem to get their forwards score consistently, averaging a goal per match from 4 attempts per match , which purely statistically, is about in line with football, it's a bit better than the averages even. So is this an AI "issue" or is it an ME issue if such forwards would score too few?

2.png

Edited by Svenc

Share this post


Link to post
Share on other sites

I played 3 times the 1st season in English Premier League and Man City finished the 6th place the 1st time, the 7th place the 2nd time and the 5th place the 3rd time... I'd say that's a highly unrealistic result and also, every time Man City scored less goals than Everton ... :eek:

Edited by Malibden

Share this post


Link to post
Share on other sites
On 29/10/2017 at 22:52, Svenc said:

Purely stats-wise, the gap in finishing the end product isn't that big. Forwards convert about every fifth attempt they take. The difference in between players isn't merely a huge gap in how clinical they are, it is how often they get into decent position or being able to carve out additionally chances for themselves via their superior movement and on the ball skills.

Regardless of stats it doesn't make sense to slap an average conversion rate onto every player on earth. The fact is that some forwards are better at finishing than others, noticeably so. Which is why I don't really like the conversion rates argument used to justify poor finishing. If the game is coded to be biased towards these rates it would be a flaw imo.

Share this post


Link to post
Share on other sites
4 hours ago, Cap'nRad said:

Regardless of stats it doesn't make sense to slap an average conversion rate onto every player on earth. The fact is that some forwards are better at finishing than others, noticeably so. Which is why I don't really like the conversion rates argument used to justify poor finishing. If the game is coded to be biased towards these rates it would be a flaw imo.


It's not as simple as that, indeed. There's forwards that completely rely on tap-ins etc. falling to them (the Poacher type), others make space all for themselves, a few thrive off counter attacks when they have space to surge into, others are a more static type, and there's naturally also finishing technique, which a computer sim imo will always struggle some to replicate (though FM arguably has a decent mechanics in the PPMs in place for that which could be expanded/better utilized). However, whilst they may hide the specifics, those numbers long-term are no lie either. On any competitive senior level in any sports, the difference is settled in slight margins and percentages, not in decisively advantages. Anything hugely sticking out you're more likely to see in a player's youth level progress, as that's oft not competitive levels due to huge gaps in quality in between teams and players. e.g. players starting out at their local youth team before getting scouted, scoring at complete fantasy rates by any actually competitive level.

What's oft forgotten is that keepers on their respective level tend to be competitive too and that most scenarios sees them in a significant edge over the forward as defending is easier than attacking. All the keeper needs to do is to deflect the shot for it not getting converted, whilst the forward has to trap the ball at typically some pace, control it, get it on target in a way that is also a challengenge for the keeper, all whilst typically being pressured / having a limited time frame to make his decision. I think that's worth pointing out too as the logics oft is unless a keeper would be world class he wouldn't be able to stop the world class forwards, however the keepers that tend to compete against world class players still tend to be top level keepers rather than semi-amateurs or something (even if they on occasion may go by the nick names of "Callamity James" around the pub). If they weren't competitive, no manager would dare to field them / no club dare to sign or keep them long. If clubs would keep such players, they'd promptly be relegated to less competitive levels, after all. Such rarely comes to pass, the closest peek we get are instances where clubs are relegated not due to their results on the pitch, but different means. E.g. Juventus in Serie B a couple seasons ago. but as that was but one level below, they didn't break all records.


On topic, possession numbers have always been difficult for AI anyways, even for AI managerse who have such traits edited into the db (this is connected to how AI works.) However, those are stats that may at best tell you something about a style of playing anyway. I found more recent you're going to find more curious if you take a look into the shot rankings, in particular of "unbalanced" leagues and compare (practically every top division in Europe), as they are actually indicicative of some performance. Football isn't about passing the ball around the park, it's about scoring goals.

Edited by Svenc

Share this post


Link to post
Share on other sites

As much as i enjoy FM the gap between the sim and real football has increased exponentially. Top team now favor possesion based football and most teams that are inferior will play very cautious formations in which they give posession away easily. 

