* This is translated by a language translator and then manually corrected, there may be some loss of meaning after translation
Test setup: The league has 4 teams, A, C, D team is set to 4231 Meta tactics (https://fm-arena.com/thread/12667-god-of-chaos-v1/), B team is set to 4141 weak tactics, all attributes of all players are set to 10, locking various attributes.
Modify the player attributes of Team A (except for the goalkeeper, that is, adjust 10 players) to the corresponding value. Statistical the goal difference.
The C and D teams use a very strong 4231 tactics, which tests the players' defensive ability and performance under pressure from strong teams. The B team uses the weak 4141 tactics to test the players' ability to Beat up weak teams and score lots of goals.
It took 10 days to test, two computers together. 2 of the attributes were tested by the help of the community netizen , and the other 70 were tested by me. Inspired by this https://fm-arena.com/table/26-player-attributes-testing/ (November 2023 version)I thought, What if the attribute is 1 ?
1.How to read this table?
The standard(Control) group represents the goal difference of Team A when all attributes are 10, which is 20.1.
## Look at Technique 20, which represents Team A (apart from the goalkeeper) skill attribute increases from 10 to 20, and Statistical the goal difference of Team A, it's 16.5. 16.5 < 20.1 (standard) In other words, the Technique has increased by 10, but the goal difference has decreased. But not by much.
## Look at Vision 1 and Vision 20, 20.0 and 38.5 respectively. 20.0 < 20.1 < 38.5 In other words, when Vision is reduced from 10 to 1, so there is no change. However, when Vision was increased from 10 to 20, resulting in a goal difference increase of (38.5-20.1)=18.4
2. Noteworthy attributes
Lowering the Work rate 1 , it have very serious consequences (goal difference -110),
and it should be ensured that a player has at least 6 and preferably 10 Work rate. A player with a Work rate is not desirable. But continuing to increase Work rate (from 10->20) is low effect. The difference between Work rate 10 and Work rate 20 is relative small.
The 6-point attribute seems to be the threshold for some attributes.
Passing, Crossing, Marking, First touch, Positioning, Decision, Teamwork, Off the ball, Bravery, are all attributes that can be considered useless.
Technique and Flair, Higher values lead to worse performance. Spoiler
Yes, it's a scam. To Flair. There are two guesses, one is that "fancy tricks" cause players to spend more "frames" on the movement, resulting in a decrease in efficiency. The other is that more "fancy tricks" contribute not enough to the mathematical expectation when attacking, while the mathematical expectation of being scored when defending increases more than attacking scored
Core attributes are these: Special : Work rate (need to reach 10, higher is useless), Pace, Acceleration, Jumping reach, Dribbling, Balance, Concentration, Anticipation, Determination，Agility.
According to the comparison of attribute 20 with attribute 10: Pace + Acceleration 2 most important cores, Jumping reach + Dribbling 2 secondary cores, Balance can be considered a level 3 attribute, level 4 attribute Anticipation and Concentration
Other attributes, which are although have effect, but much less effective.
Note that the difference between attributes is only used to show the degree of differentiation and cannot be used for direct quantified comparisons.(Because it's not standardized) . For example, "difference from the standard value +40" is not twice as useful as "difference from the standard value +20" .
3.Sample
According to the Law of Large Numbers in statistics, as long as the sample is large enough, the value you get will be closer and closer to the mathematical expectation
In terms of important attributes, there are about 950-1080 Match samples. According to statistical estimates, the margin error of attribute is ≤ 3.
In the non-important attributes, there are about 450 Match samples. According to statistical estimates, the margin error of attribute is ≤ 8
4.margin error
Is that a conclusion that only applies to 4231 and certain opponents? You know, tactics can vary a lot. Moreover, 4231 players are not evenly distributed, for example, there is only 1 striker. And tactics, duties, and commands all affect the importance of attributes. There is also the test of the Go-on-holiday, even if the game is set to maximum detail, will it get unreal values?
——Go-on-holiday have been studied, and I won't repeat it in a lot of words, but I choose to trust their findings
——Taking into account different tactics and different opponents, an additional 3,600 match were played, To test two other tactics, one is Default tiki-taka 433. And Default counter-attack 442. They will have different standard values and different attributes deviating from the standard, but "importance level" of each attribute not change.
For example, in the tactics where Passing is more important, Passing (around 24 in the table) is still not important than dribbling (more than 50 in the table) or Finishing (33 in the table) in any tactic. This also shows that the game is not complicated to do a separate set of calculations for each tactic, but a relatively universal calculation method. And this method of calculation, my personal observation, is more conducive to offensive.
——Similarly, the problem of player distribution does not have a "decisive" effect, but only a statistical perturbation of some value.
5.Multiple variables
some comments said that the performance of attributes is A multi-variable problem, for example, the A attribute is evaluated at the same time as the B attribute.
Some people will think that testing a player who First touch 10 and Finishing 15, because the player can't stop the ball and it doesn't matter how high the Finishing is, so this test is wrong.
However, if an attribute is weighted with one or more other attributes, all associated attributes must be found. In the end all testing is either pointless or too tedious to do. Since it is statistically possible to tell the difference, let's use this result as a conclusion
Now, imagine a mathematical problem. Player A's Composure is 10, provides 20% of the goal scoring ability. Finishing 10, providing 10 accuracy shots on goal per game. The end result is 10x20%=2 balls.
If this is a multivariate problem, then Composure 20, assuming it provides 50% of the goal scoring ability. The end result is 10x50%=5 balls. So 2 becomes 5. The difference "5-2=3" may or may not be obvious. Those that are not obvious are actually the effect of interaction is not good, since it is so small, it can simply be ignored. If there is an effect, it must be reflected in the statistical results.
Unless it is using some more complex mechanism, for example, it detects composure 20, and if the shot is not ≥15, the goal scoring ability % does not increase. I'm not going to consider that possibility.
6.don't bring the logic of reality into the game
For example, there is such a reply: there is no "speed 20" and other attributes are 10 in reality . Or, "Should use balabala" attributes to design all forwards/centers/back guards to reflect the real situation.
The results, as long as the control variable method is followed, make no difference (since the game seems unlikely to use the "more complex" mechanics mentioned above).
7. Why not test each player individually, such as the striker
Because it's so much work, And the impact each player has on the team's goal difference is likely to be the difference between the actual measured value and the standard value, which is always disturbed by larger random perturbations, because a tactic is mostly 1-3 strikers, who make up less than 30% of the 11 men. If you want to measure the true value, you need a very, very large number of samples
8. The goalkeeper? Similar to the above reasons. mishap If you want to measure the true value, you need a very, very large number of samples
9. Why testing just one season The result is completely inaccurate? (from reply) The random perturbation in the game is very large, which means that if you only play 30 to 40 games (a season), the result is likely to deviate significantly from the true value
One might also suspect that the test is not rigorous enough, it's not scientific enough, it's not an actual season, it's a nice ideal environment, and so on, and then argue that the empirical results under a lot of interference are more accurate than the results obtained under the control variable method
10.thoughts
Like the previous "1CA Win Champion", many people will dismiss this as a conspiracy theory before testing reveals the game mechanics. After the revelation, some players believe that this kind of testing is just excessive pursuit of intensity at the expense of game life. But that's not the purpose of the test.
My goal in testing this is the same as the goal of many gamers: to force the game company to continuously improve the old mechanics it has used for years, to change the fundamental engine flaws, and to prevent players from playing a "placebo simulator". Why a placebo simulator? There are people who take real football and put a lot of time and energy into the game, there are a lot of players who put a lot of energy into the video, I also enjoy these passionate things But, As a result, the game engine does not do these functions at all, these players believe and carefully prepared, essentially reflected in the game engine is useless. The game relies on some very simple data and then randomizes it to make players think they are playing a real simulation. Then the game company can just change the appearance of their game, sell it for generations, and advertise that they are making a simulation full of interactivity
What's more, the authorities have banned discussions on official forums and continue to mislead consumers. You can only find information in unofficial forums.
Like the game training system I tested, I only played the FM2024 generation , and now only a 300 hours player, and many previous generations, plus the latest generation, have a lot of players who spend a lot of time researching and sharing their training findings. As a result, these players actually played a placebo simulator
There are even some emotionally invested players who bury their heads in the sand and refuse to believe (forgive me for being overly aggressive) the results of these tests that reveal Physical man is Superman. These players who have invested money and effort should enjoy the same experience as the simulation theory, not the illusion of a fake mechanic.
It's not the player's fault, it's the game company's fault. players love football simulation games, doesn't mean you should accept a flawed game, especially when the game's biggest selling point and publicity point is simulation. You can choose not to play. Or you can stay and protest together in one way or another and force the game company to change. You can also play the game as much as you want, without deliberately using these "intensity mechanics" and have fun.
Awesome work! I'm lacking superlatives to thank you!
The extremes of 1 in some of these attributes are staggering. At the same time, this is also good news! It means you can't just take a player with 20Pac/Acc/Dri and dominate everything. Well-rounded players are still desirable.
I wouldn't have expected that being single footed has such a big impact. This comes in stark contrast to some mainstream advice that it's not worth improving the weak-foot. It also means that it's worth comparing boosting off-foot VS traits like: 'Avoid weaker foot'
The gap between low determination, anticipation, concentration & dribbling is also surprising.
Possebrew said: Awesome work! I'm lacking superlatives to thank you!
The extremes of 1 in some of these attributes are staggering. At the same time, this is also good news! It means you can't just take a player with 20Pac/Acc/Dri and dominate everything. Well-rounded players are still desirable.
I wouldn't have expected that being single footed has such a big impact. This comes in stark contrast to some mainstream advice that it's not worth improving the weak-foot. It also means that it's worth comparing boosting off-foot VS traits like: 'Avoid weaker foot'
The gap between low determination, anticipation, concentration & dribbling is also surprising.
Decision being worthless left me flabbergasted! Expand
The difference between onefooted and both footed is really not that big and considering CA cost I'd argue it's better to invest that into some actual attributes. Could maybe use a test with value of say 6 and 10 or something.
Yarema said: The difference between onefooted and both footed is really not that big and considering CA cost I'd argue it's better to invest that into some actual attributes. Could maybe use a test with value of say 6 and 10 or something. Expand
If I read the screenshot right Non-dominant feet 1 has 2.1 Both feet 20 has 20.1
That's a larger gap than OTB 1 vs OTB 20, or STR 1 vs STR 20.
Comments from community players "Do a rough calculation ,The two-speed (pace Acceleration) ratio can reach about 40% , Except for the ones labeled , Other attributes can be said to be useless , unreasonable"
harvestgreen22 said: Comments from community players "Do a rough calculation ,The two-speed (pace Acceleration) ratio can reach about 40% , Except for the ones labeled , Other attributes can be said to be useless , unreasonable" Expand
Would be interesting to see what happens when Work-Rate is 1 and ACC/PAC are 20. Which one trumps out?
some comments said that the performance of attributes is A multi-variable problem, for example, the A attribute is evaluated at the same time as the B attribute.
