I thought I could speed up the process by predicting which attributes matter and which don't, but it turns out I have to throw all my preconceptions out the window and just resort to good ol' trial and error.
A team of these (hiddens ~8-13; DC jump 17) achieves 4th-5th in the Premier League:
I wonder how many people even here actually realize that you can't win the Premier League unless you have all your players at 18+ pace/acc even when you're complimenting it with a few of the key attributes such as dribbling and concentration.
So for instance if I decrease stamina from '13' to '8', my team goes from 4th/5th to bottom half. You might think (as I did), just increase pace/acc to 16 to compensate. But actually in this case doing that does hardly anything. In fact, even if you increase them to 17, you still won't win, but you'll come a close 2nd.
Now if you have perfect hidden attributes then you can have 17.. 16.. even 15 pace/acc, you can win the Premier League no problem, but we're talking about what is realistic here.
Some things to note:
- 17 jump on DC (or some other player perhaps?) is necessary, big difference between 15 and 17 here. - Surprisingly it does not appear that Jumping Reach is an all-or-nothing attribute. There is a clear and important benefit to all players having moderate instead of low Jumping Reach. - Agility, work rate and composure are the least crucial of the remaining attributes, hence why I have lowered them to 10-11 with no adverse consequence for my team position - A lot of attributes are set at '8' because that's what you can roughly expect your minimums to be for a player once you're in the Premier League (even if it's a newgen you signed 3 years ago in League 2 say). This is to make the template as accurate to reality as possible, and yes '8' makes a difference to '1'. Of course players will often have a '3' in something, but you can either train this up or compensate for it in other ways and sometimes it just doesn't matter (i.e. set piece attributes). - Some essential attributes are at '8', but only because the required amount wasn't higher than '8'. This includes vision, finishing, strength for instance.
Now I have tried using HarvestGreen, Orion, my own, etc. data to try and optimize for positional differences. But everything I tried just made the result worse, even gentle adjustments. So I am now faced with 2 options, after being satisfied I've whittled it down to the essential attributes as a whole:
a) Do dozens or hundreds of positional trial and error changes, or b) Give up on it and just go forward with what I have so far
The more I think about it, the more I am convinced to choose (b). I can always return to positional refinement later. Keep in mind I've already tested a lot of different things, and I know from my 1 CA testing that certain attributes are simply universal (i.e. pace, acc, drib is as essential for DC as it is for AMR).
GeorgeFloydOverdosed said: A team of these (hiddens ~8-13; DC jump 17) achieves 4th-5th in the Premier League: Expand
This is surprising to me. I looked at how effective a player like this would be using your blended weights and they don't appear to be anywhere near good enough to achieve a top-four finish.
I created a player using the same attribute values you listed (see screenshot) and calculated his rating at all outfield positions (I gave him 17 jumping reach when calculating centre-back rating). Here's what I got:
| Position | Rating | |----|---------| | FB | 74.4% | | CB | 76.98% | | WB | 73.97% | | DM | 67.56% | | W | 65.42% | | MC | 68.73% | | AM | 69.26% | | ST | 62.37% |
Don't give too much thought to the actual values because my weights are arbitrary but here's the rating of Newcastle's best XI on May 4th, 2025 in my current save (the season is over, they've finished in 4th place):
| Player | Position | Rating | |----|--------|--------| | Nick Pope | GK | 75.65% | | Ferdi Kadıoğlu | DL | 85.35% | | Gianluca Mancini | DC | 77.97% | | Sven Botman | DC | 78.32% | | Tino Livramento | DR | 78.93% | | Alexander Isak | ML | 79.31% | | Joe Willock | MC | 74.17% | | Sandro Tonali | MC | 76.27% | | Joelinton | MR | 74.78% | | Bryan Mbeumo | ST | 69.2% | | Anthony Gordon | ST | 68.2% |
The low-CA player you posted is at least 3-4% worse off than this real team that finished 4th in my game (in 5th place was an even stronger Manchester United). If I make Newcastle play with a defensive midfielder, it picks Tonali and Willock who score 74% each.