In football manager defensive or counter mentality will sometimes have more possesion than an attacking mentality which should never be the case.

 

the fact that teams like Barcelona average a 50% possesion per season should really give the developers some food for thought. The ME still has a lot of work before it can be even remotely close to real life football

Edited by andu1

Share this post


Link to post
Share on other sites
5 minutes ago, andu1 said:

In football manager defensive or counter mentality will sometimes have more possesion than an attacking mentality which should never be the case.

 

Tottenham - Liverpool three weeks ago: 36/64% and 4-1 the final score. In this game, who was on the defensive and used a counter mentality?

Share this post


Link to post
Share on other sites
56 minutes ago, Ji-Sung Park said:

Tottenham - Liverpool three weeks ago: 36/64% and 4-1 the final score. In this game, who was on the defensive and used a counter mentality?

the point of this topic is that top teams do not have a possession football that resembles real life. I think it's down to the ME as I tried to thinker with managers stats with the ingame editor and i couldn't improve the possesion of their teams one bit.

Share this post


Link to post
Share on other sites
2 minutes ago, andu1 said:

the point of this topic is that top teams do not have a possession football that resembles real life. I think it's down to the ME as I tried to thinker with managers stats with the ingame editor and i couldn't improve the possesion of their teams one bit.

Your statement to which I replied is false. High possession should not be equated to "attacking" football. This is a hard-to-kill myth going around.

Share this post


Link to post
Share on other sites
7 minutes ago, Ji-Sung Park said:

Your statement to which I replied is false. High possession should not be equated to "attacking" football. This is a hard-to-kill myth going around.

indeed, but that has nothing to do with the issue where teams like man city, barcelona... play nothing like their real counterparts. and, it kills the game for some people.

Share this post


Link to post
Share on other sites
16 minutes ago, Ji-Sung Park said:

Your statement to which I replied is false. High possession should not be equated to "attacking" football. This is a hard-to-kill myth going around.

Again, the discussion was about the average possession that top teams have in game. Attacking football  is subjective , but one thing that most top teams have in common in real life is that when they face an inferior team they tend to have superior possession stats than in game..

Share this post


Link to post
Share on other sites
26 minutes ago, andu1 said:

Again, the discussion was about the average possession that top teams have in game. Attacking football  is subjective , but one thing that most top teams have in common in real life is that when they face an inferior team they tend to have superior possession stats than in game..

Alright andu1, that's enough water cooler talk for today, I have a pile of 96 scout reports to sift through.

Share this post


Link to post
Share on other sites
10 hours ago, Svenc said:

It's not as simple as that, indeed. There's forwards that completely rely on tap-ins etc. falling to them (the Poacher type), others make space all for themselves, a few thrive off counter attacks when they have space to surge into, others are a more static type, and there's naturally also finishing technique, which a computer sim imo will always struggle some to replicate (though FM arguably has a decent mechanics in the PPMs in place for that which could be expanded/better utilized). However, whilst they may hide the specifics, those numbers long-term are no lie either. On any competitive senior level in any sports, the difference is settled in slight margins and percentages, not in decisively advantages. Anything hugely sticking out you're more likely to see in a player's youth level progress, as that's oft not competitive levels due to huge gaps in quality in between teams and players. e.g. players starting out at their local youth team before getting scouted, scoring at complete fantasy rates by any actually competitive level.

What's oft forgotten is that keepers on their respective level tend to be competitive too and that most scenarios sees them in a significant edge over the forward as defending is easier than attacking. All the keeper needs to do is to deflect the shot for it not getting converted, whilst the forward has to trap the ball at typically some pace, control it, get it on target in a way that is also a challengenge for the keeper, all whilst typically being pressured / having a limited time frame to make his decision. I think that's worth pointing out too as the logics oft is unless a keeper would be world class he wouldn't be able to stop the world class forwards, however the keepers that tend to compete against world class players still tend to be top level keepers rather than semi-amateurs or something (even if they on occasion may go by the nick names of "Callamity James" around the pub). If they weren't competitive, no manager would dare to field them / no club dare to sign or keep them long. If clubs would keep such players, they'd promptly be relegated to less competitive levels, after all. Such rarely comes to pass, the closest peek we get are instances where clubs are relegated not due to their results on the pitch, but different means. E.g. Juventus in Serie B a couple seasons ago. but as that was but one level below, they didn't break all records.