Some people will think that testing a player who First touch 10 and Finishing 15, because the player can't stop the ball and it doesn't matter how high the Finishing is, so this test is wrong.
However, if an attribute is weighted with one or more other attributes, all associated attributes must be found. In the end all testing is either pointless or too tedious to do. Since it is statistically possible to tell the difference, let's use this result as a conclusion
Now, imagine a mathematical problem. Player A's Composure is 10, provides 20% of the goal scoring ability. Finishing 10, providing 10 accuracy shots on goal per game. The end result is 10x20%=2 balls.
If this is a multivariate problem, then Composure 20, assuming it provides 50% of the goal scoring ability. The end result is 10x50%=5 balls. So 2 becomes 5. The difference "5-2=3" may or may not be obvious. Those that are not obvious are actually the effect of interaction is not good, since it is so small, it can simply be ignored. If there is an effect, it must be reflected in the statistical results.
Unless it is using some more complex mechanism, for example, it detects composure 20, and if the shot is not ≥15, the goal scoring ability % does not increase. I'm not going to consider that possibility. Expand
But that is the flaw with these tests. If only the attribute 'Off the ball' or 'Flair' is increased, but the player doesn't have the physical or technical attributes to take advantage of the 'Off the ball' or 'Flair', of course it is going to look like a bad attribute. If these attributes are multiplicative of other attributes, and the other attributes are low, then of course these multiplicative attributes will look bad.
Your finishing & composure attributes example is not good. The truth is we don't know how much impact those two attributes have together, to say if it makes no difference, little difference, or a lot of difference.
The only way of making it clear is tedious work: You (or someone else) needs to run tests in groups of three attributes, raising them or lowering them together. This way, we'll gain insight into every 3 attribute combination, and by doing the average of every 3 attribute combination that have the same 2 attributes, we can also gain insight into which 2 attribute combination is good.
If you are correct, then the 3 attribute combination of 'Acceleration'+'Pace'+'Jumping Reach' should perform much better than any other combination, including 'Acceleration'+'Pace'+'Strength', or 'Passing'+'Decisions'+'Work Rate'.
Jolt said: But that is the flaw with these tests. If only the attribute 'Off the ball' or 'Flair' is increased, but the player doesn't have the physical or technical attributes to take advantage of the 'Off the ball' or 'Flair', of course it is going to look like a bad attribute. If these attributes are multiplicative of other attributes, and the other attributes are low, then of course these multiplicative attributes will look bad.
Your finishing & composure attributes example is not good. The truth is we don't know how much impact those two attributes have together, to say if it makes no difference, little difference, or a lot of difference.
The only way of making it clear is tedious work: You (or someone else) needs to run tests in groups of three attributes, raising them or lowering them together. This way, we'll gain insight into every 3 attribute combination, and by doing the average of every 3 attribute combination that have the same 2 attributes, we can also gain insight into which 2 attribute combination is good.
If you are correct, then the 3 attribute combination of 'Acceleration'+'Pace'+'Jumping Reach' should perform much better than any other combination, including 'Acceleration'+'Pace'+'Strength', or 'Passing'+'Decisions'+'Work Rate'. Expand
In another test (tactics not this one), 1. I increase Passing, Crossing, Marking, Dribbling, Tackling, Finishing to 12, Technique Keep 10 unchanged. get A value of A
2. I increase Passing, Crossing, Marking, Dribbling, Tackling, Finishing to 12, Then reduce Technique to 5. get a value of B.
The result is B > A As I understand it as Technique can use Passing, Crossing, Marking, Dribbling, Tackling, Finishing, But reducing it Technique improve performance
Possebrew said: Awesome work! I'm lacking superlatives to thank you!
The extremes of 1 in some of these attributes are staggering. At the same time, this is also good news! It means you can't just take a player with 20Pac/Acc/Dri and dominate everything. Well-rounded players are still desirable.
I wouldn't have expected that being single footed has such a big impact. This comes in stark contrast to some mainstream advice that it's not worth improving the weak-foot. It also means that it's worth comparing boosting off-foot VS traits like: 'Avoid weaker foot'
The gap between low determination, anticipation, concentration & dribbling is also surprising.
Decision being worthless left me flabbergasted! Expand
I have a different take on this test result, As I mentioned above, " Note that the difference between attributes is only used to show the degree of differentiation and cannot be used for direct quantified comparisons.(Because it's not standardized) . For example, "difference from the standard value +40" is not twice as useful as "difference from the standard value +20" . " This result is not "standardized." For example (hypothetical) a +40 attribute can be three times more important than a +20 attribute And a attribute of -40 is three times worse than an attribute of -20 attribute
that means , Ignore the Work rate, Paceand Acceleration have an amazing 171 and 161, which means that the importance of physical attributes is not limited to 61.47% in the picture, the importance of physical attributes must be much greater than 61.47%
Yarema said: The difference between onefooted and both footed is really not that big and considering CA cost I'd argue it's better to invest that into some actual attributes. Could maybe use a test with value of say 6 and 10 or something. Expand
harvestgreen22 said: I have a different take on this test result, As I mentioned above, " Note that the difference between attributes is only used to show the degree of differentiation and cannot be used for direct quantified comparisons.(Because it's not standardized) . For example, "difference from the standard value +40" is not twice as useful as "difference from the standard value +20" . " This result is not "standardized." For example (hypothetical) a +40 attribute can be three times more important than a +20 attribute And a attribute of -40 is three times worse than an attribute of -20 attribute
that means , Ignore the Work rate, Paceand Acceleration have an amazing 171 and 161, which means that the importance of physical attributes is not limited to 61.47% in the picture, the importance of physical attributes must be much greater than 61.47% Expand
" Note that the difference between attributes is only used to show the degree of differentiation and cannot be used for direct quantified comparisons.(Because it's not standardized) . For example, "difference from the standard value +40" is not twice as useful as "difference from the standard value +20" . "
Or to put it in a better way, It's easier to score 2 goals in 1 match, But it's not just 500% harder to score 10 goals than to score 2 goals , it's a lot harder than 500%
so ,with standard Control Group 20.1, a stats with Goal difference (GD) 60.0 is much much more important than a stats with Goal difference (GD) 40.0 The closer a Stats is to 20.1, the less important it is
@harvestgreen22 With this data, I wonder if the training schedules data can have Dribbling and Balance growth added with Pace, Acceleration and Jumping Reach.
Concentration, Anticipation, Determination，Agility can be used to separate ties in the transfer market but not a focus in training?
Thank you for this test. As suspected a little bit of weak foot ability is good, but after that doesn't seem to be worth the cost. And for that reason I don't really like players who are good with both feet.
There might be an argument to be made to train weak foot of players who are terrible at it.
Jolt said: But that is the flaw with these tests. If only the attribute 'Off the ball' or 'Flair' is increased, but the player doesn't have the physical or technical attributes to take advantage of the 'Off the ball' or 'Flair', of course it is going to look like a bad attribute. If these attributes are multiplicative of other attributes, and the other attributes are low, then of course these multiplicative attributes will look bad.
Your finishing & composure attributes example is not good. The truth is we don't know how much impact those two attributes have together, to say if it makes no difference, little difference, or a lot of difference.
The only way of making it clear is tedious work: You (or someone else) needs to run tests in groups of three attributes, raising them or lowering them together. This way, we'll gain insight into every 3 attribute combination, and by doing the average of every 3 attribute combination that have the same 2 attributes, we can also gain insight into which 2 attribute combination is good.
If you are correct, then the 3 attribute combination of 'Acceleration'+'Pace'+'Jumping Reach' should perform much better than any other combination, including 'Acceleration'+'Pace'+'Strength', or 'Passing'+'Decisions'+'Work Rate'. Expand
I'm trying more combinations, and maybe some of them are interacting with Technique and Flair
Han106 said: @harvestgreen22 With this data, I wonder if the training schedules data can have Dribbling and Balance growth added with Pace, Acceleration and Jumping Reach.
Concentration, Anticipation, Determination，Agility can be used to separate ties in the transfer market but not a focus in training? Expand
Yes. Seeing if these secondary important attributes can also grow should also be a measure of the quality of the training Recently, I didn't have the energy to re-test the training program to see how these attributes grew as a percentage of total growth , as I am busy testing various attributes
Jolt said: But that is the flaw with these tests. If only the attribute 'Off the ball' or 'Flair' is increased, but the player doesn't have the physical or technical attributes to take advantage of the 'Off the ball' or 'Flair', of course it is going to look like a bad attribute. If these attributes are multiplicative of other attributes, and the other attributes are low, then of course these multiplicative attributes will look bad.
Your finishing & composure attributes example is not good. The truth is we don't know how much impact those two attributes have together, to say if it makes no difference, little difference, or a lot of difference.
The only way of making it clear is tedious work: You (or someone else) needs to run tests in groups of three attributes, raising them or lowering them together. This way, we'll gain insight into every 3 attribute combination, and by doing the average of every 3 attribute combination that have the same 2 attributes, we can also gain insight into which 2 attribute combination is good.
If you are correct, then the 3 attribute combination of 'Acceleration'+'Pace'+'Jumping Reach' should perform much better than any other combination, including 'Acceleration'+'Pace'+'Strength', or 'Passing'+'Decisions'+'Work Rate'. Expand
More testing of the effects of multivariables，
1. Passing传球15 Crossing传中15 Dribbling盘带15 Tackling抢断15 Finishing射门15 First touch停球15 Heading头球15 longshot远射15
Technique技术10 Flair才华10 other其他属性10
Goal difference 净胜球≈73.7
2. Passing传球15 Crossing传中15 Dribbling盘带15 Tackling抢断15 Finishing射门15 First touch停球15 Heading头球15 longshot远射15
Technique技术15 Flair才华10 other其他属性10
Goal difference 净胜球≈60.3
3. Passing传球15 Crossing传中15 Dribbling盘带15 Tackling抢断15 Finishing射门15 First touch停球15 Heading头球15 longshot远射15
Summary: 7. Under Control variable, From 1, 2, and 3, you can see, Even in the multivariable case, considering multiple technical type of attributes,
technical type of attributes 15 + Technique 10 (73.7) > technical type of attributes 15+ Technique 15 (60.3) ，This means that in this case, Technique is bad technical type of attributes 15 + Flair 10 (73.7) ≈/> technical type of attributes 15 + Flair 10 (71.7) ，This means that in this case, Flair have little/no influence
8. In the case of 4, 5, and 6, it's a little special, Dribbling 20 + Technique 20 (44.2) < Dribbling 20 + Technique 10 (53.6) Higher Technique is having a negative effect with Dribbling This means that in this case, Technique is bad
Passing 20 + Technique 20 (27.9) ≈/> Passing 20 + Technique 10 (24.5) Higher Technique may or may not have a positive effect with Passing Not much difference. It is also possible that the sample size is not large enough and resulting in random error
Finishing 20 + Technique 20 (33.7) ≈ Finishing 20 + Technique 10 (33.3) Higher Technique had no effect on Finishing passing
9. What is surprising is number 5, which means that it is possible for Technique to have some positive Stats and negative effects on others
harvestgreen22 said: I have a different take on this test result, As I mentioned above, " Note that the difference between attributes is only used to show the degree of differentiation and cannot be used for direct quantified comparisons.(Because it's not standardized) . For example, "difference from the standard value +40" is not twice as useful as "difference from the standard value +20" . " This result is not "standardized." For example (hypothetical) a +40 attribute can be three times more important than a +20 attribute And a attribute of -40 is three times worse than an attribute of -20 attribute
that means , Ignore the Work rate, Paceand Acceleration have an amazing 171 and 161, which means that the importance of physical attributes is not limited to 61.47% in the picture, the importance of physical attributes must be much greater than 61.47% Expand
I feel the focus might be too much on identifying the "best attributes" for enhancing performance. However, the significant negative impact of attributes like Work Rate at 1 underscores the importance of avoiding negative attributes just as much as maximizing positive ones.