If you holidayed with a tactic which isn't considered overpowered, I see this as a bit of an indictment of the weights we've been talking about.
Of course, if you've crunched the numbers yourself and actually find your team is considered much stronger than your closest competition (ideally without the use of Genie Scout thumbing the scale mysteriously) then I'll hold my hands up. This has left me confused though.
EDIT: It strikes me that this might be what you meant when you wrote this:
GeorgeFloydOverdosed said: Now I have tried using HarvestGreen, Orion, my own, etc. data to try and optimize for positional differences. But everything I tried just made the result worse, even gentle adjustments. Expand
DOUBLE EDIT: Ignore the CA in the Lewdandowski screenshot. Ignore everything except the ability values and the overall rating. I took the real player and modified his attributes just to see what the formula would spit out.
LightningFlik said: This is surprising to me. I looked at how effective a player like this would be using your blended weights and they don't appear to be anywhere near good enough to achieve a top-four finish.
I created a player using the same attribute values you listed (see screenshot) and calculated his rating at all outfield positions (I gave him 17 jumping reach when calculating centre-back rating). Here's what I got:
| Position | Rating | |----|---------| | FB | 74.4% | | CB | 76.98% | | WB | 73.97% | | DM | 67.56% | | W | 65.42% | | MC | 68.73% | | AM | 69.26% | | ST | 62.37% |
Don't give too much thought to the actual values because my weights are arbitrary but here's the rating of Newcastle's best XI on May 4th, 2025 in my current save (the season is over, they've finished in 4th place):
| Player | Position | Rating | |----|--------|--------| | Nick Pope | GK | 75.65% | | Ferdi Kadıoğlu | DL | 85.35% | | Gianluca Mancini | DC | 77.97% | | Sven Botman | DC | 78.32% | | Tino Livramento | DR | 78.93% | | Alexander Isak | ML | 79.31% | | Joe Willock | MC | 74.17% | | Sandro Tonali | MC | 76.27% | | Joelinton | MR | 74.78% | | Bryan Mbeumo | ST | 69.2% | | Anthony Gordon | ST | 68.2% |
The low-CA player you posted is at least 3-4% worse off than this real team that finished 4th in my game (in 5th place was an even stronger Manchester United). If I make Newcastle play with a defensive midfielder, it picks Tonali and Willock who score 74% each.
If you holidayed with a tactic which isn't considered overpowered, I see this as a bit of an indictment of the weights we've been talking about.
Of course, if you've crunched the numbers yourself and actually find your team is considered much stronger than your closest competition (ideally without the use of Genie Scout thumbing the scale mysteriously) then I'll hold my hands up. This has left me confused though.
EDIT: It strikes me that this might be what you meant when you wrote this:
DOUBLE EDIT: Ignore the CA in the Lewdandowski screenshot. Ignore everything except the ability values and the overall rating. I took the real player and modified his attributes just to see what the formula would spit out. Expand
Rain said: What is this screenshot from? Expand An analysis tool I'm working on. It's like those websites that ask you to export all your players as a webpage for analysis, except this reads data straight from the game (because who's got time for that?)
This makes it more like Genie Scout, but without the ludicrous startup time. Since then I've added some editing tools that make it similar to FMRTE too, except mine is free.
It's Linux only at the moment though (because that's all I use and the only machine I have to develop on, so I can't test the memory reading code on anything else).
captain3 said: In my GS I can't filter by division since all the players belong to "Jupiter Pro League". Does anyone have this problem? Expand
Yeah, I don't how many FM versions it's had this problem, but it's like that now for the FM24 version. This and the long load time and lack of full weighting customization are making me consider FMRTE again.
LightningFlik said: This is surprising to me. I looked at how effective a player like this would be using your blended weights and they don't appear to be anywhere near good enough to achieve a top-four finish.
