Attribute Testing - Football Manager 24

by Zippo, Jan 5, 2025

For most attributes, you can figure out if they’re defensive or offensive based on goals for/goals against, but jumping reach isn’t as clear. Do the +14 goals come mainly from set pieces or open play? It’s probably set pieces, right? This just reinforces how jumping reach doesn’t matter much for strikers, like the ykykyk findings from a few years ago, especially since most top tactics rely on low crosses. Having 2-3 defensive players with good jumping reach should be plenty to score from set pieces and defend them effectively.

If my thinking is off here, let me know.

0

kvasir said: For most attributes, you can figure out if they’re defensive or offensive based on goals for/goals against, but jumping reach isn’t as clear. Do the +14 goals come mainly from set pieces or open play? It’s probably set pieces, right? This just reinforces how jumping reach doesn’t matter much for strikers, like the ykykyk findings from a few years ago, especially since most top tactics rely on low crosses. Having 2-3 defensive players with good jumping reach should be plenty to score from set pieces and defend them effectively.

If my thinking is off here, let me know.


Are u playing the game? JR strikers are by far the ones that most score.

0

JEL7879 said: Are u playing the game? JR strikers are by far the ones that most score

They score headings from set pieces, yes. I honestly saw more headings from late winger runs than from strikers in this game.

3

JEL7879 said: Are u playing the game? JR strikers are by far the ones that most score.

I haven’t noticed much difference in open play. A striker with 20 Acc/Pace and 10 JR doesn’t seem to score a lot less than one with 20/20/20 Acc/Pace/JR. The big difference comes from set pieces. My CB gets 15-20 goals per season from them. So yeah, in my opinion, that’s how Haaland racks up 60-80 goals a season without much effort in-game—he’s the main aerial threat on set pieces, plus his insane Acc/Pace helps.

0

BulldozerJokic said: They score headings from set pieces, yes. I honestly saw more headings from late winger runs than from strikers in this game.

Yeah, sometimes the wingers come at the far post scoring headers, while having 10 jumping reach.

2

It's still crazy the difference between JR 8 and JR 20 is so small. Is the test tactic using default set pieces?

0

If we exclude obviously OP acceleration and pace, then jumping reach is still one of the best attributes. Lets not get carried away by how broken the first two are.

0

Is it possible to test Agility from a GK point of view?

0

AIK said: Is it possible to test Agility from a GK point of view?

The RNG of the test is about (+/- 1.5 Point) (+/- 1.5 Goals For) (+/- 1.5 Goals Against)

Look at the result of Agility test for 10 outfield positions, it's (Points +8), (Goals For +6), (Goals Against -6) which means it's about (Points +0.8), (Goals For +0.6), (Goals Against -0.6) per 1 outfield position.

So testing Agility only for 1 position doesn't make sense because the difference won't overcome the RNG of the test.

Even that GK position is a bit different than any other outfield position but I really doubt that changing Agility attribute from 8 to 20 only for 1 position would overcome the test RNG and btw, we test each attribute about 9,000 matches and despite that there's still RNG (+/- 1.5 Point) (+/- 1.5 Goals For) (+/- 1.5 Goals Against). :)

0

@Zippo Going off what you've found so far, minus acceleration and pace, what would you say are the go to attributes? is there a certain value threshold that we should look for? e.g +15 points

Great work as always

0

Zippo said: The RNG of the test is about (+/- 1.5 Point) (+/- 1.5 Goals For) (+/- 1.5 Goals Against)

Look at the result of Agility test for 10 outfield positions, it's (Points +8), (Goals For +6), (Goals Against -6) which means it's about (Points +0.8), (Goals For +0.6), (Goals Against -0.6) per 1 outfield position.

So testing Agility only for 1 position doesn't make sense because the difference won't overcome the RNG of the test.