Consider this hypothetical scenario: If everyone on the team has a Work Rate below 5, attributes like 20 in Acceleration and Pace could become irrelevant. It's like having a team of Usain Bolts who prefer to chat rather than run because they find the effort tedious. This principle likely applies to other attributes as well.
Here's how I think about it:
- Any attribute that pulls your team's performance below the control group's 20.1 goal difference is critical to consider.
The sequence of attribute impact might work like this (and it's how I would design such a system):
1. Decision Check: First, the game checks if the player decides to act or not. 2. Off-the-Ball Intelligence: Then, it determines if the player knows where to move. 3. Pace and Acceleration: These attributes determine how quickly they can get to that position. 4. First Touch: Checks if the player can control the ball when they receive it. 5. Dribbling: Decides how effective the player is with the ball in terms of movement. 6. Agility and Balance: These determine how well the player can avoid being tackled or recover if they are. 7. Vision and Decision: Another check to see if the player chooses to shoot or pass. 8. Type of Action: Decides the specifics of the pass or shot. 9. Shot Outcome: If shooting, this is measured against the goalkeeper's attributes to see if a goal is scored.
Each of these steps would be influenced by the relative attributes of the players involved, not just the absolute values. For instance, if all players had 20 in Acceleration and Pace, the value of these attributes would diminish because it's the differential between players that matters, not the absolute attribute itself.
It's also crucial to distinguish whether an attribute is more relevant for the interaction between players (like a striker versus a defender) or for the individual's own performance metrics (like shooting accuracy).
harvestgreen22 said: * This is translated by a language translator and then manually corrected, there may be some loss of meaning after translation
Test setup: The league has 4 teams, A, C, D team is set to 4231 Meta tactics (https://fm-arena.com/thread/12667-god-of-chaos-v1/), B team is set to 4141 weak tactics, all attributes of all players are set to 10, locking various attributes.
Modify the player attributes of Team A (except for the goalkeeper, that is, adjust 10 players) to the corresponding value. Statistical the goal difference.
The C and D teams use a very strong 4231 tactics, which tests the players' defensive ability and performance under pressure from strong teams. The B team uses the weak 4141 tactics to test the players' ability to Beat up weak teams and score lots of goals.
It took 10 days to test, two computers together. 2 of the attributes were tested by the help of the community netizen , and the other 70 were tested by me. Inspired by this https://fm-arena.com/table/26-player-attributes-testing/ (November 2023 version)I thought, What if the attribute is 1 ?
1.How to read this table?
The standard(Control) group represents the goal difference of Team A when all attributes are 10, which is 20.1.
## Look at Technique 20, which represents Team A (apart from the goalkeeper) skill attribute increases from 10 to 20, and Statistical the goal difference of Team A, it's 16.5. 16.5 < 20.1 (standard) In other words, the Technique has increased by 10, but the goal difference has decreased. But not by much.
## Look at Vision 1 and Vision 20, 20.0 and 38.5 respectively. 20.0 < 20.1 < 38.5 In other words, when Vision is reduced from 10 to 1, so there is no change. However, when Vision was increased from 10 to 20, resulting in a goal difference increase of (38.5-20.1)=18.4
2. Noteworthy attributes
Lowering the Work rate 1 , it have very serious consequences (goal difference -110),
and it should be ensured that a player has at least 6 and preferably 10 Work rate. A player with a Work rate is not desirable. But continuing to increase Work rate (from 10->20) is low effect. The difference between Work rate 10 and Work rate 20 is relative small.
The 6-point attribute seems to be the threshold for some attributes.
Passing, Crossing, Marking, First touch, Positioning, Decision, Teamwork, Off the ball, Bravery, are all attributes that can be considered useless.
Technique and Flair, Higher values lead to worse performance. Spoiler
Yes, it's a scam. To Flair. There are two guesses, one is that "fancy tricks" cause players to spend more "frames" on the movement, resulting in a decrease in efficiency. The other is that more "fancy tricks" contribute not enough to the mathematical expectation when attacking, while the mathematical expectation of being scored when defending increases more than attacking scored
Core attributes are these: Special : Work rate (need to reach 10, higher is useless), Pace, Acceleration, Jumping reach, Dribbling, Balance, Concentration, Anticipation, Determination，Agility.
According to the comparison of attribute 20 with attribute 10: Pace + Acceleration 2 most important cores, Jumping reach + Dribbling 2 secondary cores, Balance can be considered a level 3 attribute, level 4 attribute Anticipation and Concentration
Other attributes, which are although have effect, but much less effective.
Note that the difference between attributes is only used to show the degree of differentiation and cannot be used for direct quantified comparisons.(Because it's not standardized) . For example, "difference from the standard value +40" is not twice as useful as "difference from the standard value +20" .
3.Sample
According to the Law of Large Numbers in statistics, as long as the sample is large enough, the value you get will be closer and closer to the mathematical expectation
In terms of important attributes, there are about 950-1080 Match samples. According to statistical estimates, the margin error of attribute is ≤ 3.
In the non-important attributes, there are about 450 Match samples. According to statistical estimates, the margin error of attribute is ≤ 8
4.margin error
Is that a conclusion that only applies to 4231 and certain opponents? You know, tactics can vary a lot. Moreover, 4231 players are not evenly distributed, for example, there is only 1 striker. And tactics, duties, and commands all affect the importance of attributes. There is also the test of the Go-on-holiday, even if the game is set to maximum detail, will it get unreal values?
——Go-on-holiday have been studied, and I won't repeat it in a lot of words, but I choose to trust their findings
——Taking into account different tactics and different opponents, an additional 3,600 match were played, To test two other tactics, one is Default tiki-taka 433. And Default counter-attack 442. They will have different standard values and different attributes deviating from the standard, but "importance level" of each attribute not change.
For example, in the tactics where Passing is more important, Passing (around 24 in the table) is still not important than dribbling (more than 50 in the table) or Finishing (33 in the table) in any tactic. This also shows that the game is not complicated to do a separate set of calculations for each tactic, but a relatively universal calculation method. And this method of calculation, my personal observation, is more conducive to offensive.
——Similarly, the problem of player distribution does not have a "decisive" effect, but only a statistical perturbation of some value.
5.Multiple variables
some comments said that the performance of attributes is A multi-variable problem, for example, the A attribute is evaluated at the same time as the B attribute.
Some people will think that testing a player who First touch 10 and Finishing 15, because the player can't stop the ball and it doesn't matter how high the Finishing is, so this test is wrong.
However, if an attribute is weighted with one or more other attributes, all associated attributes must be found. In the end all testing is either pointless or too tedious to do. Since it is statistically possible to tell the difference, let's use this result as a conclusion
Now, imagine a mathematical problem. Player A's Composure is 10, provides 20% of the goal scoring ability. Finishing 10, providing 10 accuracy shots on goal per game. The end result is 10x20%=2 balls.
If this is a multivariate problem, then Composure 20, assuming it provides 50% of the goal scoring ability. The end result is 10x50%=5 balls. So 2 becomes 5. The difference "5-2=3" may or may not be obvious. Those that are not obvious are actually the effect of interaction is not good, since it is so small, it can simply be ignored. If there is an effect, it must be reflected in the statistical results.
Unless it is using some more complex mechanism, for example, it detects composure 20, and if the shot is not ≥15, the goal scoring ability % does not increase. I'm not going to consider that possibility.
6.don't bring the logic of reality into the game
For example, there is such a reply: there is no "speed 20" and other attributes are 10 in reality . Or, "Should use balabala" attributes to design all forwards/centers/back guards to reflect the real situation.
The results, as long as the control variable method is followed, make no difference (since the game seems unlikely to use the "more complex" mechanics mentioned above).
7. Why not test each player individually, such as the striker
Because it's so much work, And the impact each player has on the team's goal difference is likely to be the difference between the actual measured value and the standard value, which is always disturbed by larger random perturbations, because a tactic is mostly 1-3 strikers, who make up less than 30% of the 11 men. If you want to measure the true value, you need a very, very large number of samples
8. The goalkeeper? Similar to the above reasons. mishap If you want to measure the true value, you need a very, very large number of samples
9. Why testing just one season The result is completely inaccurate? (from reply) The random perturbation in the game is very large, which means that if you only play 30 to 40 games (a season), the result is likely to deviate significantly from the true value
One might also suspect that the test is not rigorous enough, it's not scientific enough, it's not an actual season, it's a nice ideal environment, and so on, and then argue that the empirical results under a lot of interference are more accurate than the results obtained under the control variable method
10.thoughts
Like the previous "1CA Win Champion", many people will dismiss this as a conspiracy theory before testing reveals the game mechanics. After the revelation, some players believe that this kind of testing is just excessive pursuit of intensity at the expense of game life. But that's not the purpose of the test.
My goal in testing this is the same as the goal of many gamers: to force the game company to continuously improve the old mechanics it has used for years, to change the fundamental engine flaws, and to prevent players from playing a "placebo simulator". Why a placebo simulator? There are people who take real football and put a lot of time and energy into the game, there are a lot of players who put a lot of energy into the video, I also enjoy these passionate things But, As a result, the game engine does not do these functions at all, these players believe and carefully prepared, essentially reflected in the game engine is useless. The game relies on some very simple data and then randomizes it to make players think they are playing a real simulation. Then the game company can just change the appearance of their game, sell it for generations, and advertise that they are making a simulation full of interactivity
What's more, the authorities have banned discussions on official forums and continue to mislead consumers. You can only find information in unofficial forums.
