FM20 Experiment - Goalkeeper as striker in Liverpool (Inspired by Phil Jones ✕ Paris Saint-Germain)

[Luizinho]

17 minutes ago, russell9 said:

So after 23 appearances, he has a better conversion rate than both Harry Kane and Sergio Aguero. (albeit slightly skewed by penalties)

Can someone from SI confirm what is happening here? I've lost a lot of faith in the game after seeing this thread...

[HUNT3R]

3 minutes ago, Luizinho said:

albeit slightly skewed by penalties

10% isn't slight, btw. If you exclude penalties the rate is 16%.

[Tiger666]

How many of Aguero and Kane's goals were penalties?

[Zemahh]

54 minutes ago, russell9 said:

screenshots

Wow. 

So, to sum it up:

A 17 y/o 1-star Current Ability reserve goalkeeper has:

• Better Average Minutes Per Goal than Aubameyang, Vardy and Mousset
• Same Shots On Target percentage as Aubameyang
• THE BEST Conversion Rate percentage in the league

How does one explain this?

[davehanson]

47 minutes ago, HUNT3R said:

10% isn't slight, btw. If you exclude penalties the rate is 16%.

So would be the same as Vardy and Jesus then? 

Not a great advert for the ME on FM20 whichever way you want to look at it. There is no way that a 17 yo goalkeeper should be scoring goals like that in the Premier league. Literally the ONLY goals he should be scoring are penalties and goals from set-pieces. 

Edited October 28, 2020 by davehanson

[steam just is]

As soon as 21 comes out this is the first test! 

[naterego93]

What a joke, I'm gonna put my mom as a striker on this & she'll probably have a 20 goal season, "FM 20's ME is better" my arse!

[Zemahh]

Just now, steam just is said:

As soon as 21 comes out this is the first test! 

Sadly, this looks like it's part of a bigger problem and not a measly bug that could just get "fixed". Firstly, one would need to do some extensive testing to confirm this can actually happen regularly and is not just some sort of a one-in-a-thousand fluke season, but if it does, it's probably something that would need major match engine changes to get fixed. Either that, or position familiarity ratings would need to become much more relevant than they currently are.

@Rashidi Would be curious to hear your take on this. I know you focus very much on specific attribute values and have repeated many times how important they are, so how can a "striker" with 7s and 8s across the board have that much success in the Premier League, where average defences have Positioning, Anticipation, Marking and Acceleration of 15 and higher? In theory, this guy should not even be able to have a sniff aside from an occasional set piece goal, regardless of who's assisting him.

[Luizinho]

1 hour ago, HUNT3R said:

10% isn't slight, btw. If you exclude penalties the rate is 16%.

I assume 16% will be still higher than Kane after you exclude his penalties...

Can anyone else from SI shed some light here on whats going on here please? 

Edited October 28, 2020 by Luizinho

[ferrarinseb]

1 hour ago, Zemahh said:

Wow. 

So, to sum it up:

A 17 y/o 1-star Current Ability reserve goalkeeper has:

• Better Average Minutes Per Goal than Aubameyang, Vardy and Mousset
• Same Shots On Target percentage as Aubameyang
• THE BEST Conversion Rate percentage in the league

How does one explain this?

The 2nd one concerns me the most. That was purely striker instinct and a young GK matching it really not comforting.

[russell9]

1 hour ago, Zemahh said:

Sadly, this looks like it's part of a bigger problem and not a measly bug that could just get "fixed". Firstly, one would need to do some extensive testing to confirm this can actually happen regularly and is not just some sort of a one-in-a-thousand fluke season, but if it does, it's probably something that would need major match engine changes to get fixed. Either that, or position familiarity ratings would need to become much more relevant than they currently are.

I just did another run and the results are quite similar so far. To be honest results like this should not be reproducible even in a millionth run it should be absolutely impossible in a reasonable robust system.

First friendly match, Ben already score and man of the match...

 

[russell9]

Second friendly, another MOM for Ben.

 

Edited October 28, 2020 by russell9

[russell9]

Fast forward to last friendly, Ben hattrick.

[russell9]

First league match, Ben score and not a penalty.

[russell9]

Second match, Ben score one and miss a penalty.

[russell9]

I'll just stop the rerun here I think that's enough. I'll back to finish the first test later when I'm free just to see if Ben could win golden boot 

[Dbuk1]

The only surprising thing is everybody seems shocked. 

[steam just is]

22 minutes ago, Dbuk1 said:

The only surprising thing is everybody seems shocked. 

I am! I honestly thought this was impossible. It makes a mockery of the match. A few weeks before the release this isn’t good? Needs an explanation ASAP.

