"We see that work rate reduces in-match condition % identically to stamina. So stamina cannot be more important than work rate, or at least not more than ~20% more important say."
And you have the weights at 18 and 12 respectively? This is why I always ask questions because you come out with statements like this where it's one way for a few days then it completely flips, but you are 100% certain each time. You had one set of weights one day, I found them strange and enquire about the reasoning behind them, then you completely switch the system with massive changes and somehow it's my fault for asking. The "truth" shouldn't switch around so much, sure there is some inaccuracy and with more data we can hone in better.
I generally like that you are pursuing these questions, but the attitude "I am right and everyone else is wrong and stupid" is kind of offputting, especially when you do 180 as often as you do. Like do the tests and lets discuss if it makes sense, what are the limitations, can you approach it differently etc. You've found some interesting stuff that is unfortunately the victim of the level of discussion in here.
I don't think it was stated that a GK is only 20% of an outfield player.
Why you keep using league position as the end point still baffles me. Yes it has less variance because there are only 20 possible spots compared to lets say 100 possible points. You have to go really out of your way not to be in a +-3 places range. Additionally there is a lot more volatility in the middle of the table than at very top or bottom.
There was decent data for GKs from FM Arena, Harvestgreen, Orion and probably others and they mostly match with what you found so it's not exactly groundbreaking. Still valuable additional data from a different approach.
Stellur said: @ZaZ Just some quick questions: is there a way to change the training intensity of the youth team? I have been trying to find it but cannot seem to find it anywhere, and another question is: what is the quickest way to get to the calendar for the youth team to give them the schedule Expand Click "Squad", then look in training tile - there are 2-4 dots in the top right corner. Each of those dots represents a team: senior, reserves, u21, u18 ... Click on the dot that you want to manage and then open the tile. From there you can edit whatever you want just like for the main team including intensity.
tam1236 said: Sure, with better concentration, dribbling and finishing they would be much better. But until there is no other team with good, physical players they are getting along. And, maybe that's not my coaches, but default FM schedules that give not enough quickness who sucks? CA points are really badly distributed: leadership (0 importance, unless...) is so expenxive - why? cooperation (0 importance) expensive - why? And physicals, most important, are just cheap - I can have decent player of 45 CA. When I see a CB 24 yo., 160 CA, pa, acc 10-11 I think something is really wrong in this game. It reminds me, that some time ago a club from my town won a domestic championship with a technically very weak but big, strong and quick team (and doing mainly physical training). So it happens irl, but of course not on a very high level Expand And that was my whole point. That other attributes aren't completely useless and shouldn't be valued at 0.
A 24 yo CB with 160 CA and having 11 acc is completely normal. We got so used to abusing the broken training system that we somehow forgot that. What is actually wrong in this game is that you can engineer a 18+ pace player out of "almost anyone".
tam1236 said: ... but I was really very, very curious, so I transferred in an editor the team from a table above to Arsenal (of course get rid off regular Arsenal players). And simulated whole season. They finished at 3rd place, after Chelsea and Liverpool (83-81-79 pts) and over MU (71). /Lost all cups/. Average CA of basic 11 was 103. Second part was transferring same FC Vaduz but 2 seasons later - stronger team, some weaker players sold, some came from an academy - average CA 119 (I blocked attributes) - and it's not "engineered distribution" - all from coaching and playing. They won PL over Chelsea and MC (88-83-80 pts) /won FA, lost Europa League/. Expand Now add 40 CA to them in nonessential attributes and sim again. If possible a fair assessment, not all CA into the absolute worst attributes you can think of.
Also the not engineered distribution is sort of true but not really. Using the meta schedules to boost physicals to levels that were obviously not intended is kind of engineering them, although they do happen "naturally" as opposed to straight up editing them.
GeorgeFloydOverdosed said: Here we see that speed, work rate, among others, have non-linear effects. Expand Sorry what? Other than work rate, the other attributes are pretty much linear.