Like the game training system I tested, I only played the FM2024 generation , and now only a 300 hours player, and many previous generations, plus the latest generation, have a lot of players who spend a lot of time researching and sharing their training findings. As a result, these players actually played a placebo simulator
There are even some emotionally invested players who bury their heads in the sand and refuse to believe (forgive me for being overly aggressive) the results of these tests that reveal Physical man is Superman. These players who have invested money and effort should enjoy the same experience as the simulation theory, not the illusion of a fake mechanic.
It's not the player's fault, it's the game company's fault. players love football simulation games, doesn't mean you should accept a flawed game, especially when the game's biggest selling point and publicity point is simulation. You can choose not to play. Or you can stay and protest together in one way or another and force the game company to change. You can also play the game as much as you want, without deliberately using these "intensity mechanics" and have fun. Expand
My guess for why high Technical and Decisions is detrimental is because the game does not know what actions are the most effective. If you look at technical attributes, you will see that dribbling is more efficient than passing or shooting, possibly because pace and acceleration are over tuned in the engine, and dribbling makes use of both attributes. So, characters with high technical will have increased chance of success for passes and shots, while characters with high decisions will find better opportunities to pass and shoot. Unfortunately, the most effective is to pass and shoot as little as possible, unless you have those attributes much higher than dribbling attributes (including pace and acceleration).
So, in short, if I am not wrong: Pass 20 + Technique 20 + Other 10 > Pass 20 + Technique 10 + Other 10. Pass 20 + Technique 10 + Dribbling 20 + Pace 20 + Acceleration 20 + Other 10 > Pass 20 + Technique 20 + Dribbling 20 + Pace 20 + Acceleration 20 + Other 10.
The same is probably true for long shots or finishing.
ZaZ said: My guess for why high Technical and Decisions is detrimental is because the game does not know what actions are the most effective. If you look at technical attributes, you will see that dribbling is more efficient than passing or shooting, possibly because pace and acceleration are over tuned in the engine, and dribbling makes use of both attributes. So, characters with high technical will have increased chance of success for passes and shots, while characters with high decisions will find better opportunities to pass and shoot. Unfortunately, the most effective is to pass and shoot as little as possible, unless you have those attributes much higher than dribbling attributes (including pace and acceleration).
So, in short, if I am not wrong: Pass 20 + Technique 20 + Other 10 > Pass 20 + Technique 10 + Other 10. Pass 20 + Technique 10 + Dribbling 20 + Pace 20 + Acceleration 20 + Other 10 > Pass 20 + Technique 20 + Dribbling 20 + Pace 20 + Acceleration 20 + Other 10.
The same is probably true for long shots or finishing. Expand
1. Passing传球15 Crossing传中15 Dribbling盘带15 Tackling抢断15 Finishing射门15 First touch停球15 Heading头球15 longshot远射15
Technique技术10 Flair才华10 other其他属性10
Goal difference 净胜球≈73.7
2. Passing传球15 Crossing传中15 Dribbling盘带15 Tackling抢断15 Finishing射门15 First touch停球15 Heading头球15 longshot远射15
Technique技术15 Flair才华10 other其他属性10
Goal difference 净胜球≈60.3
3. Passing传球15 Crossing传中15 Dribbling盘带15 Tackling抢断15 Finishing射门15 First touch停球15 Heading头球15 longshot远射15
Technique技术10 Flair才华15 other其他属性10
Goal difference 净胜球≈71.7
4. Passing传球15 Crossing传中15 Dribbling盘带15 Tackling抢断15 Finishing射门15 First touch停球15 Heading头球15 longshot远射15
Technique技术15 Flair才华15 other其他属性10
Goal difference 净胜球≈58.9
I added a situation (situation 4.), They're both at 15, Goal difference the lowest. don't know how to explain that mechanism I guess it's not as complicated as we might think: It's a bug,Or the mechanics aren't right
Thank you so much for all the great work! It is great that you have bothered to do all these tests so we can know more about which attributes are more useful in the game! This is all greatly appreciated!
However, I do have one question which, depending on your answer, may explain why you are finding that some of the attributes have a negative impact on performance (e.g., technique): - How are you dealing with the current ability (CA) of players not being the same after you raise (or lower) one (multiple) of the attributes? -> When you set all players' attributes to 10, their CA needs to be set to a certain level such that the attributes remain stable... when you change one of the attributes to 20 (100 in the "centi-attributes" that the game actually uses and that you can see in the save game editor), you need a higher CA amount for the attributes to remain stable (and this higher CA amount will depend on the player's position given that, for example, finishing has a larger CA weight for forwards than for defenders). Otherwise, the game will lower attributes across the board so that the player's CA remains the same... -> This means that if you are running a league for your tests that takes any amount of game time and you don't either (1) make sure that the player's CA is what is needed to maintain the attributes you set initially or (2) have some mechanism to ensure that the player's attributes remain frozen (it is my understanding that some save game editors may be able to do this), then the game will be changing other attributes to compensate for the attribute that you maxed out (or set to 1). -> This would mean that when, for example, you set technique to 20, other attributes are lowered below 10 to compensate (you may not always be able to see this when attributes are represented 1-20 in-game, but this may still be happening under the hood in the attributes represented 1-100 which you can see in savegame editors). -> If you are not accounting for this attribute re-balancing, then your finding that technique has a negative impact on goal difference in your test may not mean that technique literally has a negative impact on player performance, but rather that setting technique to 20 is not worth the reduction in other attributes that it causes...
Anyway, I would be interested in knowing how you dealt with attribute re-balancing due to the player's CA in your tests.
------------------- In addition, I'd like to suggest that you also try testing "aggression" and "leadership." Both may have some influence in performance. Aggression may interact with certain attributes, such as tackling or bravery.
Also could be interesting if you tested the important matches attribute (having in mind that the influence of this attribute depends on how important the match is - see https://www.youtube.com/watch?v=ECe0ygapG-8&t=1404s)
Thank you so much for all the great work! It is great that you have bothered to do all these tests so we can know more about which attributes are more useful in the game! This is all greatly appreciated!
However, I do have one question which, depending on your answer, may explain why you are finding that some of the attributes have a negative impact on performance (e.g., technique): - How are you dealing with the current ability (CA) of players not being the same after you raise (or lower) one (multiple) of the attributes? -> When you set all players' attributes to 10, their CA needs to be set to a certain level such that the attributes remain stable... when you change one of the attributes to 20 (100 in the "centi-attributes" that the game actually uses and that you can see in the save game editor), you need a higher CA amount for the attributes to remain stable (and this higher CA amount will depend on the player's position given that, for example, finishing has a larger CA weight for forwards than for defenders). Otherwise, the game will lower attributes across the board so that the player's CA remains the same... -> This means that if you are running a league for your tests that takes any amount of game time and you don't either (1) make sure that the player's CA is what is needed to maintain the attributes you set initially or (2) have some mechanism to ensure that the player's attributes remain frozen (it is my understanding that some save game editors may be able to do this), then the game will be changing other attributes to compensate for the attribute that you maxed out (or set to 1). -> This would mean that when, for example, you set technique to 20, other attributes are lowered below 10 to compensate (you may not always be able to see this when attributes are represented 1-20 in-game, but this may still be happening under the hood in the attributes represented 1-100 which you can see in savegame editors). -> If you are not accounting for this attribute re-balancing, then your finding that technique has a negative impact on goal difference in your test may not mean that technique literally has a negative impact on player performance, but rather that setting technique to 20 is not worth the reduction in other attributes that it causes...
Anyway, I would be interested in knowing how you dealt with attribute re-balancing due to the player's CA in your tests.
------------------- In addition, I'd like to suggest that you also try testing "aggression" and "leadership." Both may have some influence in performance. Aggression may interact with certain attributes, such as tackling or bravery.
Also could be interesting if you tested the important matches attribute (having in mind that the influence of this attribute depends on how important the match is - see https://www.youtube.com/watch?v=ECe0ygapG-8&t=1404s) Expand
I manually changed the "Current CA" to the "Recommended CA" in every test every time (I don't know if the translator misunderstands your meaning.)
like this: I used the in-game modifier. Did that make a difference to the results? Or do I need other non-in-game tools?
This removes my objection... I am still having trouble believing in some of the results (just like the results in the other FM Arena testing - https://fm-arena.com/table/26-player-attributes-testing/), but I don't have an alternative theory that can explain them at the moment.
At this point it would be more interesting to test position rating systems to see if it can consistently produce winners against a + 2 or + 4 acc/pac team with the secondary meta attributes + stamina/wr set to 10 & the rest of the worthless attributes set to 1. ie A team with of 14 acc/pac with anything goes vs 16/16 acc/pac or 18 acc/pac team.
Like often times you are picking between a 15/15 AMRL that has 16 dribbling or 15 jumping reach, and a 17/17 player that has 3 jumping reach & 10 dribbling.
bigloser said: At this point it would be more interesting to test position rating systems to see if it can consistently produce winners against a + 2 or + 4 acc/pac team with the secondary meta attributes + stamina/wr set to 10 & the rest of the worthless attributes set to 1. ie A team with of 14 acc/pac with anything goes vs 16/16 acc/pac or 18 acc/pac team.
Like often times you are picking between a 15/15 AMRL that has 16 dribbling or 15 jumping reach, and a 17/17 player that has 3 jumping reach & 10 dribbling. Expand
from experience, 17/17 player is still better. How often are there a 15/15 with 15JR + 15 DRB? its usually like a super Dribbly 15/15/15 or a 15/15/8 with 16 JR.
Not that it matters that much because hidden attributes is just as important, so a good looking player might just be a dude if pressure, injury proness or dirtiness is bad.
* This is translated by a language translator and then manually corrected, there may be some loss of meaning after translation
Test setup: The league has 4 teams, A, C, D team is set to 4231 Meta tactics (https://fm-arena.com/thread/12667-god-of-chaos-v1/), B team is set to 4141 weak tactics, all attributes of all players are set to 10, locking various attributes.
Modify the player attributes of Team A (except for the goalkeeper, that is, adjust 10 players) to the corresponding value.
Statistical the goal difference.
The C and D teams use a very strong 4231 tactics, which tests the players' defensive ability and performance under pressure from strong teams.
The B team uses the weak 4141 tactics to test the players' ability to Beat up weak teams and score lots of goals.
It took 10 days to test, two computers together. 2 of the attributes were tested by the help of the community netizen , and the other 70 were tested by me.
Inspired by this https://fm-arena.com/table/26-player-attributes-testing/ (November 2023 version)I thought, What if the attribute is 1 ?
1.How to read this table?
The standard(Control) group represents the goal difference of Team A when all attributes are 10, which is 20.1.
## Look at Technique 20, which represents Team A (apart from the goalkeeper) skill attribute increases from 10 to 20, and Statistical the goal difference of Team A, it's 16.5.
16.5 < 20.1 (standard)
In other words, the Technique has increased by 10, but the goal difference has decreased. But not by much.