Yeah, the fact that most top level goalkeepers have excellent shot stopping is a very good argument for conversion rates and one that I like better, as it shows the balance of the game. And like you said, in fm strikers do play and finish differently based on their stats and ppms. I think this is a better, more accurate way to explain poor finishing, as it mirrors real life. 

 

As for possession, I think the m.e lacks a few things which influence possession in real life. Teams with high mentalities in fm usually go more direct in the buildup, and slow down in the final third, while irl it's the opposite. The directness of play often results in lower possession numbers. And there's this really obnoxious thing in fm where teams with defensive mentalities or "retain possession" ti are so hard to dispossess, even if you've isolated their player on the ball. In truth it's a lot of m.e stuff that results in the weird possession and passing numbers so until there's a new m.e and the more attacking mentalities are coded to be more possession based, less direct, and more realistic, the issue will probably still remain.

 

Edited by Cap'nRad

Share this post


Link to post
Share on other sites

The unrealistic possession numbers isn’t the only issue.

Just look at how many goals Barcelona and Real Madrid scored in the season, it’s like 2 times less than these teams usually score in real life, also, the fact that Atletico Madrid scored more than Barca and Real is ridiculous...

Share this post


Link to post
Share on other sites
8 hours ago, Karimza said:

The unrealistic possession numbers isn’t the only issue.

Just look at how many goals Barcelona and Real Madrid scored in the season, it’s like 2 times less than these teams usually score in real life, also, the fact that Atletico Madrid scored more than Barca and Real is ridiculous...


Goals. As hinted above: To gauge the latter you have to go a step back on that. To simplify, what increases the chances of scoring goals? Getting attempts off. It's arguably not totally new to FM 2018, but thanks to the Shot rankings (and numerous sites that publish such stats in real football), we now know:

1) In unbalanced leagues the top teams are completely running away here (with the exception of Atletico, who excel in the counter piece to this, which is the shots they concede, which FM doesn't have a stats for).
2) On FM, such utterly domination in attempts (AI managed, either way) rarely comes to pass, so how could they score significantly more goals? IN-game too, the average SOT conversion is about roughly 1 in 3-4, and so in this full match sim both Barcelona and Real hovered around the 25 goals mark scored from 17 matches. I've posted a few in the impression thread, including Celtic, which in real football run away off the league unlike any other side in any Top division (that is, in real football , in-game Aberdeen oft surpass them)

OylRDt9.jpg

There can be a multitude of reasons for that. Tighter defending, the balance in between player quality and tactics getting a few adjusted (e.g. superior players don't tore defenses a new one simply by their presence as such anymore), but also Ai tactics (AI managers shift gears during matches accordingly to scorelines, so on occasion may too readily shift a gear down after taking leads). On FM's level of play, I'd still be somewhat suspicious off such rankings, no less because it has (arguably) too easily been possible to "spam" "on target attempts" from set pieces, which would be no sign of superiority/outplaying anybody as such, and mostly they naturally aren't converted (headers under pressure in packed boxes are tough). Still something that is way more curious than the possession numbers, as it is indicative of some key performance data, rather than just a "style of playing". Not only does the advantage melt regularly (IRL: 7+ shots on target for the top vs 4 for the league averge). The likes of Eibar can record just as many shots as the top sides (oft about 6 per match). Ronaldo, btw. took 53 shots in those 17 matches, which is less than half of what he gets in real football too. As I have already seen AI forwards on surpassing 5 attempts per match (where they would also score consistently), that's at least somewhat connected to how he's being utilized here too.