I created a player using the same attribute values you listed (see screenshot) and calculated his rating at all outfield positions (I gave him 17 jumping reach when calculating centre-back rating). Here's what I got:
| Position | Rating | |----|---------| | FB | 74.4% | | CB | 76.98% | | WB | 73.97% | | DM | 67.56% | | W | 65.42% | | MC | 68.73% | | AM | 69.26% | | ST | 62.37% |
Don't give too much thought to the actual values because my weights are arbitrary but here's the rating of Newcastle's best XI on May 4th, 2025 in my current save (the season is over, they've finished in 4th place):
| Player | Position | Rating | |----|--------|--------| | Nick Pope | GK | 75.65% | | Ferdi Kadıoğlu | DL | 85.35% | | Gianluca Mancini | DC | 77.97% | | Sven Botman | DC | 78.32% | | Tino Livramento | DR | 78.93% | | Alexander Isak | ML | 79.31% | | Joe Willock | MC | 74.17% | | Sandro Tonali | MC | 76.27% | | Joelinton | MR | 74.78% | | Bryan Mbeumo | ST | 69.2% | | Anthony Gordon | ST | 68.2% |
The low-CA player you posted is at least 3-4% worse off than this real team that finished 4th in my game (in 5th place was an even stronger Manchester United). If I make Newcastle play with a defensive midfielder, it picks Tonali and Willock who score 74% each.
If you holidayed with a tactic which isn't considered overpowered, I see this as a bit of an indictment of the weights we've been talking about.
Of course, if you've crunched the numbers yourself and actually find your team is considered much stronger than your closest competition (ideally without the use of Genie Scout thumbing the scale mysteriously) then I'll hold my hands up. This has left me confused though.
EDIT: It strikes me that this might be what you meant when you wrote this:
DOUBLE EDIT: Ignore the CA in the Lewdandowski screenshot. Ignore everything except the ability values and the overall rating. I took the real player and modified his attributes just to see what the formula would spit out. Expand I only semi-understand what you are saying, so forgive me if I'm wrong in assuming that what you are essentially pointing out is that the team of players I said achieve 4th, are not in line with the 'Blended' ratings file I posted, and that your own Newcastle result of 4th (same position) yet with players with significantly higher % rating is further evidence that the Blended file is out of whack.
If that is what you mean, then you are correct. The 'Blended' ratings file is only at best an a guesstimate of what will win you the Premier league. The 15 pace/acc team template is concrete and precise about what achieves ~4th-5th.
Here's the problems though if one is to simply ditch the guesstimates and go with just what is found through brute force trial and error:
1) Attributes have to be realistically achievable. No newly promoted team is going to have a full roster of 18+ pace/acc players, in fact they'll likely have none. Optimals for certain positions won't even exist. And of what is available, you have to set a purchase value limit on it (say £1mil/player for newly promoted team?).
2) The most optimal configuration is the most inflexible. If one uses a 17 pace/acc template as the basis, a single player dipping to 16 could be the difference between 5th and relegation. If one uses my 15 pace/acc template, dips anywhere have the same consequences. I've actually been doing a realistic test of Luton, giving them the best players according to that new 15 pace/acc template that cost less than £500k and they survived relegation but not by much (in my view this is actually pretty good as it shows it's at least viable realistically). When I tried doing it with free transfers only before that, they were relegated.
3) Aiming for avoiding relegation, 4th/5th, or 1st. Long term success or short term best. With the 15 pace/acc template, the idea is that 1st season you come 4th-5th (or at least survive relegation), and then you train up that 15 pace/acc to 18 pace/acc to come 1st through the high pace/acc alone. If you just go with 16 pace/acc players regardless of mentals/technicals, you will get relegated. If you go with old players with better stats and cheap but no future, you will finish mid-high in 1st season but won't win your 2nd season.
4) Genie Scout is particularly bad in dealing with inflexibility. If your players absolutely need at least 14 pace/acc, even 100% pace/acc weight isn't going to guarantee that, only extra filtering for that will. But if you're going to filter anyway, then you may as well set pace/acc to its actual weight rather than just trying to ensure 14+ pace/acc bubbles to the top. But then also sometimes we just won't bother with filters.. so realistically we actually really do need flexibility.