## Look at Vision 1 and Vision 20, 20.0 and 38.5 respectively.
20.0 < 20.1 < 38.5
In other words, when Vision is reduced from 10 to 1, so there is no change.
However, when Vision was increased from 10 to 20, resulting in a goal difference increase of (38.5-20.1)=18.4
2. Noteworthy attributes
Lowering the Work rate 1 ,
it have very serious consequences (goal difference -110),
and it should be ensured that a player has at least 6 and preferably 10 Work rate.
A player with a Work rate is not desirable.
But continuing to increase Work rate (from 10->20) is low effect. The difference between Work rate 10 and Work rate 20 is relative small.
The 6-point attribute seems to be the threshold for some attributes.
Passing, Crossing, Marking, First touch, Positioning, Decision, Teamwork, Off the ball, Bravery, are all attributes that can be considered useless.
Technique and Flair, Higher values lead to worse performance.
Spoiler Yes, it's a scam.
To Flair. There are two guesses,
one is that "fancy tricks" cause players to spend more "frames" on the movement, resulting in a decrease in efficiency.
The other is that more "fancy tricks" contribute not enough to the mathematical expectation when attacking, while the mathematical expectation of being scored when defending increases more than attacking scored
Core attributes are these:
Special : Work rate (need to reach 10, higher is useless),
Pace, Acceleration, Jumping reach, Dribbling, Balance, Concentration, Anticipation, Determination，Agility.
According to the comparison of attribute 20 with attribute 10:
Pace + Acceleration 2 most important cores,
Jumping reach + Dribbling 2 secondary cores,
Balance can be considered a level 3 attribute,
level 4 attribute Anticipation and Concentration
Other attributes, which are although have effect, but much less effective.
Note that the difference between attributes is only used to show the degree of differentiation and cannot be used for direct quantified comparisons.(Because it's not standardized) . For example, "difference from the standard value +40" is not twice as useful as "difference from the standard value +20" .
3.Sample
According to the Law of Large Numbers in statistics, as long as the sample is large enough, the value you get will be closer and closer to the mathematical expectation
In terms of important attributes, there are about 950-1080 Match samples. According to statistical estimates, the margin error of attribute is ≤ 3.
In the non-important attributes, there are about 450 Match samples. According to statistical estimates, the margin error of attribute is ≤ 8
4.margin error
Is that a conclusion that only applies to 4231 and certain opponents?
You know, tactics can vary a lot.
Moreover, 4231 players are not evenly distributed, for example, there is only 1 striker.
And tactics, duties, and commands all affect the importance of attributes.
There is also the test of the Go-on-holiday, even if the game is set to maximum detail, will it get unreal values?
——Go-on-holiday have been studied, and I won't repeat it in a lot of words, but I choose to trust their findings
——Taking into account different tactics and different opponents, an additional 3,600 match were played,
To test two other tactics, one is Default tiki-taka 433. And Default counter-attack 442.
They will have different standard values and different attributes deviating from the standard, but "importance level" of each attribute not change.
For example, in the tactics where Passing is more important,
Passing (around 24 in the table) is still not important than dribbling (more than 50 in the table) or Finishing (33 in the table) in any tactic.
This also shows that the game is not complicated to do a separate set of calculations for each tactic, but a relatively universal calculation method.
And this method of calculation, my personal observation, is more conducive to offensive.
——Similarly, the problem of player distribution does not have a "decisive" effect, but only a statistical perturbation of some value.
5.Multiple variables
some comments said that the performance of attributes is A multi-variable problem, for example, the A attribute is evaluated at the same time as the B attribute.
Some people will think that testing a player who First touch 10 and Finishing 15, because the player can't stop the ball and it doesn't matter how high the Finishing is, so this test is wrong.
However, if an attribute is weighted with one or more other attributes, all associated attributes must be found. In the end all testing is either pointless or too tedious to do.
Since it is statistically possible to tell the difference, let's use this result as a conclusion
Now, imagine a mathematical problem. Player A's Composure is 10, provides 20% of the goal scoring ability. Finishing 10, providing 10 accuracy shots on goal per game. The end result is 10x20%=2 balls.
If this is a multivariate problem, then Composure 20, assuming it provides 50% of the goal scoring ability. The end result is 10x50%=5 balls. So 2 becomes 5. The difference "5-2=3" may or may not be obvious.
Those that are not obvious are actually the effect of interaction is not good, since it is so small, it can simply be ignored.
If there is an effect, it must be reflected in the statistical results.
Unless it is using some more complex mechanism, for example, it detects composure 20, and if the shot is not ≥15, the goal scoring ability % does not increase. I'm not going to consider that possibility.
6.don't bring the logic of reality into the game
For example, there is such a reply: there is no "speed 20" and other attributes are 10 in reality .
Or, "Should use balabala" attributes to design all forwards/centers/back guards to reflect the real situation.
The results, as long as the control variable method is followed, make no difference (since the game seems unlikely to use the "more complex" mechanics mentioned above).
7. Why not test each player individually, such as the striker
Because it's so much work,
And the impact each player has on the team's goal difference is likely to be the difference between the actual measured value and the standard value, which is always disturbed by larger random perturbations, because a tactic is mostly 1-3 strikers, who make up less than 30% of the 11 men.
If you want to measure the true value, you need a very, very large number of samples
8. The goalkeeper?
Similar to the above reasons. mishap
If you want to measure the true value, you need a very, very large number of samples
9. Why testing just one season The result is completely inaccurate? (from reply)
The random perturbation in the game is very large, which means that if you only play 30 to 40 games (a season), the result is likely to deviate significantly from the true value
One might also suspect that the test is not rigorous enough,
it's not scientific enough,
it's not an actual season,
it's a nice ideal environment,
and so on,
and then argue that the empirical results under a lot of interference are more accurate than the results obtained under the control variable method
10.thoughts
Like the previous "1CA Win Champion", many people will dismiss this as a conspiracy theory before testing reveals the game mechanics. After the revelation, some players believe that this kind of testing is just excessive pursuit of intensity at the expense of game life. But that's not the purpose of the test.
My goal in testing this is the same as the goal of many gamers: to force the game company to continuously improve the old mechanics it has used for years, to change the fundamental engine flaws, and to prevent players from playing a "placebo simulator". Why a placebo simulator? There are people who take real football and put a lot of time and energy into the game, there are a lot of players who put a lot of energy into the video, I also enjoy these passionate things
But,
As a result, the game engine does not do these functions at all, these players believe and carefully prepared, essentially reflected in the game engine is useless. The game relies on some very simple data and then randomizes it to make players think they are playing a real simulation. Then the game company can just change the appearance of their game, sell it for generations, and advertise that they are making a simulation full of interactivity
What's more, the authorities have banned discussions on official forums and continue to mislead consumers. You can only find information in unofficial forums.
Like the game training system I tested, I only played the FM2024 generation , and now only a 300 hours player, and many previous generations, plus the latest generation, have a lot of players who spend a lot of time researching and sharing their training findings.
As a result, these players actually played a placebo simulator
There are even some emotionally invested players who bury their heads in the sand and refuse to believe (forgive me for being overly aggressive) the results of these tests that reveal Physical man is Superman. These players who have invested money and effort should enjoy the same experience as the simulation theory, not the illusion of a fake mechanic.
It's not the player's fault, it's the game company's fault. players love football simulation games, doesn't mean you should accept a flawed game, especially when the game's biggest selling point and publicity point is simulation. You can choose not to play. Or you can stay and protest together in one way or another and force the game company to change. You can also play the game as much as you want, without deliberately using these "intensity mechanics" and have fun.
Awesome work! I'm lacking superlatives to thank you!
The extremes of 1 in some of these attributes are staggering.
At the same time, this is also good news!
It means you can't just take a player with 20Pac/Acc/Dri and dominate everything.
Well-rounded players are still desirable.
I wouldn't have expected that being single footed has such a big impact. This comes in stark contrast to some mainstream advice that it's not worth improving the weak-foot. It also means that it's worth comparing boosting off-foot VS traits like: 'Avoid weaker foot'
The gap between low determination, anticipation, concentration & dribbling is also surprising.
Decision being worthless left me flabbergasted!
Possebrew said: Awesome work! I'm lacking superlatives to thank you!
The extremes of 1 in some of these attributes are staggering.
At the same time, this is also good news!
It means you can't just take a player with 20Pac/Acc/Dri and dominate everything.
Well-rounded players are still desirable.
I wouldn't have expected that being single footed has such a big impact. This comes in stark contrast to some mainstream advice that it's not worth improving the weak-foot. It also means that it's worth comparing boosting off-foot VS traits like: 'Avoid weaker foot'
The gap between low determination, anticipation, concentration & dribbling is also surprising.
Decision being worthless left me flabbergasted!
The difference between onefooted and both footed is really not that big and considering CA cost I'd argue it's better to invest that into some actual attributes. Could maybe use a test with value of say 6 and 10 or something.
Yarema said: The difference between onefooted and both footed is really not that big and considering CA cost I'd argue it's better to invest that into some actual attributes. Could maybe use a test with value of say 6 and 10 or something.
If I read the screenshot right
Non-dominant feet 1 has 2.1
Both feet 20 has 20.1
That's a larger gap than OTB 1 vs OTB 20, or STR 1 vs STR 20.
add 210cm 75kg
Comments from community players
"Do a rough calculation ,The two-speed (pace Acceleration) ratio can reach about 40% ,
Except for the ones labeled , Other attributes can be said to be useless , unreasonable"
harvestgreen22 said:
Comments from community players
"Do a rough calculation ,The two-speed (pace Acceleration) ratio can reach about 40% ,
Except for the ones labeled , Other attributes can be said to be useless , unreasonable"
Would be interesting to see what happens when Work-Rate is 1 and ACC/PAC are 20.
Which one trumps out?
add
Non-dominant feet = 6
Non-dominant feet = 10
harvestgreen22 said: 5.Multiple variables
some comments said that the performance of attributes is A multi-variable problem, for example, the A attribute is evaluated at the same time as the B attribute.
Some people will think that testing a player who First touch 10 and Finishing 15, because the player can't stop the ball and it doesn't matter how high the Finishing is, so this test is wrong.
However, if an attribute is weighted with one or more other attributes, all associated attributes must be found. In the end all testing is either pointless or too tedious to do.
Since it is statistically possible to tell the difference, let's use this result as a conclusion
Now, imagine a mathematical problem. Player A's Composure is 10, provides 20% of the goal scoring ability. Finishing 10, providing 10 accuracy shots on goal per game. The end result is 10x20%=2 balls.
If this is a multivariate problem, then Composure 20, assuming it provides 50% of the goal scoring ability. The end result is 10x50%=5 balls. So 2 becomes 5. The difference "5-2=3" may or may not be obvious.
Those that are not obvious are actually the effect of interaction is not good, since it is so small, it can simply be ignored.
If there is an effect, it must be reflected in the statistical results.
