Did some analysis on ykykyk weights vs. the FM Arena attribute testing. Sorry if this has been discussed before and I missed it...
Basically, I wanted to compare the two, so I aggregated the value of each attribute in YKYKYK (with and without GK position). Then created an average for the attribute across all positions. You'll see that in the first two columns, sorted from highest to lowest. P/A are top, leadership on the bottom.
Then took the attribute testing from this site, and ranked that from most impact to least (and made a little %). Here's the results.
This suggests that Jumping Reach, Concentration and Strength, are key “defensive position heavy weight” is a key defensive performance attribute.
Balance and Finishing are big “offensive position heavy weight” and key offensive performance attributes.
So maybe we’re looking at 3 Tiers of interesting attributes -
Tier 1 - ykykyk down to, say, Aggression (or Decisions, if you’d like) PLUS for offense Balance and Finishing, and then for defense, Jumping Reach, Concentration, Strength.
I am following it just a few more details would be great. How did you simulate, how did you determine which attributes are important for defenders and attackers.
Ps I will look at your work on the other thread. I'm really interested in both and interested in what are key attributes for keepers.
I've always been a bit sceptical about ykykyky's ratings because their findings were only based on that single ZaZ tactic. The attribute testing, being done by this website, is obviously also based on a single tactic though.
We've also seen on other tests that attributes, and subsequently good players, do have a bigger impact on a match compared to any tactic though. So in theory these findings should be somewhat universal, regardless of tactic.
There are also some attributes that I believe have low-to-zero impact on the result of any game, as they're only used to determine player's pro- or regression throughout their careers such as Determination and Natural Fitness.
In my personal experience, I've used both ykykyky and Mark's ratings file for Genie Scout. I tend to lean towards Mark's ratings because they seem to consistently identify high-performing players for my team. It's all about finding what works best for your style of play, but it's great to have multiple resources like these at our disposal.
skawkclsrn said: I've always been a bit sceptical about ykykyky's ratings because their findings were only based on that single ZaZ tactic. The attribute testing, being done by this website, is obviously also based on a single tactic though.
We've also seen on other tests that attributes, and subsequently good players, do have a bigger impact on a match compared to any tactic though. So in theory these findings should be somewhat universal, regardless of tactic.
There are also some attributes that I believe have low-to-zero impact on the result of any game, as they're only used to determine player's pro- or regression throughout their careers such as Determination and Natural Fitness.
In my personal experience, I've used both ykykyky and Mark's ratings file for Genie Scout. I tend to lean towards Mark's ratings because they seem to consistently identify high-performing players for my team. It's all about finding what works best for your style of play, but it's great to have multiple resources like these at our disposal. Expand
The Machine Learning weights never felt right to me. I always felt they underrated physicals. On average they put less weight on Acceleration, Agility, Strength, and Jumping to just the default CA weights. By comparison they put a lot more emphases on Aggression, Bravery, Work Rate, and Dribbling.
On average they put about 15% of the total score to pace and acceleration. I'm not sure what the correct percentage would be, but my guess is it might be 30-40 depending on the position.
jackofarcades said: The Machine Learning weights never felt right to me. I always felt they underrated physicals. On average they put less weight on Acceleration, Agility, Strength, and Jumping to just the default CA weights. By comparison they put a lot more emphases on Aggression, Bravery, Work Rate, and Dribbling.
On average they put about 15% of the total score to pace and acceleration. I'm not sure what the correct percentage would be, but my guess is it might be 30-40 depending on the position. Expand
also, I think the ratings might have to be modified for this FM because of how much more important jumping is all of a sudden. Actually, jumping might have become more important since FM 23 midway, after attribute testings were already done. I noticed that players with high Target man rating in ykykyk always performed better, even when using them as AF, because they would just get into positions and score more headers. I had Simmeone Giovanni score 56 goals in Siere A alone with a pretty mid Roma side becuase he has very good pace, acceleration and jumping. Also, mentality seem even weaker this FM based on the attribute testing, especially with the nerf to anticipation.
Did some analysis on ykykyk weights vs. the FM Arena attribute testing. Sorry if this has been discussed before and I missed it...
Basically, I wanted to compare the two, so I aggregated the value of each attribute in YKYKYK (with and without GK position). Then created an average for the attribute across all positions. You'll see that in the first two columns, sorted from highest to lowest. P/A are top, leadership on the bottom.
Then took the attribute testing from this site, and ranked that from most impact to least (and made a little %). Here's the results.