[thiagoanjo]

Em 27/10/2020 em 05:23, Viking disse:

Well... I'm 52 and could'nt run 5 meters straight to save my life, and I would score a goal per game with those players behind me.

But on a serious note, and just out of curiosity: whats Bens versitality?

Say it for Firmino

[thiagoanjo]

1 hora atrás, Dbuk1 disse:

The only surprising thing is everybody seems shocked. 

So, do you think this is even close to simulate a reality?

[thiagoanjo]

23 horas atrás, HUNT3R disse:

 

@russell9 How are you simulating these matches? Are you playing them yourself? Instant Result? Going on holiday?

What possible difference could be if it is holiday or instant result?

[HUNT3R]

10 minutes ago, thiagoanjo said:

What possible difference could be if it is holiday or instant result?

I didn't say it made a difference. I just asked how the matches were simulated. I was just interested to know.

To try and answer your question - Holidaying can see your assistant fiddle with your tactic. I don't use instant result, so I don't know what that could possibly alter.

[thiagoanjo]

36 minutos atrás, HUNT3R disse:

I didn't say it made a difference. I just asked how the matches were simulated. I was just interested to know.

To try and answer your question - Holidaying can see your assistant fiddle with your tactic. I don't use instant result, so I don't know what that could possibly alter.

I was just curious, the way you asked I felt the game could process match engine differently if Holiday or Instant Result calculation.

Also, how I merge content? My answers are terribly flooding, I thought it was auto-merged, Im sorry

Edited October 28, 2020 by thiagoanjo

[bananatableleg]

4 hours ago, steam just is said:

As soon as 21 comes out this is the first test! 

I've got some bad news, friend

[bananatableleg]

3 hours ago, Zemahh said:

 or position familiarity ratings would need to become much more relevant than they currently are.

As far as I'm aware, there's minimal relevancy to them. A decisions penalty when there's not much evidence that the ME makes good decisions in the first place is either a tiny penalty for playing out of position, or position familiarity actually doesn't matter, at all. Some people have just been aware of this for a little longer, I suppose. For the record, Firmino is my favorite wing back in FM20

[WillHoward42]

Would be interesting to see if it had the same affects, using the same tactic and same goal keeper up front, but at a lower prem team.

 

[Deego619]

FM is a game of numbers and math, it simply can't take into account that Goalkeepers do not have the awareness to play up front, never mind a youth team goalkeeper. 

I guess the only real life reference that comes to mind is to David James' performance up front for Man City some years back.

 I can't imagine this guy outscoring Aubameyang even if being supplied by Mane/Salah et all.

But FM simply can't see it this way, because again it's a game of math and probability. If Liverpool have 10 world class players and 1 passenger, the Liverpool team will win the majority of dice rolls all over the pitch, which may result in the game generating a scenario where the passenger (youth goalkeeper) scores because he's the right guy at the right time. 

I feel we need to take this a bit less seriously than this thread is doing, FM is a game and cannot possibly replicate real life and the millions of variables involved. If your youth team GK scored 10 goals, Frimino may have scored 20. It seems to be probability at work over anything else. Interesting experiment nevertheless.

[thiagoanjo]

45 minutos atrás, bananatableleg disse:

As far as I'm aware, there's minimal relevancy to them. A decisions penalty when there's not much evidence that the ME makes good decisions in the first place is either a tiny penalty for playing out of position, or position familiarity actually doesn't matter, at all. Some people have just been aware of this for a little longer, I suppose. For the record, Firmino is my favorite wing back in FM20

For roles there is no penalty, for position there is, but after this thread I think it is less than I thought

@Deego619 I think the attributes should have more impact then, cause the low  forwards attributes for a GK (also the miss position) should have made them bad at it

 

Edited October 28, 2020 by thiagoanjo

[steam just is]

36 minutes ago, Deego619 said:

FM is a game of numbers and math, it simply can't take into account that Goalkeepers do not have the awareness to play up front, never mind a youth team goalkeeper. 

I guess the only real life reference that comes to mind is to David James' performance up front for Man City some years back.

 I can't imagine this guy outscoring Aubameyang even if being supplied by Mane/Salah et all.

But FM simply can't see it this way, because again it's a game of math and probability. If Liverpool have 10 world class players and 1 passenger, the Liverpool team will win the majority of dice rolls all over the pitch, which may result in the game generating a scenario where the passenger (youth goalkeeper) scores because he's the right guy at the right time. 

I feel we need to take this a bit less seriously than this thread is doing, FM is a game and cannot possibly replicate real life and the millions of variables involved. If your youth team GK scored 10 goals, Frimino may have scored 20. It seems to be probability at work over anything else. Interesting experiment nevertheless.