Unless it is using some more complex mechanism, for example, it detects composure 20, and if the shot is not ≥15, the goal scoring ability % does not increase. I'm not going to consider that possibility.
But that is the flaw with these tests. If only the attribute 'Off the ball' or 'Flair' is increased, but the player doesn't have the physical or technical attributes to take advantage of the 'Off the ball' or 'Flair', of course it is going to look like a bad attribute. If these attributes are multiplicative of other attributes, and the other attributes are low, then of course these multiplicative attributes will look bad.
Your finishing & composure attributes example is not good. The truth is we don't know how much impact those two attributes have together, to say if it makes no difference, little difference, or a lot of difference.
The only way of making it clear is tedious work: You (or someone else) needs to run tests in groups of three attributes, raising them or lowering them together. This way, we'll gain insight into every 3 attribute combination, and by doing the average of every 3 attribute combination that have the same 2 attributes, we can also gain insight into which 2 attribute combination is good.
If you are correct, then the 3 attribute combination of 'Acceleration'+'Pace'+'Jumping Reach' should perform much better than any other combination, including 'Acceleration'+'Pace'+'Strength', or 'Passing'+'Decisions'+'Work Rate'.
add height , Consistency ,Non-dominant feet
Jolt said: But that is the flaw with these tests. If only the attribute 'Off the ball' or 'Flair' is increased, but the player doesn't have the physical or technical attributes to take advantage of the 'Off the ball' or 'Flair', of course it is going to look like a bad attribute. If these attributes are multiplicative of other attributes, and the other attributes are low, then of course these multiplicative attributes will look bad.
Your finishing & composure attributes example is not good. The truth is we don't know how much impact those two attributes have together, to say if it makes no difference, little difference, or a lot of difference.
The only way of making it clear is tedious work: You (or someone else) needs to run tests in groups of three attributes, raising them or lowering them together. This way, we'll gain insight into every 3 attribute combination, and by doing the average of every 3 attribute combination that have the same 2 attributes, we can also gain insight into which 2 attribute combination is good.
If you are correct, then the 3 attribute combination of 'Acceleration'+'Pace'+'Jumping Reach' should perform much better than any other combination, including 'Acceleration'+'Pace'+'Strength', or 'Passing'+'Decisions'+'Work Rate'.
In another test (tactics not this one),
1. I increase Passing, Crossing, Marking, Dribbling, Tackling, Finishing to 12,
Technique Keep 10 unchanged.
get A value of A
2. I increase Passing, Crossing, Marking, Dribbling, Tackling, Finishing to 12,
Then reduce Technique to 5.
get a value of B.
The result is B > A
As I understand it as Technique can use Passing, Crossing, Marking, Dribbling, Tackling, Finishing,
But reducing it Technique improve performance
Possebrew said: Awesome work! I'm lacking superlatives to thank you!
The extremes of 1 in some of these attributes are staggering.
At the same time, this is also good news!
It means you can't just take a player with 20Pac/Acc/Dri and dominate everything.
Well-rounded players are still desirable.
I wouldn't have expected that being single footed has such a big impact. This comes in stark contrast to some mainstream advice that it's not worth improving the weak-foot. It also means that it's worth comparing boosting off-foot VS traits like: 'Avoid weaker foot'
The gap between low determination, anticipation, concentration & dribbling is also surprising.
Decision being worthless left me flabbergasted!
I have a different take on this test result,
As I mentioned above,
"
Note that the difference between attributes is only used to show the degree of differentiation and cannot be used for direct quantified comparisons.(Because it's not standardized) . For example, "difference from the standard value +40" is not twice as useful as "difference from the standard value +20" .
"
This result is not "standardized."
For example (hypothetical)
a +40 attribute can be three times more important than a +20 attribute
And
a attribute of -40 is three times worse than an attribute of -20 attribute
that means ,
Ignore the Work rate,
Paceand Acceleration have an amazing 171 and 161,
which means that the importance of physical attributes is not limited to 61.47% in the picture,
the importance of physical attributes must be much greater than 61.47%
Yarema said: The difference between onefooted and both footed is really not that big and considering CA cost I'd argue it's better to invest that into some actual attributes. Could maybe use a test with value of say 6 and 10 or something.
Non-dominant feet = 1 2.1
Non-dominant feet = 6 11.7
Non-dominant feet = 10 12.7
Left Feet = Right Feet = 20 20.1
harvestgreen22 said: I have a different take on this test result,
As I mentioned above,
"
Note that the difference between attributes is only used to show the degree of differentiation and cannot be used for direct quantified comparisons.(Because it's not standardized) . For example, "difference from the standard value +40" is not twice as useful as "difference from the standard value +20" .
"
This result is not "standardized."
For example (hypothetical)
a +40 attribute can be three times more important than a +20 attribute
And
a attribute of -40 is three times worse than an attribute of -20 attribute
that means ,
Ignore the Work rate,
Paceand Acceleration have an amazing 171 and 161,
which means that the importance of physical attributes is not limited to 61.47% in the picture,
the importance of physical attributes must be much greater than 61.47%
"
Note that the difference between attributes is only used to show the degree of differentiation and cannot be used for direct quantified comparisons.(Because it's not standardized) . For example, "difference from the standard value +40" is not twice as useful as "difference from the standard value +20" .
"
Or to put it in a better way,
It's easier to score 2 goals in 1 match,
But it's not just 500% harder to score 10 goals than to score 2 goals , it's a lot harder than 500%
so ,with standard Control Group 20.1,
a stats with Goal difference (GD) 60.0 is much much more important than a stats with Goal difference (GD) 40.0
The closer a Stats is to 20.1, the less important it is
**** SI
@harvestgreen22 With this data, I wonder if the training schedules data can have Dribbling and Balance growth added with Pace, Acceleration and Jumping Reach.
Concentration, Anticipation, Determination，Agility can be used to separate ties in the transfer market but not a focus in training?
harvestgreen22 said: Non-dominant feet = 1 2.1
Non-dominant feet = 6 11.7
Non-dominant feet = 10 12.7
Left Feet = Right Feet = 20 20.1
Thank you for this test. As suspected a little bit of weak foot ability is good, but after that doesn't seem to be worth the cost. And for that reason I don't really like players who are good with both feet.
There might be an argument to be made to train weak foot of players who are terrible at it.
Jolt said: But that is the flaw with these tests. If only the attribute 'Off the ball' or 'Flair' is increased, but the player doesn't have the physical or technical attributes to take advantage of the 'Off the ball' or 'Flair', of course it is going to look like a bad attribute. If these attributes are multiplicative of other attributes, and the other attributes are low, then of course these multiplicative attributes will look bad.
Your finishing & composure attributes example is not good. The truth is we don't know how much impact those two attributes have together, to say if it makes no difference, little difference, or a lot of difference.
The only way of making it clear is tedious work: You (or someone else) needs to run tests in groups of three attributes, raising them or lowering them together. This way, we'll gain insight into every 3 attribute combination, and by doing the average of every 3 attribute combination that have the same 2 attributes, we can also gain insight into which 2 attribute combination is good.
If you are correct, then the 3 attribute combination of 'Acceleration'+'Pace'+'Jumping Reach' should perform much better than any other combination, including 'Acceleration'+'Pace'+'Strength', or 'Passing'+'Decisions'+'Work Rate'.
I'm trying more combinations, and maybe some of them are interacting with Technique and Flair
Like
Passing20，Finishing20，Dribbling20，Technique20，Flair20，other 10 : 70.7
Passing20，Finishing20，Dribbling20，Technique20，Flair10，other 10 : 80.4
Passing20，Finishing20，Dribbling20，Technique10，Flair20，other 10 : testing
Passing20，Technique20，Flair10，other 10 :testing
Finishing20，Technique20，Flair10，other 10 :testing
Dribbling20，Technique20，Flair10，other 10 :testing
Passing20，Technique10，Flair20，other 10 :testing
Finishing20，Technique10，Flair20，other 10 :testing
Dribbling20，Technique10，Flair20，other 10 :testing
Han106 said: @harvestgreen22 With this data, I wonder if the training schedules data can have Dribbling and Balance growth added with Pace, Acceleration and Jumping Reach.
Concentration, Anticipation, Determination，Agility can be used to separate ties in the transfer market but not a focus in training?
Yes.
Seeing if these secondary important attributes can also grow should also be a measure of the quality of the training
Recently, I didn't have the energy to re-test the training program to see how these attributes grew as a percentage of total growth , as I am busy testing various attributes
Jolt said: But that is the flaw with these tests. If only the attribute 'Off the ball' or 'Flair' is increased, but the player doesn't have the physical or technical attributes to take advantage of the 'Off the ball' or 'Flair', of course it is going to look like a bad attribute. If these attributes are multiplicative of other attributes, and the other attributes are low, then of course these multiplicative attributes will look bad.
Your finishing & composure attributes example is not good. The truth is we don't know how much impact those two attributes have together, to say if it makes no difference, little difference, or a lot of difference.
The only way of making it clear is tedious work: You (or someone else) needs to run tests in groups of three attributes, raising them or lowering them together. This way, we'll gain insight into every 3 attribute combination, and by doing the average of every 3 attribute combination that have the same 2 attributes, we can also gain insight into which 2 attribute combination is good.
If you are correct, then the 3 attribute combination of 'Acceleration'+'Pace'+'Jumping Reach' should perform much better than any other combination, including 'Acceleration'+'Pace'+'Strength', or 'Passing'+'Decisions'+'Work Rate'.
More testing of the effects of multivariables，
1.
Passing传球15
Crossing传中15
Dribbling盘带15
Tackling抢断15
Finishing射门15
First touch停球15
Heading头球15
longshot远射15
Technique技术10
Flair才华10
other其他属性10
Goal difference 净胜球≈73.7
2.
Passing传球15
Crossing传中15
Dribbling盘带15
Tackling抢断15
Finishing射门15
First touch停球15
Heading头球15
longshot远射15
Technique技术15
Flair才华10
other其他属性10
Goal difference 净胜球≈60.3
3.
Passing传球15
Crossing传中15
Dribbling盘带15
Tackling抢断15
Finishing射门15
First touch停球15
Heading头球15
longshot远射15
Technique技术10
Flair才华15
other其他属性10
Goal difference 净胜球≈71.7
4.
Dribbling盘带20
Technique技术20
other其他属性10
Goal difference 净胜球≈44.2
Dribbling盘带20
Technique技术10
other其他属性10
Goal difference 净胜球=53.6
5.
Passing传球20
Technique技术20
other其他属性10
Goal difference 净胜球≈27.9
Passing传球20
Technique技术10
other其他属性10
Goal difference 净胜球=24.5
6.
Finishing传球20
Technique技术20
other其他属性10
Goal difference 净胜球≈33.7
Finishing传球20
Technique技术10
other其他属性10
Goal difference 净胜球=33.3
Summary:
7.