Also, looking at all this a second time, interesting attribute differences between weight and impact on GD in simulations come from:
Jumping Reach
Balance
Concentration
Finishing
Strength
This suggests that Jumping Reach, Concentration and Strength, are key “defensive position heavy weight” is a key defensive performance attribute.
Balance and Finishing are big “offensive position heavy weight” and key offensive performance attributes.
So maybe we’re looking at 3 Tiers of interesting attributes -
Tier 1 - ykykyk down to, say, Aggression (or Decisions, if you’d like) PLUS for offense Balance and Finishing, and then for defense, Jumping Reach, Concentration, Strength.
Tier 2 - ykykyk everything else between, say, Teamwork and Tackling
Tier 3 - everything else.
Then all you’d have to do is figure out how much a 17 in, say, Determination or Flair is worth, compared to a 12 in Pace, vs. a 14 in 1st Touch…
Now explain it to me like I'm five years old
I am following it just a few more details would be great. How did you simulate, how did you determine which attributes are important for defenders and attackers.
Ps I will look at your work on the other thread. I'm really interested in both and interested in what are key attributes for keepers.
I'm thinking through this again. Ignore this.
I have spent the last 2 hours looking for this 'ykykyk weights' and can't find them anywhere, any pointers please?
So what attributes are the most important for each role in your opinion?
Y15ARG said: I have spent the last 2 hours looking for this 'ykykyk weights' and can't find them anywhere, any pointers please?
ykykyky balanced ratings file link
I've always been a bit sceptical about ykykyky's ratings because their findings were only based on that single ZaZ tactic.
The attribute testing, being done by this website, is obviously also based on a single tactic though.
We've also seen on other tests that attributes, and subsequently good players, do have a bigger impact on a match compared to any tactic though. So in theory these findings should be somewhat universal, regardless of tactic.
There are also some attributes that I believe have low-to-zero impact on the result of any game, as they're only used to determine player's pro- or regression throughout their careers such as Determination and Natural Fitness.
In my personal experience, I've used both ykykyky and Mark's ratings file for Genie Scout. I tend to lean towards Mark's ratings because they seem to consistently identify high-performing players for my team. It's all about finding what works best for your style of play, but it's great to have multiple resources like these at our disposal.
skawkclsrn said: I've always been a bit sceptical about ykykyky's ratings because their findings were only based on that single ZaZ tactic.
The attribute testing, being done by this website, is obviously also based on a single tactic though.
We've also seen on other tests that attributes, and subsequently good players, do have a bigger impact on a match compared to any tactic though. So in theory these findings should be somewhat universal, regardless of tactic.
There are also some attributes that I believe have low-to-zero impact on the result of any game, as they're only used to determine player's pro- or regression throughout their careers such as Determination and Natural Fitness.
In my personal experience, I've used both ykykyky and Mark's ratings file for Genie Scout. I tend to lean towards Mark's ratings because they seem to consistently identify high-performing players for my team. It's all about finding what works best for your style of play, but it's great to have multiple resources like these at our disposal.
Is the "mark ratings" not ykykyk balanced?
Floppyaams said: Is the "mark ratings" not ykykyk balanced?
My ratings came out before the machine learning ratings and are slightly different.
The Machine Learning weights never felt right to me. I always felt they underrated physicals. On average they put less weight on Acceleration, Agility, Strength, and Jumping to just the default CA weights. By comparison they put a lot more emphases on Aggression, Bravery, Work Rate, and Dribbling.
On average they put about 15% of the total score to pace and acceleration. I'm not sure what the correct percentage would be, but my guess is it might be 30-40 depending on the position.
jackofarcades said: The Machine Learning weights never felt right to me. I always felt they underrated physicals. On average they put less weight on Acceleration, Agility, Strength, and Jumping to just the default CA weights. By comparison they put a lot more emphases on Aggression, Bravery, Work Rate, and Dribbling.
On average they put about 15% of the total score to pace and acceleration. I'm not sure what the correct percentage would be, but my guess is it might be 30-40 depending on the position.
also, I think the ratings might have to be modified for this FM because of how much more important jumping is all of a sudden. Actually, jumping might have become more important since FM 23 midway, after attribute testings were already done. I noticed that players with high Target man rating in ykykyk always performed better, even when using them as AF, because they would just get into positions and score more headers. I had Simmeone Giovanni score 56 goals in Siere A alone with a pretty mid Roma side becuase he has very good pace, acceleration and jumping. Also, mentality seem even weaker this FM based on the attribute testing, especially with the nerf to anticipation.