Even that GK position is a bit different than any other outfield position but I really doubt that changing Agility attribute from 8 to 20 only for 1 position would overcome the test RNG and btw, we test each attribute about 9,000 matches and despite that there's still RNG (+/- 1.5 Point) (+/- 1.5 Goals For) (+/- 1.5 Goals Against). :)


Ah, makes sense. Thanks for elaborating.

0

CBP87 said: @Zippo Going off what you've found so far, minus acceleration and pace, what would you say are the go to attributes? is there a certain value threshold that we should look for? e.g +15 points

Great work as always


I think the most effective approach to pick players would be something like this:

Let's say you're choosing between 2 players:

"Player A" 16 Acceleration, 16 Pace, 8 Anticipation, 8 Work Rate
"Player B" 15 Acceleration, 15 Pace, 18 Anticipation, 18 Work Rate

Which of them to pick?

According to our attribute test improving each your outfield player's Acceleration attribute by 1 point improves your team result by 5.3 points on 38 matches distance. Ok, but where did I get that 5.3 value for the Acceleration attribute? It's quite simple just look at the test result of Acceleration attribute and you find that increasing its value from 8 to 20 improves the result by 64 points then increasing the Acceleration value only by 1 point would increase the result for 5.3 points. For simplicity's sake we assume that the relationship is linear here.

Using the same approach as above we can find out, let's call them "Weights", of other attributes Pace, Anticipation and Work Rate.

So the weights for Acceleration, Pace, Anticipation and Pace would be:

Acceleration = 5.3
Pace = 5.3
Anticipation = 1.5
Work Rate = 1

Now, when we know the weights of the attributes we can calculate the "usefulness" values of "Player A" and "Player B"

"Player A" = (16 Acceleration) x 5.3, (16 Pace) x 5.3, (8 Anticipation) x 1.5 , (8 Work Rate) x 1 = 189 W
"Player B" = (15 Acceleration) x 5.3, (15 Pace) x 5.3, (18 Anticipation) x 1.5 , (18 Work Rate) x 1 = 204 W

It comes out that the "Player B" is a slight better pick than "Player A" despite the fact that "Player A" is faster than "Player B"

3

Zippo said: I think the most effective approach to pick players would be something like this:

Let's say you're choosing between 2 players:

"Player A" 16 Acceleration, 16 Pace, 8 Anticipation, 8 Work Rate
"Player B" 15 Acceleration, 15 Pace, 18 Anticipation, 18 Work Rate

Which of them to pick?

According to our attribute test improving each your outfield player's Acceleration attribute by 1 point improves your team result by 5.3 points on 38 matches distance. Ok, but where did I get that 5.3 value for the Acceleration attribute? It's quite simple just look at the test result of Acceleration attribute and you find that increasing its value from 8 to 20 improves the result by 64 points then increasing the Acceleration value only by 1 point would increase the result for 5.3 points. For simplicity's sake we assume that the relationship is linear here.

Using the same approach as above we can find out, let's call them "Weights", of other attributes Pace, Anticipation and Work Rate.

So the weights for Acceleration, Pace, Anticipation and Pace would be:

Acceleration = 5.3
Pace = 5.3
Anticipation = 1.5
Work Rate = 1

Now, when we know the weights of the attributes we can calculate the "usefulness" values of "Player A" and "Player B"

"Player A" = (16 Acceleration) x 5.3, (16 Pace) x 5.3, (8 Anticipation) x 1.5 , (8 Work Rate) x 1 = 189 W
"Player B" = (15 Acceleration) x 5.3, (15 Pace) x 5.3, (18 Anticipation) x 1.5 , (18 Work Rate) x 1 = 204 W

It comes out that the "Player B" is a slight better pick than "Player A" despite the fact that "Player A" is faster than "Player B"


that's insane difference. -1 to acc and pace is just slightly worse than +10 to work rate and anticipation

0

I remember reading in another thread that other than decorative attributes, the ones that matter are divided into 2. I think it was capped and uncapped. Capped attributes have diminished effects beyond a certain point while uncapped attributes will give an effect for each point until 20. Are these gonna be tested too?