Sorry but disagree. He was using 2 GKs up front. That’s a joke. and If it can’t simulate a kid keeper who really  shouldn’t score for Liverpool then what are we all doing?? And,,,,,it wasn’t just tap ins - he was scoring worldys. And...someone tried CDM role with a passing attribute of 3 and he had 90% passing success rate. 
actually completely disagree with your point. If we let this go unanswered the match becomes an irrelevance and therefore the game. 
I just hope is an unspotted FM20 bug. But this is a very serious issue. 

[Deego619]

  

16 minutes ago, steam just is said:

Sorry but disagree. He was using 2 GKs up front. That’s a joke. and If it can’t simulate a kid keeper who really  shouldn’t score for Liverpool then what are we all doing?? And,,,,,it wasn’t just tap ins - he was scoring worldys. And...someone tried CDM role with a passing attribute of 3 and he had 90% passing success rate. 
actually completely disagree with your point. If we let this go unanswered the match becomes an irrelevance and therefore the game. 
I just hope is an unspotted FM20 bug. But this is a very serious issue. 

I've agreed that it's not realistic if you read my post again, but if you have 10 world class players and one passenger, you'll win those RNG battles overall across the pitch, it's how AI works and why this experiment is getting results.

The GKs scoring worldies and playing like Xavi, is another argument however about how FM shows these events to us, I believe there was a custom unofficial patch that enhanced that, it would be interesting for this experiment to use that and see what happens. 

@thiagoanjo I remember reading in the threads about the custom ME that was doing the rounds that FM calculates in the back end what is supposed to happen using the attributes as a base, and then the results of this are displayed to us via watching the match (which is what the mods were saying to convey that the unofficial ME didn't actually change anything in regards to results). This is why I'm thinking if Liverpool have 10 world class players, that they win enough RNG battles in the ME logic for the ME to say "Oh there needs to be a goal now" and the young GK is in the right place at the right time. Scoring worldies and the like, that is an issue.

I'm more so trying to explain why I feel it happens rather than trying to discredit the work here which is super interesting. SI should review this at least.

Edited October 28, 2020 by Deego619

[Ronaldo Beckham]

1 hour ago, WillHoward42 said:

Would be interesting to see if it had the same affects, using the same tactic and same goal keeper up front, but at a lower prem team.

 

I am going to start a save with Burnley and either play Ben Mee or James Tarkowski up front and see how they do or I will look at buying Phil Jones and seeing how he does when he aint playing for a great team like Liverpool. 

I will post my progress in the challenge/experiment I set up earlier. 

Here is the thread as I don't want to bombard this guy's experiment.

 

Edited October 28, 2020 by Ronaldo Beckham

[hehehemann]

Anyone tried this with other versions of FM ? Does it happen on FM18 which many regard as the better ME?

[steam just is]

25 minutes ago, Deego619 said:

  

I've agreed that it's not realistic if you read my post again, but if you have 10 world class players and one passenger, you'll win those RNG battles overall across the pitch, it's how AI works and why this experiment is getting results.

The GKs scoring worldies and playing like Xavi, is another argument however about how FM shows these events to us, I believe there was a custom unofficial patch that enhanced that, it would be interesting for this experiment to use that and see what happens. 

@thiagoanjo I remember reading in the threads about the custom ME that was doing the rounds that FM calculates in the back end what is supposed to happen using the attributes as a base, and then the results of this are displayed to us via watching the match (which is what the mods were saying to convey that the unofficial ME didn't actually change anything in regards to results). This is why I'm thinking if Liverpool have 10 world class players, that they win enough RNG battles in the ME logic for the ME to say "Oh there needs to be a goal now" and the young GK is in the right place at the right time. Scoring worldies and the like, that is an issue.

I'm more so trying to explain why I feel it happens rather than trying to discredit the work here which is super interesting. SI should review this at least.

I get what you’re saying. My reply also said that he played with 2 youth keepers. So that’s 9 world class players and 2 duds up front. I understand it’s not real life and obviously it’s just equations ( all life is maths if we’re getting deep). But after 30 years of writing management sims you would have thought such anomalies would have been cleared up. 

[Deego619]

5 minutes ago, steam just is said:

But after 30 years of writing management sims you would have thought such anomalies would have been cleared up. 

We'll agree on this. I believe the ME is long overdue an overhaul and in 2020 the logic can surely be more detailed than "Oh this team has 9 world class players and the other team has 11 average ones with a lower rep, therefore team A should win and score goals somehow".