As for possession. Same as any stat (the shots on target too), in the end, is is a symptom, not a cause. The same that goes in the last paragraph about shots on target / FOR can hold true here. Some if it relates to inherent ME play. I'd argue that the ball has traditionally been quickly played out of the back to rarely move back has as much of an influence here as anything. As also noted by @El Payaso, in real football, in particular for possession based sides, the centre backs collect some of the most passes on the pitch, rather than guys up front. On FM this never much comes to pass, as a) never do CBs much recycle/shift the ball between them. And b) once it's played out of the back, it rarely moves back again. Cautious needs to be applied to other areas too. From my experience, the pressure on the ball carrier has been some improved, which includes closing down of the keeper and midfielders.

As such, sides would find it harder to amass huge possession numbers (in my brief playing experience, I too found that my previous ploys of simply having a few guys sitting deeper wasn't enough anymore for record numbers on every match). Compare this to early FM 2016 in the screenshot below, where pressing in particular in deep areas around teh OP backline was this toned, much worse sides wasted time keeping the ball in their backline. As possession numbers aren't in any kind of way in itself an indicator of performance (unlike SHOTS, which should be a much bigger concern here, see above). Possession in itself means jack.

Z0taOYK.jpg

It's still an issue. Some of it will be connected to how AI operates -- you can get a possession based style going on any mentality btw. [speaking of which, attacking mentalities aren't inherently about direct play, however, and this is semantical as much as limitation: attacking football arguably is about getting the ball goalside to score, and as such it's been implemented into FM for the more attacking mentalities]. Some of the possession "issue" may relate to FM 18 ME inherent traits (already seen other guys posting how they struggled to impose their game as previous against lesser sides, in particular away). The tactical UI still needs an overhaul, and I'm not convinced that SI did make things easier for the AI a couple releases ago, as they have complicated what "duty" governs too (basically, support duty guys are a few more likely to recycle the ball, attack duty guys a few more likely to get the ball forward, when that prior never was a thing expect for one shape setting). Unless that was done in line with how AI actually picks duties, mind. I still think it's convoluted. Heck, it's convoluted for the human brain.

Edited by Svenc

Share this post


Link to post
Share on other sites
Guest El Payaso

I was trying to ask about this in the "Ask SI anything"-thread (I mean "how they are trying to make the AI play realistic style of plays in the game?") just to see everything else being answered but my question totally ignored so by that I immediately thought that I am either on their bad books or something or that there isn't a hope to see anything happening this year either. 

And the next chapter is off-topic:

Spoiler

This issue along with the transfer system issues, which seem to still be there as you still seem to be able to pick all the players that you want without any resistance from the AI, will make the game for me at least in long term unplayable. Well luckily I don't have a lot of time to play and my Las Palmas save is progressing like the snails on our streets when it's been raining. :) Something like: "Well I bought Dolberg, wonderkid 2, wonderkid 3, wonderkid 4 etc." puts me really off as there should be competition from other AI clubs and also more will from the owning club to keep the player and also more will from the players to stay with their current club.

These two were definitely areas that I hoped would be improved, balanced or whatever but they have failed to make it. 

Would be interesting to see actually how badly the AI is getting team instructions etc. wrong as on beta phase you can see the roles and the duties and quite often those even aren't completely wrong and there are for example Support - Attack duties combined down the flanks while the worse AI managers make those silly systems with AF and Poacher up front combined to zero support from midfield. Also it is a bit funny how many tactical errors the human users can run with keeping their feet dry and overperforming. For example any of those more famous YouTubers don't have any knowledge whatsoever on tactics and using key highlights they just keep going like steam trains.

Share this post


Link to post
Share on other sites
Guest
23 minutes ago, El Payaso said:

I was trying to ask about this in the "Ask SI anything"-thread (I mean "how they are trying to make the AI play realistic style of plays in the game?") just to see everything else being answered but my question totally ignored so by that I immediately thought that I am either on their bad books or something or that there isn't a hope to see anything happening this year either.