0

Germaniac said: I remember reading in another thread that other than decorative attributes, the ones that matter are divided into 2. I think it was capped and uncapped. Capped attributes have diminished effects beyond a certain point while uncapped attributes will give an effect for each point until 20. Are these gonna be tested too?

Those were my words earlier but more testing proved that it was a wrong assumption so I take it back.

1

@Zippo : TLDR: Please include all hidden attributes to the test

The new attribute testing is more informative than the previous version. I believe that most players in the test league typically had attributes ranging between 13 and 15, marking them as top-tier players. Increasing the attributes by +5 elevated them to super elite status. However, now we are increasing the attribute from 8 to 20, which includes most players that folks would recruit.

I appreciate all the hard work that has gone into this. My only request is to test all hidden attributes, not just Consistency. I understand that depending on how the test league is set up, certain hidden attributes like "Important Matches" might not be properly tested. However, testing whatever possible will greatly enhance our understanding of game mechanics and improve player recruitment strategies.

0

kvasir said: For most attributes, you can figure out if they’re defensive or offensive based on goals for/goals against, but jumping reach isn’t as clear. Do the +14 goals come mainly from set pieces or open play? It’s probably set pieces, right? This just reinforces how jumping reach doesn’t matter much for strikers, like the ykykyk findings from a few years ago, especially since most top tactics rely on low crosses. Having 2-3 defensive players with good jumping reach should be plenty to score from set pieces and defend them effectively.

If my thinking is off here, let me know.


Anecdotally for me at least, Strikers with high JR get the majority of their goals from set pieces and not the run of play so you are really just cannibalizing goals from your DC that need to have it anyway. So if you have a high JR CB there is not a huge net gain in team goals from having high JR on your striker.

AMLR (IFS) with big JR get more heading goals in the run of play in my experience due to the mismatches against DRLs.

So I suspect JR effectiveness is mostly matchup dependent and these tests basically show what happens if you win nearly every header.

3

Thanks for the updates!
A thing I was wondering for a while: PPMs don't seem to have a huge effect on results from other tests I've seen. But I was wondering: Since Pace and Acceleration are so effective, what about PPMs that rely on those stats, particularly "Knocks Ball Past Opponent"? Is there any difference in having something like 20 Pace / Acc vs having 20 in both plus that PPM?

0

bigloser said: Anecdotally for me at least, Strikers with high JR get the majority of their goals from set pieces and not the run of play so you are really just cannibalizing goals from your DC that need to have it anyway. So if you have a high JR CB there is not a huge net gain in team goals from having high JR on your striker.

AMLR (IFS) with big JR get more heading goals in the run of play in my experience due to the mismatches against DRLs.

So I suspect JR effectiveness is mostly matchup dependent and these tests basically show what happens if you win nearly every header.


exactly described my experience as well

0

bigloser said: Anecdotally for me at least, Strikers with high JR get the majority of their goals from set pieces and not the run of play so you are really just cannibalizing goals from your DC that need to have it anyway. So if you have a high JR CB there is not a huge net gain in team goals from having high JR on your striker.

AMLR (IFS) with big JR get more heading goals in the run of play in my experience due to the mismatches against DRLs.

So I suspect JR effectiveness is mostly matchup dependent and these tests basically show what happens if you win nearly every header.


Could be that most tactics are playing with low crosses as well.

My strikers never scores on set pieces. But they score maaaaany goals in open play.

0

I want to thank EVERYONe who has contributed to this post and a HUGE thanks to the initial post. This fundamentally changes how I will scout players and also makes it easier (my opinion). Knowing that there are certain attributes that either add very little or nothing at all in the match engine saves time evaluating potential signings. I do hope though in FM 2025 that ALL of the attributes as well as hidden traits help shape who a player is and how they play in the match engine.

0

About the GK Attributes, since you included some physicals in your tests such as Agility, Acceleration, Pace, are you going to test Jumping reach for them?