Edited October 28, 2020 by Deego619

[HUNT3R]

9 minutes ago, Deego619 said:

We'll agree on this. I believe the ME is long overdue an overhaul and in 2020 the logic can surely be more detailed than "Oh this team has 9 world class players and the other team has 11 average ones with a lower rep, therefore team A should win and score goals somehow".

And with 2-3 million lines in the ME - it is far more detailed than you're making it out to be. 

[bananatableleg]

19 minutes ago, steam just is said:

But after 30 years of writing management sims you would have thought such anomalies would have been cleared up. 

If you put bandaids on top of bandaids on top of bandaids, you end up with a mountain of soggy, half peeling bandaids all muddled up. I'd be shocked if there wasn't a major need to go back in and streamline everything

Edited October 28, 2020 by bananatableleg
grammar am hard

[Deego619]

3 minutes ago, HUNT3R said:

And with 2-3 million lines in the ME - it is far more detailed than you're making it out to be. 

If quantity of code is used as a measure of quality in SI, we have found our problem right here.

[HUNT3R]

15 minutes ago, Deego619 said:

If quantity of code is used as a measure of quality in SI, we have found our problem right here.

I am just pointing out that it's not a few dice rolls and then the ME says goal and one random lucky contestant gets awarded a goal. Attributes are taken into account and a lot of decisions made/calculated every second for the entire 90 mins.

 

[santy001]

At this point its just something of mild interest in the same way as someone posting a 100 game winning streak is. Unless the save is uploaded, the tactics are analysed, the games checked thoroughly then it remains nothing more than a minor interest because no one else can verify what has happened.

 

[Zemahh]

1 minute ago, santy001 said:

Unless the save is uploaded, the tactics are analysed, the games checked thoroughly then it remains nothing more than a minor interest because no one else can verify what has happened.

Well, the Match Engine Bugs forum is currently closed and the thread was moved from General Discussion, where it could actually attract sufficient interest to get looked at, into Specific Playing Styles, which I highly doubt developers are checking regularly. I've been told they've been linked this thread, but so far there has been no response or request for PKMs of any kind, so there's no reason to believe they're interested in investigating this.

[steam just is]

9 minutes ago, santy001 said:

At this point its just something of mild interest in the same way as someone posting a 100 game winning streak is. Unless the save is uploaded, the tactics are analysed, the games checked thoroughly then it remains nothing more than a minor interest because no one else can verify what has happened.

 

Hmmmmm not sure about that. With SI about to take in several million from combined sales of games. Customers deserve an answer.

would you buy a fundamentally flawed game? I assume not 

does this theoretically suggest FM is fundamentally flawed? Yes

im not accusing anyone of anything btw. I’m just very concerned and would prefer those closer to SI like researchers not to be so dismissive of our questions by saying it’s only of mild interest.

so whilst we wait for the proof of the anomaly (and a second challenge has uncovered similar problems)I think it would be sensible for SI to make a comment. 

 

[santy001]

The way in which people in this thread have been throwing at mentions to different forum goers who are moderators, SI staff and so on - people have to remember that other people aren't actually at your beck and call. 

By the same being uploaded, by pkm's being uploaded those with curiosity can look into it further. Maybe even take over the team and play a few games and see if they get the same result. The staff at SI, would I expect with FM21 rapidly on the horizon, have more pressing work than trying to slowly extract the details to investigate it fully. Not to mention that people have their own lives outside of their working capacities.

If the OP doesn't want to share what he's done, and no one else is replicating it... then it just goes down as a cool story bro.

- - -

Trying to conflate your personal view desire for an answer somehow with a burning desire for an answer from the entire playerbase is disingenuous. 

You have to remember, I'm not an SI employee. So I can hold any opinion I want and there's no real consideration for anything else. An unverified, unsubstantiated outcome is only ever going to be of mild interest because you don't know all the factors. If SI spent their time running down every oddity they see on the forums they'd never get chance to make another FM game in all likelihood.

There's no guarantees anyone will get an explicit response from developers, but anecdotally, the threads that have more information, have all relevant materials available and can be reviewed by others tend to have a better chance. 

[thiagoanjo]

1 hora atrás, Deego619 disse:

  

I've agreed that it's not realistic if you read my post again, but if you have 10 world class players and one passenger, you'll win those RNG battles overall across the pitch, it's how AI works and why this experiment is getting results.

The GKs scoring worldies and playing like Xavi, is another argument however about how FM shows these events to us, I believe there was a custom unofficial patch that enhanced that, it would be interesting for this experiment to use that and see what happens. 