 

Since FM2013, I have been asking whether there is any chance of players actually understanding their sizes relative to each other so that they would stop playing high balls to small guys marked by big defenders. Nobody has ever said anything, so I can definitely feel you.

 

For me, this is one of the biggest gripes in the match engine. There is no way to ask your players to keep the ball on the ground. Just look at any match stats and you will find that your striker has attempted approximately 20 headers, often more than your central defenders. If he's a small guy, he will get rubbish ratings unless he scores. And your players will never, ever understand that if you got Aguero up front against Sergio Ramos, it's just NOT a good idea to play high balls to him. They will try it over and over again, and Sergio Ramos will gladly take the ball every time. Also, there is no way to ask your players to stop doing this.

Share this post


Link to post
Share on other sites
3 hours ago, Svenc said:

never do CBs much recycle/shift the ball between them. And b) once it's played out of the back, it rarely moves back again.

these two points are crucial for how the match plays out in the ME and why we see such discrepancy between irl man city and one in the game.

Once the ball is played out from the back it rarely goes back - there is a good reason for it. If you observe real football match on whatever level, you will notice why the ball gets back in the defensive third. Simply because the opposition prevents the team on the ball to go forward. In FM, this behaviour doesn't exist. The teams don't play the defensive phase as a unit and the team on the ball always has space to move the ball forward without the need to go back through the back line or the pivot (which largely makes these roles redundant).

It is useless to expect Man City in FM to play anything like man city in real because there simply is no need. Guardiola doesn't recycle the possession and shift the ball around because he is emotionally attached to it. He does it because he has no space in more advanced zones so he needs to create that space. In FM, there is no need to create that space as the space is provided by poor defending. If you go to youtube and search for some match from 60's you will see the ball doesn't go much back once it is in the middle third. For the same reason as in FM - no need. The defensive phase is the key that prevents the game from representing lifelike football in FM. 

Share this post


Link to post
Share on other sites
Guest El Payaso
3 minutes ago, MBarbaric said:

 

I would also add to this especially the defenders' tendency to go hoof ball under any kind of pressure while for example with 'play out of defense' you would expect defenders to try to follow this instruction in all situations and take risks with the ball because that is what they are instructed to do. 

Strikers' defensive positioning (which you mention here between the lines for sure) is also playing a role as this prevents the attacking side's defenders to get high up the pitch and take part to the build ups. This flips passing stats to haywire as in FM support duty strikers are summing up pass amounts that midfielders and even playmakers sum in real life and things go even crazier with playmakers even though this is a bit better compared to woeful FM 2017 where these stats were totally out of control. And while players in these positions are doing it the centre backs rarely even touch the ball. 

But I think that we who actually care about these things are in minority and our years of persistent asking for answer never gets any attention or answers. Maybe because those big masses who don't care about detail or want to put effort in playing the game are more important in terms of business. 

Would like to claim that this is the last FM I will buy but that is never going to be the case as I always support SI by buying it and every autumn I am hopeful that they have managed (or are willing?) to improve the AI but it never seems to happen. 

Share this post


Link to post
Share on other sites
5 minutes ago, El Payaso said:

Would like to claim that this is the last FM I will buy but that is never going to be the case as I always support SI by buying it and every autumn I am hopeful that they have managed (or are willing?) to improve the AI but it never seems to happen. 

That's what I thought until last year.

5 minutes ago, El Payaso said:

But I think that we who actually care about these things are in minority and our years of persistent asking for answer never gets any attention or answers. Maybe because those big masses who don't care about detail or want to put effort in playing the game are more important in terms of business. 

Absolutely we are in minority. Most of the people don't even see these things. And, to be honest, I understand that. Most of the people just want to take their favourite team where they rightly belong :D And I understand SI as well. It is obvious where is the money and they are a business after all. Creating a plausible football match is a difficult task and if you improve the defence you need to revamp the attack which is way more difficult to represent in computer code. Defence is a well-defined system with rules and laws of its own that would be much easier to code to play out to perfection. Attack, on the other hand, is much less predictable, fluid and diverse. A nightmare for computer code I'd imagine.