Obviously we aren't looking for the 6th decimal, there is margin of error. And they might even behave slightly nonlinearly but a linear function can be a very good approximation to reduce complexity.
skawkclsrn said: TL;DR: I randomised all 60 attributes of every player in a 20-club league every 75 seconds and let the game holiday through ~16,000 full-engine matches. Pace and Acceleration are worth ~3x anything else. Everything technical measured at or below zero. According to these resutls no attribute has a threshold, they pay linearly 1-20. Once you divide by what a point costs in CA, the best buys are Work Rate, Jumping Reach and the hidden personality attributes, which are (mostly) free.
Method FM's engine is C++ with MSVC RTTI, so every object carries its class name behind the vtable pointer, so you can find every player and scoreline by name instead of hunting offsets. Loop: write a random attribute vector into every squad, sleep 75s while the game holidays, read back and harvest scorelines, repeat every 75s to take into account promoted teams.
Warning that cost me 120,000 matches Match Detail Level decides which simulator runs. On None everything resolves through the Quick Match Engine and effects come out 10x too small in the wrong order. Set the competition to All Competitive Matches. The quick path has its own QME_STAT_MANAGER sitting beside Poisson and Normal distribution objects, so it samples statistics rather than simulating football.
Dataset 15,951 full-engine matches. 47 visible + 5 hidden + 8 personality attributes randomised. CA and PA pinned at 150, morale pinned at 15. Attributes are stored 1-100 and displayed /5; personality attributes sit on a different record at plain 1-20.
Reading the numbers CA is pinned, so ability is a fixed budget and every weight answers "given fixed ability, where is it best spent?" Positioning at -9.6 does not mean Positioning harms you, rather it means CA spent there is not spent on Pace, and that trade costs about 10 points a season.
Validation Fitted on 5,011 matches, tested on 5,011 never used for fitting. Correlation with actual goal difference: Measured weights 0.392; Pace+Acceleration only 0.365; per-position 0.329; CA only 0.318; everything except the speed pair -0.043 Favourable quarter of the test set averages +1.79 goal difference, unfavourable -0.73. A two-segment fit searching every knee from 6 to 16 found no thresholds anywhere.
Global weights, CA cost and value Format: attribute, points per season, (CA cost of one point, points per CA point). CA cost measured separately by ramping a single attribute on an isolated player, R2=0.93. Hidden and personality attributes cost no CA at all. Ratios for near-free attributes are huge because the denominator is tiny, and read those as "free", not as literal multiples. Work Rate is the standout: nearly as valuable as Pace per point of ability, at a third of the cost. Decisions is the worst buy in the game, while the second most expensive attribute and negative. Pressure is worth 14.9 points a season and costs nothing.
Positional weights (0-20 within each group) Acceleration outranks Pace for CB and FB and the order flips further forward which seems plausible. The positional tables predicted worse out of sample than the pooled one (0.329 vs 0.392).
For goalkeepers, read the top two or three only. Reflexes and Agility separate cleanly; below that it is noise, and Handling ranking under Corners is not a finding. One keeper per match against ten outfielders means that column has about a tenth of the statistical power.
If an attribute point has a value of 10 from 4 to 5 and 1 from 14 to 15, how exactly do you set a fixed value for scoring in GS? That is the fundamental issue. You need linear scaling at least in the vast majority of values. You can't just say oh it's an average of lets say 3 because it'll be right only in a narrow bracket or maybe even never.
What you can do for attributes that have minimal or no value above a certain point is to set it's value to 0 (or whatever minimal value the calculations show) with a caveat that it only applies above X and that you need to filter for it. I am not disputing that some attributes might act that way, but that assigning them some value (by feel?) is misleading. And again FM-Arena tests showed that going from 18 to 19 to 20 pace is more or less just as valuable, it keeps it's scaling. Can you win with 18? Sure, but 20 is still better.
I mean it's kind of the basis of all these weighting systems, because if a point of pace was vastly different from 9 to 10 as opposed to 15 to 16 there would be no point in having a fixed value for an attribute weight.
So you either accept they are pretty much linear (with a few exceptions) or throw GS ratings out the window.