@thiagoanjo I remember reading in the threads about the custom ME that was doing the rounds that FM calculates in the back end what is supposed to happen using the attributes as a base, and then the results of this are displayed to us via watching the match (which is what the mods were saying to convey that the unofficial ME didn't actually change anything in regards to results). This is why I'm thinking if Liverpool have 10 world class players, that they win enough RNG battles in the ME logic for the ME to say "Oh there needs to be a goal now" and the young GK is in the right place at the right time. Scoring worldies and the like, that is an issue.

I'm more so trying to explain why I feel it happens rather than trying to discredit the work here which is super interesting. SI should review this at least.

Yeah, I dont think ME works like that (general attributes x general attributes), I think even if the goalkeeper is in the right position (it shouldnt be, since he has almost no off the ball, has the penalty to play out of position, etc etc) he couldnt score (often) since lack of good enough attributes to make his finishing/shots beat the GK or even go in the right direction

[Wavelberry]

Some interesting stuff here. Obviously what you'd want to see is what @Ronaldo Beckham has suggested doing or other goalkeepers from other teams to get a true sense of whether it's Liverpool's squad that's making it easy or whether it's the tactic etc. Would have to branch out to really see how good or bad this is.

Would Ben score as many goals for Norwich for example? Or Levante? Or SPAL? 

[Tiger666]

10 minutes ago, Wavelberry said:

Would Ben score as many goals for Norwich for example? Or Levante? Or SPAL? 

No he wouldn't, but if he's scoring as many goals as the regular Norwich, Levante or SPAL strikers, it's a problem.

[Wavelberry]

3 minutes ago, Tiger666 said:

No he wouldn't, but if he's scoring as many goals as the regular Norwich, Levante or SPAL strikers, it's a problem.

Fair point. That would be interesting to see.

[steam just is]

56 minutes ago, santy001 said:

The way in which people in this thread have been throwing at mentions to different forum goers who are moderators, SI staff and so on - people have to remember that other people aren't actually at your beck and call. 

By the same being uploaded, by pkm's being uploaded those with curiosity can look into it further. Maybe even take over the team and play a few games and see if they get the same result. The staff at SI, would I expect with FM21 rapidly on the horizon, have more pressing work than trying to slowly extract the details to investigate it fully. Not to mention that people have their own lives outside of their working capacities.

If the OP doesn't want to share what he's done, and no one else is replicating it... then it just goes down as a cool story bro.

- - -

Trying to conflate your personal view desire for an answer somehow with a burning desire for an answer from the entire playerbase is disingenuous. 

You have to remember, I'm not an SI employee. So I can hold any opinion I want and there's no real consideration for anything else. An unverified, unsubstantiated outcome is only ever going to be of mild interest because you don't know all the factors. If SI spent their time running down every oddity they see on the forums they'd never get chance to make another FM game in all likelihood.

There's no guarantees anyone will get an explicit response from developers, but anecdotally, the threads that have more information, have all relevant materials available and can be reviewed by others tend to have a better chance. 

I’m sorry but where have I suggested that SI or related people should be at our beck and call? I said with a few weeks until 21 I think it’s advisable SI make a comment. Once again I’m not accusing anyone of anything but look at the reaction Santy! With the  level of heat this is generating I think it would be sensible for SI to make a simple comment which calms everyone’s doubts. 
I refer you the the points I made earlier.....the suggested anomaly, however unsubstantiated (and look at the whole thread...it’s not isolated) hints at a major flaw. It would be in SIs interests to nip this in the bud surely.

Edited October 28, 2020 by steam just is

[davehanson]

@russell9 is there any chance you could just upload the save? Leave your manager as he is, in charge of Liverpool, so people can look at the game and tactics?

 

[bananatableleg]

2 minutes ago, steam just is said:

with a few weeks until 21 I think it’s advisable SI make a comment

To be fair, it takes a few updates before the new versions are worth playing anyway, in my opinion. I really don't think there's much rush to satisfy  the people who will actually care about threads such as this one, as opposed to the pre-order crowd who would buy FM21 if it was literally just a roster update

/speculation

[Christmas]

9 hours ago, steam just is said:

Sorry but disagree. He was using 2 GKs up front. That’s a joke. and If it can’t simulate a kid keeper who really  shouldn’t score for Liverpool then what are we all doing?? And,,,,,it wasn’t just tap ins - he was scoring worldys. And...someone tried CDM role with a passing attribute of 3 and he had 90% passing success rate. 
actually completely disagree with your point. If we let this go unanswered the match becomes an irrelevance and therefore the game. 
I just hope is an unspotted FM20 bug. But this is a very serious issue. 

This is key point. We all just spent thousands of hours, me over around 12,000 hours on what is just a clicky button booboo game.

Makes me want to throw off bridge.