Now let's distill this to the main dilemma. My Genie Scout rating files and HarvestGreen's results (and training) have a heavy physical focus. The idea is to get to that threshold of ~17-18 pace/acc ASAP so no one can catch you and you win everything. But my thinking is, given you need to survive at least the first season or two without that, you have to first work out a way of getting lower pace/acc players to work for you. And it turns out it is possible, but it may be unrealistic or too inflexible. And you kind of have to go one way or the other, because if you pacemaxx (especially with full rest training) you will never get those mentals/technicals back, and more importantly it also turns out that simply taking my template and upping the pace/acc to 16 doesn't make it do much better.
I'm still mulling over this. The Luton result shows that low pace/acc is viable, but not as much I'd hoped. I've got a few ideas, and hopefully at least one of them will sure things up.
GeorgeFloydOverdosed said: you are essentially pointing out is that the team of players I said achieve 4th, are not in line with the 'Blended' ratings file I posted, and that your own Newcastle result of 4th (same position) yet with players with significantly higher % rating is further evidence that the Blended file is out of whack. Expand Yeah, precisely. It would have been lovely if the weighted formula produced similar ratings for your artificial player and a real-world example but the real team is clearly regarded as much better.
If you're seeing a certain synergy between some attributes or significant performance gains/drop-off after crossing attribute thresholds, it would suggest that the actual formula is more complicated than just summing up weighted values. It sounds obvious but I wouldn't put it past Sports Interactive given how many corners they've cut.
I think Harvest Green in one of his spreadsheets had a formula which changed weights depending on how high or low the values were; I might try to implement that and compare the players that way.
( edited 8 hours, 15 min ago by GeorgeFloydOverdosed )
Share this post#190
Link to the post:
LightningFlik said: Yeah, precisely. It would have been lovely if the weighted formula produced similar ratings for your artificial player and a real-world example but the real team is clearly regarded as much better.
If you're seeing a certain synergy between some attributes or significant performance gains/drop-off after crossing attribute thresholds, it would suggest that the actual formula is more complicated than just summing up weighted values. It sounds obvious but I wouldn't put it past Sports Interactive given how many corners they've cut.
I think Harvest Green in one of his spreadsheets had a formula which changed weights depending on how high or low the values were; I might try to implement that and compare the players that way. Expand I think what you possibly have in mind is the precise formula for CA which FM Scout did an article on, and EBFM tried to work out but even he couldn't fully figure it out. It is true that essentially the more all over the place your attributes are, the more/less CA it costs.
You may also have in mind, or at least I'll point out, HarvestGreen's synergies he found between certain attributes (invented example: say finishing + composure to be worth more than either separately). He also hypothesizes that for many attributes, the number is proficiency as well as tendency and that that may explain a few things.
For the moment, I'm putting CA aside and just considering what attributes and their numbers are actually required to succeed - an attribute floor to then build on.
Most simply put, yes the way attributes are being calculated is complex. Messy even. I've dug deep into this, finding everything I could about what SI staff have said in regards to how attributes work under the hood and analyzing how things worked in the old CM games. Turned out most of the stuff the SI staff said was rubbish, so I ended up triangulating between the data from HarvestGreen, Orion, and my own 1 CA testing. But even Orion's data I ended up factoring out, and now I'm seeing that what remains isn't the whole truth either.
To communicate this, I have to step back and present the bigger picture to you. Either take my word for it that these results are all substantially different, or go and take a look at the data yourself:
1) ykykyk05251's FM21/FM22 tests that purportedly used mass testing using AI doing knockout tournaments over 40 million matches and then pitting it against human picks in 100,000 matches to determine even ideal positional attributes. Supposedly dribbling is more important than pace for ST, stamina is most important for DL/DR, and decisions is 40% weight for every position.
2) Zippo's tests which seem to have been ongoing for a few years and claims that composure is pointless above 8 and concentration + determination combined is less important than stamina.
3) HarvestGreen's data which says that decisions is ~0% benefit and that concentration is more important than stamina (at both 10 > 18 and 10 > 1).
4) Orion's data that different positions favor different things, but stamina doesn't feature in the top 8 for DL/DR.
5) My 1 CA testing which shows you can win the Premier League with a team of 1 CA players with it mostly just being about having high pace/acc/drib/jump, but also some positional quirks like high concentration for DC/DL/DR but unnecessary on other positions.