Under Control variable,
From 1, 2, and 3, you can see,
Even in the multivariable case, considering multiple technical type of attributes,
technical type of attributes 15 + Technique 10 (73.7) > technical type of attributes 15+ Technique 15 (60.3) ，This means that in this case, Technique is bad
technical type of attributes 15 + Flair 10 (73.7) ≈/> technical type of attributes 15 + Flair 10 (71.7) ，This means that in this case, Flair have little/no influence
8.
In the case of 4, 5, and 6, it's a little special,
Dribbling 20 + Technique 20 (44.2) < Dribbling 20 + Technique 10 (53.6)
Higher Technique is having a negative effect with Dribbling
This means that in this case, Technique is bad
Passing 20 + Technique 20 (27.9) ≈/> Passing 20 + Technique 10 (24.5)
Higher Technique may or may not have a positive effect with Passing
Not much difference.
It is also possible that the sample size is not large enough and resulting in random error
Finishing 20 + Technique 20 (33.7) ≈ Finishing 20 + Technique 10 (33.3)
Higher Technique had no effect on Finishing passing
9.
What is surprising is number 5,
which means that it is possible for Technique to have some positive Stats and negative effects on others
here you go, a real meta player.
https://fminside.net/players/5-fm-243/2000052135-abdallah-sima (note: dvide the numbers by 5 for the actual game value.)
harvestgreen22 said: I have a different take on this test result,
As I mentioned above,
"
Note that the difference between attributes is only used to show the degree of differentiation and cannot be used for direct quantified comparisons.(Because it's not standardized) . For example, "difference from the standard value +40" is not twice as useful as "difference from the standard value +20" .
"
This result is not "standardized."
For example (hypothetical)
a +40 attribute can be three times more important than a +20 attribute
And
a attribute of -40 is three times worse than an attribute of -20 attribute
that means ,
Ignore the Work rate,
Paceand Acceleration have an amazing 171 and 161,
which means that the importance of physical attributes is not limited to 61.47% in the picture,
the importance of physical attributes must be much greater than 61.47%
I feel the focus might be too much on identifying the "best attributes" for enhancing performance. However, the significant negative impact of attributes like Work Rate at 1 underscores the importance of avoiding negative attributes just as much as maximizing positive ones.
Consider this hypothetical scenario: If everyone on the team has a Work Rate below 5, attributes like 20 in Acceleration and Pace could become irrelevant. It's like having a team of Usain Bolts who prefer to chat rather than run because they find the effort tedious. This principle likely applies to other attributes as well.
Here's how I think about it:
- Any attribute that pulls your team's performance below the control group's 20.1 goal difference is critical to consider.
The sequence of attribute impact might work like this (and it's how I would design such a system):
1. Decision Check: First, the game checks if the player decides to act or not.
2. Off-the-Ball Intelligence: Then, it determines if the player knows where to move.
3. Pace and Acceleration: These attributes determine how quickly they can get to that position.
4. First Touch: Checks if the player can control the ball when they receive it.
5. Dribbling: Decides how effective the player is with the ball in terms of movement.
6. Agility and Balance: These determine how well the player can avoid being tackled or recover if they are.
7. Vision and Decision: Another check to see if the player chooses to shoot or pass.
8. Type of Action: Decides the specifics of the pass or shot.
9. Shot Outcome: If shooting, this is measured against the goalkeeper's attributes to see if a goal is scored.
Each of these steps would be influenced by the relative attributes of the players involved, not just the absolute values. For instance, if all players had 20 in Acceleration and Pace, the value of these attributes would diminish because it's the differential between players that matters, not the absolute attribute itself.
It's also crucial to distinguish whether an attribute is more relevant for the interaction between players (like a striker versus a defender) or for the individual's own performance metrics (like shooting accuracy).
harvestgreen22 said: * This is translated by a language translator and then manually corrected, there may be some loss of meaning after translation
Test setup: The league has 4 teams, A, C, D team is set to 4231 Meta tactics (https://fm-arena.com/thread/12667-god-of-chaos-v1/), B team is set to 4141 weak tactics, all attributes of all players are set to 10, locking various attributes.
Modify the player attributes of Team A (except for the goalkeeper, that is, adjust 10 players) to the corresponding value.
Statistical the goal difference.
The C and D teams use a very strong 4231 tactics, which tests the players' defensive ability and performance under pressure from strong teams.
The B team uses the weak 4141 tactics to test the players' ability to Beat up weak teams and score lots of goals.
It took 10 days to test, two computers together. 2 of the attributes were tested by the help of the community netizen , and the other 70 were tested by me.
Inspired by this https://fm-arena.com/table/26-player-attributes-testing/ (November 2023 version)I thought, What if the attribute is 1 ?
1.How to read this table?
The standard(Control) group represents the goal difference of Team A when all attributes are 10, which is 20.1.
## Look at Technique 20, which represents Team A (apart from the goalkeeper) skill attribute increases from 10 to 20, and Statistical the goal difference of Team A, it's 16.5.
16.5 < 20.1 (standard)
In other words, the Technique has increased by 10, but the goal difference has decreased. But not by much.
## Look at Vision 1 and Vision 20, 20.0 and 38.5 respectively.
20.0 < 20.1 < 38.5
In other words, when Vision is reduced from 10 to 1, so there is no change.
However, when Vision was increased from 10 to 20, resulting in a goal difference increase of (38.5-20.1)=18.4
2. Noteworthy attributes
Lowering the Work rate 1 ,
it have very serious consequences (goal difference -110),
and it should be ensured that a player has at least 6 and preferably 10 Work rate.
A player with a Work rate is not desirable.
But continuing to increase Work rate (from 10->20) is low effect. The difference between Work rate 10 and Work rate 20 is relative small.
The 6-point attribute seems to be the threshold for some attributes.
Passing, Crossing, Marking, First touch, Positioning, Decision, Teamwork, Off the ball, Bravery, are all attributes that can be considered useless.
Technique and Flair, Higher values lead to worse performance.
Spoiler Yes, it's a scam.
To Flair. There are two guesses,
one is that "fancy tricks" cause players to spend more "frames" on the movement, resulting in a decrease in efficiency.
The other is that more "fancy tricks" contribute not enough to the mathematical expectation when attacking, while the mathematical expectation of being scored when defending increases more than attacking scored
Core attributes are these:
Special : Work rate (need to reach 10, higher is useless),
Pace, Acceleration, Jumping reach, Dribbling, Balance, Concentration, Anticipation, Determination，Agility.
According to the comparison of attribute 20 with attribute 10:
Pace + Acceleration 2 most important cores,
Jumping reach + Dribbling 2 secondary cores,
Balance can be considered a level 3 attribute,
level 4 attribute Anticipation and Concentration
Other attributes, which are although have effect, but much less effective.
Note that the difference between attributes is only used to show the degree of differentiation and cannot be used for direct quantified comparisons.(Because it's not standardized) . For example, "difference from the standard value +40" is not twice as useful as "difference from the standard value +20" .
3.Sample
According to the Law of Large Numbers in statistics, as long as the sample is large enough, the value you get will be closer and closer to the mathematical expectation
In terms of important attributes, there are about 950-1080 Match samples. According to statistical estimates, the margin error of attribute is ≤ 3.
In the non-important attributes, there are about 450 Match samples. According to statistical estimates, the margin error of attribute is ≤ 8
4.margin error
Is that a conclusion that only applies to 4231 and certain opponents?
You know, tactics can vary a lot.
Moreover, 4231 players are not evenly distributed, for example, there is only 1 striker.
And tactics, duties, and commands all affect the importance of attributes.
There is also the test of the Go-on-holiday, even if the game is set to maximum detail, will it get unreal values?
——Go-on-holiday have been studied, and I won't repeat it in a lot of words, but I choose to trust their findings
——Taking into account different tactics and different opponents, an additional 3,600 match were played,
To test two other tactics, one is Default tiki-taka 433. And Default counter-attack 442.
They will have different standard values and different attributes deviating from the standard, but "importance level" of each attribute not change.
For example, in the tactics where Passing is more important,
Passing (around 24 in the table) is still not important than dribbling (more than 50 in the table) or Finishing (33 in the table) in any tactic.
This also shows that the game is not complicated to do a separate set of calculations for each tactic, but a relatively universal calculation method.
And this method of calculation, my personal observation, is more conducive to offensive.
——Similarly, the problem of player distribution does not have a "decisive" effect, but only a statistical perturbation of some value.
5.Multiple variables
some comments said that the performance of attributes is A multi-variable problem, for example, the A attribute is evaluated at the same time as the B attribute.
Some people will think that testing a player who First touch 10 and Finishing 15, because the player can't stop the ball and it doesn't matter how high the Finishing is, so this test is wrong.
However, if an attribute is weighted with one or more other attributes, all associated attributes must be found. In the end all testing is either pointless or too tedious to do.
Since it is statistically possible to tell the difference, let's use this result as a conclusion
Now, imagine a mathematical problem. Player A's Composure is 10, provides 20% of the goal scoring ability. Finishing 10, providing 10 accuracy shots on goal per game. The end result is 10x20%=2 balls.
If this is a multivariate problem, then Composure 20, assuming it provides 50% of the goal scoring ability. The end result is 10x50%=5 balls. So 2 becomes 5. The difference "5-2=3" may or may not be obvious.
Those that are not obvious are actually the effect of interaction is not good, since it is so small, it can simply be ignored.
If there is an effect, it must be reflected in the statistical results.
Unless it is using some more complex mechanism, for example, it detects composure 20, and if the shot is not ≥15, the goal scoring ability % does not increase. I'm not going to consider that possibility.
6.don't bring the logic of reality into the game
For example, there is such a reply: there is no "speed 20" and other attributes are 10 in reality .
Or, "Should use balabala" attributes to design all forwards/centers/back guards to reflect the real situation.
The results, as long as the control variable method is followed, make no difference (since the game seems unlikely to use the "more complex" mechanics mentioned above).
7. Why not test each player individually, such as the striker
Because it's so much work,
And the impact each player has on the team's goal difference is likely to be the difference between the actual measured value and the standard value, which is always disturbed by larger random perturbations, because a tactic is mostly 1-3 strikers, who make up less than 30% of the 11 men.
If you want to measure the true value, you need a very, very large number of samples
8. The goalkeeper?
Similar to the above reasons. mishap
If you want to measure the true value, you need a very, very large number of samples
9. Why testing just one season The result is completely inaccurate? (from reply)
The random perturbation in the game is very large, which means that if you only play 30 to 40 games (a season), the result is likely to deviate significantly from the true value
One might also suspect that the test is not rigorous enough,
it's not scientific enough,
it's not an actual season,
it's a nice ideal environment,
and so on,
and then argue that the empirical results under a lot of interference are more accurate than the results obtained under the control variable method
10.thoughts
Like the previous "1CA Win Champion", many people will dismiss this as a conspiracy theory before testing reveals the game mechanics. After the revelation, some players believe that this kind of testing is just excessive pursuit of intensity at the expense of game life. But that's not the purpose of the test.