Share this post


Link to post
Share on other sites

A few good points @MBarbaric . I think there's still room to improve, even though that must be difficult to code. On the last PES I own some aspects of zonal defending were represented somewhat decent enough on higher levels of difficulty that, without any of the super dribbers arcade sims exaggerate anyway (everybody going Barca/Real online), you would need to pass the ball around the park quite a bit for there to open space. All within the confines of these being two by six minutes of oneway soccer, naturally -- I found it difficult to score a hats full every time even against worse teams sometimes. And yeah, as of FM, teams aren't that oftenly required to start rebuilding/reshuffling at the back again. On occasion it still looks as if it's hardcoded, as CB's oft seem to be ignored. Then again, maybe I just suck at PES. :D

Speaking about "sucking". I think some fundamentals are still important here, and SI will be trying to improve them anyway. Reason is the underlying perception of unfairness, when what is behind is not seldom defensive / ME issues (apart of tactics, naturally, which can serve to higlight them). Because, at some point, everybody deeply cares about this. It is when his side loses a match he can't comprehend how, then logs in to scream foul play. Unfortunately, in particular  in the GD forums, any kind of "quality" assessment is rare. Which probably makes the GD forums not the most ideal place to report much, as it's typically drowned in the screams of anger, frustration and despair. It wouldn't surprise me if it isn't much monitored anymore by SI staff directly working on areas that this influences. Well they have different feedback from all kinds of levels anyhow, so let's see.

That said, that hoof ball by defenders brought up by @ElPayaso I have rarely witnessed in prior. In particular if you tucked in all the possession based instructions, the decision bias to pass from my experience became this big they passed it short even when the guy next to them was then promptly caught in dangerous areas, like directly after an attack was defended, and the opposition was still all pushed up. I saw goals scored that way -- rather than the defender clearing it to the forwards staying high up the pitch, which would be the more natural choice, he would then pass it to the DM, who would be immediately dispossed, and the side would concede. I didn't even see that constant hoof ball on lower leagues, actually. It's something I personally even considered a weakness, as even on such levels you could build it cleanly from the back with ease, play keep-ball ad nauseum. Seems to be one of those cases where there are different versions/flavors of FM going around (mine probably happened to be Strawberry/Barcelona -- there must be a Crushed Wasabi/1980s Watford as well someplace because such has been reported quite oftenly). :D

Edited by Svenc

Share this post


Link to post
Share on other sites
Guest El Payaso

Just witnessed Real Madrid having 4 shots on goal against Atlético with 58% of the possession. I know that Atlético are well known for their defensive qualities but there really seems to be a lot wrong on how the AI teams attack as Real basically always should be a team that are able to find couple of those openings no matter how the opposition defends. Should be quite simple for them to make it work by just having Benzema as a support duty forward and by that linking up with Ronaldo and Asensio. Meanwhile Barcelona are scoring 5 goals in a game, good, but three of them from crosses with Dembele and Messi are dominating the big defenders of Athletic inside the penalty area. :(

Share this post


Link to post
Share on other sites

This debated again :D

As I’ve shown over recent years & in each version of FM it is possible to put together a tactic that produces similar numbers & styles as seen from the real life players & managera TFF has mentioned, he may recall the threads we were previously involved in.

The problem has always been that the AI’s use of the TC has fallen short due to what can only be perceived as a lack of a more complex thought processes, the end result is that AI managers who employ a more simplistic tactical system tend to be more successful in FM because that’s easier to replicate.

The fault here is not with the ME, it’s with the team selector which needs to be taught how to use the tools available to implement the style of play that a manager’s tactical attributes & traits says they should employ.

Share this post


Link to post
Share on other sites
Guest El Payaso

@Barside I don't know if anyone is blaming the engine here but indeed the AI's tactical choices. I've noticed that it is fairly easy to get attacking flowing reasonably well in the ME and this with decent or good players compared to the level of the league. The only thing I'm wondering here is how the AI fails to do that with world class players. 