Wigo said: Guys i read all threads about atributes testing ant meta ones... this topic is the most intresting to me and i think i have a question that nobody answers. We found that speed in general is the king, but everytime i start a new save i see that (fm26) first season Harry Kane with speed of turtle is scoring unbelievable amount of goals and is gets player of the year award. So technically is speed is everything how come this happening??? so still plenty theories in my head that maybe for some positions speed is not so overpowered as we all think? Expand Because he is a highly skilled player in a great team that farms the league. Speed is very important but isn't everything. Just as in real life he scores 30 goals per season in Tottenham and 60+ in Bayern
GeorgeFloydOverdosed said: I think the indisputable contribution of this data is that it tells us what matters in FM26, and it seems like there's not much difference to FM24.
I don't know why you're finding every attribute effect to be linear, I think something going wrong somewhere in regards to that. I suspect it could be that you're randomizing attributes of every player at every club in the league, so that there's no longer set levels to compete against (i.e. in a real league, most players will have ~14 pace for example, or most GKs will have ~14 aerial reach, producing thresholds - it's not necessarily innate to the attribute itself, but a product of the relative competition. So 13 > 14 pace isn't a threshold in Vanarama South, but it is in Premier League). It's not entirely clear though if you're randomizing every player at every club, or just one club.
The part that intrigues me the most are the position weights. Have you simply divided the general results by the CA costs for each position, or have you done comparative testing for each position here? Expand I think FM-Arena attribute tests for the past few editions show linear effect for pretty much all of the attributes.
iezzex said: Just open up tactics and theres 90% of em is absolutely same shapes and main team instructions Only diff is PI but the result is almost same everywhere with 84.5 avg pts lmao Expand If you open the tactics table and sort out by release date there are maybe half of tactics with current meta shape. The rest are different. But yes you won't see them at the top of points table because they simply don't work as well.
Filter them by formation you want to play and browse through a few until you find roles and instructions you like.
apple said: Anyone else having problems with a ridiculous amount of penalties? Expand I find that when player skill is below your opponents you commit more penalites, probably some instructions contribute as well. Thing is the best tactics make you overachieve so much that you tend to play way above the actual skill level of the team so they can't actually cope on an individual level.
You can absolutely win PL without a single player having 18 pace or acceleration, as long as their CA is high enough. And that is the entire point. CA isn't worthless, it may not be best distributed but it is very very hard for a 120 CA player to outperform 160. Even with engineered perfect attribute distribution on 120 it's close and we know those don't actually exist ingame.
Not to mention that somehow you claim that under 18 pace you can't win the league and at the same time it's weighting is almost the same as 12 other attributes. So the scoring only works once you reach the minimum thresholds or what?
Bottom line is people are looking for who is a better player right now kind of rating, not who you can train to be better in 5 seasons. You severely undervalue pace and acceleration for example because you argue that it's easy to train it up, which might be true but it takes a lot of time and some luck.
Secondly if you compare 2 players with same "key" attributes, one is 120 CA the other is 160 CA the 2nd one will perform better. Probably even significantly, those 1% here and 1% there add up. Can a team of 120s finish 7th in PL? Sure, but most of people looking for ratings aren't after that. They simply want a scoring system that will tell them player X is better than player Y, and I think you've strayed further and further from the original intention with each iteration.
GrandRegentKone said: Anyone else getting it where if you put rebalance 1 the recovery sessions only show up as rest? and it doesnt let you select it even to change it? Expand You need to have a physio employed in the team, either that or sports scientist can't remember exactly.
Over and out.
And you have the weights at 18 and 12 respectively? This is why I always ask questions because you come out with statements like this where it's one way for a few days then it completely flips, but you are 100% certain each time. You had one set of weights one day, I found them strange and enquire about the reasoning behind them, then you completely switch the system with massive changes and somehow it's my fault for asking. The "truth" shouldn't switch around so much, sure there is some inaccuracy and with more data we can hone in better.