My goal in testing this is the same as the goal of many gamers: to force the game company to continuously improve the old mechanics it has used for years, to change the fundamental engine flaws, and to prevent players from playing a "placebo simulator". Why a placebo simulator? There are people who take real football and put a lot of time and energy into the game, there are a lot of players who put a lot of energy into the video, I also enjoy these passionate things
But,
As a result, the game engine does not do these functions at all, these players believe and carefully prepared, essentially reflected in the game engine is useless. The game relies on some very simple data and then randomizes it to make players think they are playing a real simulation. Then the game company can just change the appearance of their game, sell it for generations, and advertise that they are making a simulation full of interactivity
What's more, the authorities have banned discussions on official forums and continue to mislead consumers. You can only find information in unofficial forums.
Like the game training system I tested, I only played the FM2024 generation , and now only a 300 hours player, and many previous generations, plus the latest generation, have a lot of players who spend a lot of time researching and sharing their training findings.
As a result, these players actually played a placebo simulator
There are even some emotionally invested players who bury their heads in the sand and refuse to believe (forgive me for being overly aggressive) the results of these tests that reveal Physical man is Superman. These players who have invested money and effort should enjoy the same experience as the simulation theory, not the illusion of a fake mechanic.
It's not the player's fault, it's the game company's fault. players love football simulation games, doesn't mean you should accept a flawed game, especially when the game's biggest selling point and publicity point is simulation. You can choose not to play. Or you can stay and protest together in one way or another and force the game company to change. You can also play the game as much as you want, without deliberately using these "intensity mechanics" and have fun.
My guess for why high Technical and Decisions is detrimental is because the game does not know what actions are the most effective. If you look at technical attributes, you will see that dribbling is more efficient than passing or shooting, possibly because pace and acceleration are over tuned in the engine, and dribbling makes use of both attributes. So, characters with high technical will have increased chance of success for passes and shots, while characters with high decisions will find better opportunities to pass and shoot. Unfortunately, the most effective is to pass and shoot as little as possible, unless you have those attributes much higher than dribbling attributes (including pace and acceleration).
So, in short, if I am not wrong:
Pass 20 + Technique 20 + Other 10 > Pass 20 + Technique 10 + Other 10.
Pass 20 + Technique 10 + Dribbling 20 + Pace 20 + Acceleration 20 + Other 10 > Pass 20 + Technique 20 + Dribbling 20 + Pace 20 + Acceleration 20 + Other 10.
The same is probably true for long shots or finishing.
ZaZ said: My guess for why high Technical and Decisions is detrimental is because the game does not know what actions are the most effective. If you look at technical attributes, you will see that dribbling is more efficient than passing or shooting, possibly because pace and acceleration are over tuned in the engine, and dribbling makes use of both attributes. So, characters with high technical will have increased chance of success for passes and shots, while characters with high decisions will find better opportunities to pass and shoot. Unfortunately, the most effective is to pass and shoot as little as possible, unless you have those attributes much higher than dribbling attributes (including pace and acceleration).
So, in short, if I am not wrong:
Pass 20 + Technique 20 + Other 10 > Pass 20 + Technique 10 + Other 10.
Pass 20 + Technique 10 + Dribbling 20 + Pace 20 + Acceleration 20 + Other 10 > Pass 20 + Technique 20 + Dribbling 20 + Pace 20 + Acceleration 20 + Other 10.
The same is probably true for long shots or finishing.
1.
Passing传球15
Crossing传中15
Dribbling盘带15
Tackling抢断15
Finishing射门15
First touch停球15
Heading头球15
longshot远射15
Technique技术10
Flair才华10
other其他属性10
Goal difference 净胜球≈73.7
2.
Passing传球15
Crossing传中15
Dribbling盘带15
Tackling抢断15
Finishing射门15
First touch停球15
Heading头球15
longshot远射15
Technique技术15
Flair才华10
other其他属性10
Goal difference 净胜球≈60.3
3.
Passing传球15
Crossing传中15
Dribbling盘带15
Tackling抢断15
Finishing射门15
First touch停球15
Heading头球15
longshot远射15
Technique技术10
Flair才华15
other其他属性10
Goal difference 净胜球≈71.7
4.
Passing传球15
Crossing传中15
Dribbling盘带15
Tackling抢断15
Finishing射门15
First touch停球15
Heading头球15
longshot远射15
Technique技术15
Flair才华15
other其他属性10
Goal difference 净胜球≈58.9
I added a situation (situation 4.), They're both at 15, Goal difference the lowest.
don't know how to explain that mechanism
I guess it's not as complicated as we might think: It's a bug,Or the mechanics aren't right
Hi @harvestgreen22!
Thank you so much for all the great work! It is great that you have bothered to do all these tests so we can know more about which attributes are more useful in the game! This is all greatly appreciated!
However, I do have one question which, depending on your answer, may explain why you are finding that some of the attributes have a negative impact on performance (e.g., technique):
- How are you dealing with the current ability (CA) of players not being the same after you raise (or lower) one (multiple) of the attributes?
-> When you set all players' attributes to 10, their CA needs to be set to a certain level such that the attributes remain stable... when you change one of the attributes to 20 (100 in the "centi-attributes" that the game actually uses and that you can see in the save game editor), you need a higher CA amount for the attributes to remain stable (and this higher CA amount will depend on the player's position given that, for example, finishing has a larger CA weight for forwards than for defenders). Otherwise, the game will lower attributes across the board so that the player's CA remains the same...
-> This means that if you are running a league for your tests that takes any amount of game time and you don't either (1) make sure that the player's CA is what is needed to maintain the attributes you set initially or (2) have some mechanism to ensure that the player's attributes remain frozen (it is my understanding that some save game editors may be able to do this), then the game will be changing other attributes to compensate for the attribute that you maxed out (or set to 1).
-> This would mean that when, for example, you set technique to 20, other attributes are lowered below 10 to compensate (you may not always be able to see this when attributes are represented 1-20 in-game, but this may still be happening under the hood in the attributes represented 1-100 which you can see in savegame editors).
-> If you are not accounting for this attribute re-balancing, then your finding that technique has a negative impact on goal difference in your test may not mean that technique literally has a negative impact on player performance, but rather that setting technique to 20 is not worth the reduction in other attributes that it causes...
Anyway, I would be interested in knowing how you dealt with attribute re-balancing due to the player's CA in your tests.
-------------------
In addition, I'd like to suggest that you also try testing "aggression" and "leadership." Both may have some influence in performance. Aggression may interact with certain attributes, such as tackling or bravery.
Also could be interesting if you tested the important matches attribute (having in mind that the influence of this attribute depends on how important the match is - see https://www.youtube.com/watch?v=ECe0ygapG-8&t=1404s)
mmigueis said: Hi @harvestgreen22!
Thank you so much for all the great work! It is great that you have bothered to do all these tests so we can know more about which attributes are more useful in the game! This is all greatly appreciated!
However, I do have one question which, depending on your answer, may explain why you are finding that some of the attributes have a negative impact on performance (e.g., technique):
- How are you dealing with the current ability (CA) of players not being the same after you raise (or lower) one (multiple) of the attributes?
-> When you set all players' attributes to 10, their CA needs to be set to a certain level such that the attributes remain stable... when you change one of the attributes to 20 (100 in the "centi-attributes" that the game actually uses and that you can see in the save game editor), you need a higher CA amount for the attributes to remain stable (and this higher CA amount will depend on the player's position given that, for example, finishing has a larger CA weight for forwards than for defenders). Otherwise, the game will lower attributes across the board so that the player's CA remains the same...
-> This means that if you are running a league for your tests that takes any amount of game time and you don't either (1) make sure that the player's CA is what is needed to maintain the attributes you set initially or (2) have some mechanism to ensure that the player's attributes remain frozen (it is my understanding that some save game editors may be able to do this), then the game will be changing other attributes to compensate for the attribute that you maxed out (or set to 1).
-> This would mean that when, for example, you set technique to 20, other attributes are lowered below 10 to compensate (you may not always be able to see this when attributes are represented 1-20 in-game, but this may still be happening under the hood in the attributes represented 1-100 which you can see in savegame editors).
-> If you are not accounting for this attribute re-balancing, then your finding that technique has a negative impact on goal difference in your test may not mean that technique literally has a negative impact on player performance, but rather that setting technique to 20 is not worth the reduction in other attributes that it causes...
Anyway, I would be interested in knowing how you dealt with attribute re-balancing due to the player's CA in your tests.
-------------------
In addition, I'd like to suggest that you also try testing "aggression" and "leadership." Both may have some influence in performance. Aggression may interact with certain attributes, such as tackling or bravery.
Also could be interesting if you tested the important matches attribute (having in mind that the influence of this attribute depends on how important the match is - see https://www.youtube.com/watch?v=ECe0ygapG-8&t=1404s)
I manually changed the "Current CA" to the "Recommended CA" in every test every time
(I don't know if the translator misunderstands your meaning.)
like this:
I used the in-game modifier.
Did that make a difference to the results? Or do I need other non-in-game tools?
Thank you for confirming!
This removes my objection... I am still having trouble believing in some of the results (just like the results in the other FM Arena testing - https://fm-arena.com/table/26-player-attributes-testing/), but I don't have an alternative theory that can explain them at the moment.
At this point it would be more interesting to test position rating systems to see if it can consistently produce winners against a + 2 or + 4 acc/pac team with the secondary meta attributes + stamina/wr set to 10 & the rest of the worthless attributes set to 1. ie A team with of 14 acc/pac with anything goes vs 16/16 acc/pac or 18 acc/pac team.
Like often times you are picking between a 15/15 AMRL that has 16 dribbling or 15 jumping reach, and a 17/17 player that has 3 jumping reach & 10 dribbling.
Maybe flair is like a randomness attribute
bigloser said: At this point it would be more interesting to test position rating systems to see if it can consistently produce winners against a + 2 or + 4 acc/pac team with the secondary meta attributes + stamina/wr set to 10 & the rest of the worthless attributes set to 1. ie A team with of 14 acc/pac with anything goes vs 16/16 acc/pac or 18 acc/pac team.
Like often times you are picking between a 15/15 AMRL that has 16 dribbling or 15 jumping reach, and a 17/17 player that has 3 jumping reach & 10 dribbling.
from experience, 17/17 player is still better. How often are there a 15/15 with 15JR + 15 DRB? its usually like a super Dribbly 15/15/15 or a 15/15/8 with 16 JR.
Not that it matters that much because hidden attributes is just as important, so a good looking player might just be a dude if pressure, injury proness or dirtiness is bad.
Question: are these attributes frozen? I don't use the ingame editor for my tests, so I'm not familiar with it.