Share this post


Link to post
Share on other sites
14 minutes ago, Barside said:

This debated again :D

As I’ve shown over recent years & in each version of FM it is possible to put together a tactic that produces similar numbers & styles as seen from the real life players & managera TFF has mentioned, he may recall the threads we were previously involved in.

The problem has always been that the AI’s use of the TC has fallen short due to what can only be perceived as a lack of a more complex thought processes, the end result is that AI managers who employ a more simplistic tactical system tend to be more successful in FM because that’s easier to replicate.

The fault here is not with the ME, it’s with the team selector which needs to be taught how to use the tools available to implement the style of play that a manager’s tactical attributes & traits says they should employ.

I would prefer one year dedicated to such improvement to the AI instead of pompous new features like dynamics. The game cannot replicate the current  tactical footballing trends and i think the developers should dedicate more time to this area which only briefly mentioned improvements were done every single year.

Share this post


Link to post
Share on other sites
26 minutes ago, Barside said:

The problem has always been that the AI’s use of the TC has fallen short due to what can only be perceived as a lack of a more complex thought processes, the end result is that AI managers who employ a more simplistic tactical system tend to be more successful in FM because that’s easier to replicate.

The fault here is not with the ME, it’s with the team selector which needs to be taught how to use the tools available to implement the style of play that a manager’s tactical attributes & traits says they should employ.

Better players = running circles around AI left, right and centre. :D It's still not quite as simple as that, and it's not simply about possession or styles of play anymore either way. This is big teams not domianting the leagues as they do in real football, full stop. Even in completely unbalanced leagues, such as the Scottish PL, where nobody has com even close to Celtic in half a decade running, but in-game, the likes of Aberdeen can easily outperform their "attacking" nous alone. If you go through matches individually, you'll also find evidence that on the occasion big teams are content with 0-0 draws sometimes in domestic matches they would be expected to domiante, or else Bayern going a narrow formation with both backs kept back  wouldn't make any sense (and naturally, is mostly pish easy to defend when that happens).

You may also find that it's not as simple to massively "statistically dominate" many a match anymore either way with the big sides (my experience). Otherwise, agree. Thing is, unless SI make the selection/style shaping process more straight forward, which can be an entry level barrier challenge for human players, how do you think AI could hope to cope? They're not going to become holistic thinking machines, they aren't on any strategy game on the planet, where AI is an important part of the game world. What they need is styles and strategies, ready to pick up, the way the "match plans" first introduced to Touch work, except a bit refined plus playing styles added into the mix. I'd also monitor why such nonsensical decisions can be at all possible (Bayern example), as there is no scenario in real football were Bayern would "play for the 0-0 draw" from kick off in any domestic match ever. They probably never did in the past 30+ years. If such is possible, who knows what else is.

https://uploadfiles.io/wccv9

Edited by Svenc

Share this post


Link to post
Share on other sites
11 minutes ago, El Payaso said:

@Barside I don't know if anyone is blaming the engine here but indeed the AI's tactical choices. I've noticed that it is fairly easy to get attacking flowing reasonably well in the ME and this with decent or good players compared to the level of the league. The only thing I'm wondering here is how the AI fails to do that with world class players. 

What I started to do last year I’d interpret the tactical data for Luis Enrique & transfer that into a tactic using the TC, ending up producing some very strong results & pleasing performances.

If folk are interested in highlighting the issue of top level AI managers being ineffective I would advise that showing what the human manager can do when working within the confines of the AI managers attributes is the best way of highlighting the AI’s tactical shortcomings.

Hopefully a few regulars will take on that challenge & provide the detailed feedback required in the bugs forum or at least link SI staff to any thread replicate (insert manager) thread in the tactics forum.

@Svenc I think part of the problem is that an AI manager may view a 1-0 to be as good as a 5-1 win which then puts them at greater risk of a draw or loss, no idea if I’m even remotely close in that thinking but I’ve always had the sense that the AI managers are content to aim for a win with the least amount of drama or effort.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...