I generally like that you are pursuing these questions, but the attitude "I am right and everyone else is wrong and stupid" is kind of offputting, especially when you do 180 as often as you do. Like do the tests and lets discuss if it makes sense, what are the limitations, can you approach it differently etc. You've found some interesting stuff that is unfortunately the victim of the level of discussion in here.
Why you keep using league position as the end point still baffles me. Yes it has less variance because there are only 20 possible spots compared to lets say 100 possible points. You have to go really out of your way not to be in a +-3 places range. Additionally there is a lot more volatility in the middle of the table than at very top or bottom.
There was decent data for GKs from FM Arena, Harvestgreen, Orion and probably others and they mostly match with what you found so it's not exactly groundbreaking. Still valuable additional data from a different approach.
Click "Squad", then look in training tile - there are 2-4 dots in the top right corner. Each of those dots represents a team: senior, reserves, u21, u18 ... Click on the dot that you want to manage and then open the tile. From there you can edit whatever you want just like for the main team including intensity.
It reminds me, that some time ago a club from my town won a domestic championship with a technically very weak but big, strong and quick team (and doing mainly physical training). So it happens irl, but of course not on a very high level
And that was my whole point. That other attributes aren't completely useless and shouldn't be valued at 0.
A 24 yo CB with 160 CA and having 11 acc is completely normal. We got so used to abusing the broken training system that we somehow forgot that. What is actually wrong in this game is that you can engineer a 18+ pace player out of "almost anyone".
Now add 40 CA to them in nonessential attributes and sim again. If possible a fair assessment, not all CA into the absolute worst attributes you can think of.
Also the not engineered distribution is sort of true but not really. Using the meta schedules to boost physicals to levels that were obviously not intended is kind of engineering them, although they do happen "naturally" as opposed to straight up editing them.
Sorry what? Other than work rate, the other attributes are pretty much linear.
Obviously we aren't looking for the 6th decimal, there is margin of error. And they might even behave slightly nonlinearly but a linear function can be a very good approximation to reduce complexity.
Pace and Acceleration are worth ~3x anything else.
Everything technical measured at or below zero.
According to these resutls no attribute has a threshold, they pay linearly 1-20.
Once you divide by what a point costs in CA, the best buys are Work Rate, Jumping Reach and the hidden personality attributes, which are (mostly) free.
Method
FM's engine is C++ with MSVC RTTI, so every object carries its class name behind the vtable pointer, so you can find every player and scoreline by name instead of hunting offsets.
Loop: write a random attribute vector into every squad, sleep 75s while the game holidays, read back and harvest scorelines, repeat every 75s to take into account promoted teams.
Warning that cost me 120,000 matches
Match Detail Level decides which simulator runs.
On None everything resolves through the Quick Match Engine and effects come out 10x too small in the wrong order. Set the competition to All Competitive Matches.
The quick path has its own QME_STAT_MANAGER sitting beside Poisson and Normal distribution objects, so it samples statistics rather than simulating football.
Dataset
15,951 full-engine matches.
47 visible + 5 hidden + 8 personality attributes randomised.
CA and PA pinned at 150, morale pinned at 15.
Attributes are stored 1-100 and displayed /5;
personality attributes sit on a different record at plain 1-20.
Reading the numbers
CA is pinned, so ability is a fixed budget and every weight answers "given fixed ability, where is it best spent?"
Positioning at -9.6 does not mean Positioning harms you, rather it means CA spent there is not spent on Pace, and that trade costs about 10 points a season.
Validation
Fitted on 5,011 matches, tested on 5,011 never used for fitting.
Correlation with actual goal difference:
Measured weights 0.392; Pace+Acceleration only 0.365; per-position 0.329; CA only 0.318; everything except the speed pair -0.043
Favourable quarter of the test set averages +1.79 goal difference, unfavourable -0.73.
A two-segment fit searching every knee from 6 to 16 found no thresholds anywhere.
Global weights, CA cost and value
Format: attribute, points per season, (CA cost of one point, points per CA point).
CA cost measured separately by ramping a single attribute on an isolated player, R2=0.93.
Hidden and personality attributes cost no CA at all.