0

nios said: About the GK Attributes, since you included some physicals in your tests such as Agility, Acceleration, Pace, are you going to test Jumping reach for them?

In the game you can hover the mouse cursor over "Aerial Reach" attribute and read the description.

When it comes to aerial challenges GKs use "Aerial Reach" attribute instead of "Jumping Reach" attribute as outfield players do.

2

I think what could actually matter for GK performance is how tall he is

0

Does the gk testing results mean, that the only thing I should be looking at regarding gk attributes is reflexes, agility and acceleration?

Btw. love that you test all this. Much appreciated.

0

Tidemann said: Does the gk testing results mean, that the only thing I should be looking at regarding gk attributes is reflexes, agility and acceleration?

Btw. love that you test all this. Much appreciated.


Well, you probably still want some basic level of all other attributes. And acceleration impact is very low, so I just focus on reflexes ang agility

0

AIK said: Could be that most tactics are playing with low crosses as well.

My strikers never scores on set pieces. But they score maaaaany goals in open play.


I doubt it. They still float low % crosses with that setting which is why I experimented with a high JR but slower than i normally use striker. I found that IFs and offset AMCs with abnormally high JR got on the end of those more.  If the other team has low JR central defenders it's probably just as affective. It's all about difference matchups imo.

He still scored a lot of goals, but nearly all the heading ones were on corners or ones where he wasn't marked.

0

Jolt said: That is extremely surprising. So every combination tried merely increases cumulatively with the increase of the combined chosen attributes, with no statistical difference from the sum of the chosen attributes?

EDIT: So what does testing of a 20 Pace and 20 Acceleration look like, in terms of points, goals for and goals against, as opposed to a 20 Pace 10 Acceleration, or a 10 Pace 20 Acceleration?

EDIT 2: Or perhaps even a more clear example of synergy: "Goals for" with both Crossing and Heading in 20. For an attacking team to fully exploit high heading attributes, accurate crossings with high Crossing attributes generally means higher opportunities of goal scoring, as opposed great crosses to mediocre heading players or mediocre crosses that don't frequently reach great heading players. I wonder what's the testing difference specifically goals for, between 20 Crossing 20 Heading, and 20 Crossing 10 Heading, and 10 Crossing 20 Heading.


@Zippo Could you clarify this?

What's the practical difference in points, Goals For, Goals Against, between the current tests where only one attribute gets raised, and when two with theorerical synergies get raised (Pace+Acceleration; Crossing+Heading; or others such as Heading+Jumping Reach). I'm curious to understand in practical terms what "there's no difference" means.

0

Jolt said: That is extremely surprising. So every combination tried merely increases cumulatively with the increase of the combined chosen attributes, with no statistical difference from the sum of the chosen attributes?

EDIT: So what does testing of a 20 Pace and 20 Acceleration look like, in terms of points, goals for and goals against, as opposed to a 20 Pace 10 Acceleration, or a 10 Pace 20 Acceleration?


Hi,

"20 Acc/ 20 Pace" combo gets a higher score than "20 Acc/ 10 Pace" or "10 Acc/ 20 Pace" combos.

"20 Acc/ 10 Pace" and "10 Acc/ 20 Pace" combos get similar scores.

"20 Jumping/10 Heading" combo gets a higher score than "10 Jumping/20 Heading" combo.

I hope it helps.

Cheers.

0

Guys, a small notice here. Recently, I posted a guide how to search players with FMRTE based on the fm-arena attribute testing. But we found that FMRTE has an issue when it come to the Position Ratings, it's only possible to set them for the striker position and the application evaluates any other positions only on the Striker position ratings.

However, despite the issue, it's still possible to use FMRTE for searching players based on the position rating but it might be very confusing for some people so we decided to remove the guide about FMRTE.

I just wanted to let you what happened to the guide and why it's gone.

Cheers.

0
Create an account or log in to leave a comment