Ratios for near-free attributes are huge because the denominator is tiny, and read those as "free", not as literal multiples.
Work Rate is the standout: nearly as valuable as Pace per point of ability, at a third of the cost. Decisions is the worst buy in the game, while the second most expensive attribute and negative.
Pressure is worth 14.9 points a season and costs nothing.
Positional weights (0-20 within each group)
Acceleration outranks Pace for CB and FB and the order flips further forward which seems plausible.
The positional tables predicted worse out of sample than the pooled one (0.329 vs 0.392).
For goalkeepers, read the top two or three only.
Reflexes and Agility separate cleanly; below that it is noise, and Handling ranking under Corners is not a finding.
One keeper per match against ten outfielders means that column has about a tenth of the statistical power.
Shout out
to mavarobli / mavali / mark17072
@https://fm-arena.com/thread/20087-fmsuperscout-free-open-source-scouting-tool-for-fm26-with-a-meta-score-built-on-fm-arena-attribute-testing/
Anyway, back to the actual thread. It's an interesting approach with results more or less in line with what we already know so that's good. Could maybe argue about some of the attributes especially once broken down by position, but overall seems pretty decent.
What you can do for attributes that have minimal or no value above a certain point is to set it's value to 0 (or whatever minimal value the calculations show) with a caveat that it only applies above X and that you need to filter for it. I am not disputing that some attributes might act that way, but that assigning them some value (by feel?) is misleading. And again FM-Arena tests showed that going from 18 to 19 to 20 pace is more or less just as valuable, it keeps it's scaling. Can you win with 18? Sure, but 20 is still better.
So they are not linear. Am i right? Or "linear for pretty much all of the attributes"?
https://fm-arena.com/find-comment/40302/
I mean it's kind of the basis of all these weighting systems, because if a point of pace was vastly different from 9 to 10 as opposed to 15 to 16 there would be no point in having a fixed value for an attribute weight.
So you either accept they are pretty much linear (with a few exceptions) or throw GS ratings out the window.
Because he is a highly skilled player in a great team that farms the league. Speed is very important but isn't everything. Just as in real life he scores 30 goals per season in Tottenham and 60+ in Bayern
I don't know why you're finding every attribute effect to be linear, I think something going wrong somewhere in regards to that. I suspect it could be that you're randomizing attributes of every player at every club in the league, so that there's no longer set levels to compete against (i.e. in a real league, most players will have ~14 pace for example, or most GKs will have ~14 aerial reach, producing thresholds - it's not necessarily innate to the attribute itself, but a product of the relative competition. So 13 > 14 pace isn't a threshold in Vanarama South, but it is in Premier League). It's not entirely clear though if you're randomizing every player at every club, or just one club.
The part that intrigues me the most are the position weights. Have you simply divided the general results by the CA costs for each position, or have you done comparative testing for each position here?
I think FM-Arena attribute tests for the past few editions show linear effect for pretty much all of the attributes.
Only diff is PI but the result is almost same everywhere with 84.5 avg pts lmao
If you open the tactics table and sort out by release date there are maybe half of tactics with current meta shape. The rest are different. But yes you won't see them at the top of points table because they simply don't work as well.
Filter them by formation you want to play and browse through a few until you find roles and instructions you like.
I find that when player skill is below your opponents you commit more penalites, probably some instructions contribute as well. Thing is the best tactics make you overachieve so much that you tend to play way above the actual skill level of the team so they can't actually cope on an individual level.
Not to mention that somehow you claim that under 18 pace you can't win the league and at the same time it's weighting is almost the same as 12 other attributes. So the scoring only works once you reach the minimum thresholds or what?
Secondly if you compare 2 players with same "key" attributes, one is 120 CA the other is 160 CA the 2nd one will perform better. Probably even significantly, those 1% here and 1% there add up. Can a team of 120s finish 7th in PL? Sure, but most of people looking for ratings aren't after that. They simply want a scoring system that will tell them player X is better than player Y, and I think you've strayed further and further from the original intention with each iteration.
You need to have a physio employed in the team, either that or sports scientist can't remember exactly.