6) My 15 pace/acc template, where having a bunch of mental/technical stats at ~13 is strictly necessary to come ~4-5th instead of relegated. And I find that Orion's positional adjustments, or even my own ones like concentration on only DC/DL/DR doesn't work here.
Now is everyone lying, or making serious errors here? I don't think it's either. I think that the process changes at different levels of attributes, in ways that are too difficult to predict and require brute force testing to deduce.
A hypothetical example as how this plays out in the match engine:
AML with 18 pace/acc > exceeds DC's 16 acc (speed rate) and 15 agility (direction change speed) > no further calculations, AML 'escapes' and ends up scoring against GK's reflexes
AML with 16 pace/acc > matched by DC's 16 acc and 15 agil > AML 14 balance challenged by DC 12 strength > DC outmuscle failed > AML decision (8) continue dribble (13 proficiency/tendency) > DC decision (13) slide in with both feet (tackling 13, bravery 14, aggression 15, dirtiness 15) > AML escape (14 balance, 18 agility) or AML deposed of ball (14 balance, 9 agility)
From this hypothetical we can see how when pace/acc 18+, a lot of those technicals/mentals like tackling and bravery become pointless, but if you lower the pace/acc to 16 then the reverse starts becoming true. If you lower pace/acc to 10, probably another process plays out. And yet of course exceedingly high pace/acc is always your best bet.
Then of course there are things like 'pooled' and compensatory attributes to further complicate the process, where it seems to be ok if your ST has 1 finishing, so long as someone else in your team has 7 finishing. Stamina seems necessary to preserve low pace/acc, but with 20 pace/acc you have so much to spare that you can get away with a team of 1 stamina (that's how I'm reading stamina's results anyway).
So returning to Genie Scout ratings, for newly promoted team 15 pace/acc with good mentals/technicals is what works, but what also would work if you could get it is 18+ pace/acc team with limited mentals/technicals. So how do you value mentals/technicals? Something inbetween? Then no one wins, as the 16 pace/acc thing I explain. Different set of values for each? Too much work and I think everyone likes just one simple rating set, plus I think one set is feasible. Like a good knife, it needs to be sharp but have some flex to it, so that even if you *can* have those 18 pace/acc players but miscalculate things and end up with a team of duds/injuries/etc then you'll still survive relegation at least.
@GeorgeFloydOverdosed So if you are starting a save with a small club in a small league with little reputation I'm wondering what the best play is. Do you just go for the best players you can get according to current GS rating at the start or would you look for the best potential ratings?
Rain said: @GeorgeFloydOverdosed So if you are starting a save with a small club in a small league with little reputation I'm wondering what the best play is. Do you just go for the best players you can get according to current GS rating at the start or would you look for the best potential ratings? Expand
Use current rating + take into account or filter CA-PA gap. Genie Scout potential rating will give you a rough idea how good they can become, but it will be inaccurate.
You need to first establish your club to the point that you can actually attract high potential players. Plus you aren't really developing anyone in Latvian 2nd division with bad facilities, bad coaches and likely semipro status
captain3 said: Someone can send me that print screens for the attributes weights for each role for the FM26 file? I can't upload on my GS
I dm'd them to you
captain3 said: Someone can send me that print screens for the attributes weights for each role for the FM26 file? I can't upload on my GS
1. Put file in Genie scout rating folder
2. open GS, klik rating, choose your rating rx5m3t
3. load player on GS, you will see every player rating
I thought I could speed up the process by predicting which attributes matter and which don't, but it turns out I have to throw all my preconceptions out the window and just resort to good ol' trial and error.
A team of these (hiddens ~8-13; DC jump 17) achieves 4th-5th in the Premier League:
I wonder how many people even here actually realize that you can't win the Premier League unless you have all your players at 18+ pace/acc even when you're complimenting it with a few of the key attributes such as dribbling and concentration.
So for instance if I decrease stamina from '13' to '8', my team goes from 4th/5th to bottom half. You might think (as I did), just increase pace/acc to 16 to compensate. But actually in this case doing that does hardly anything. In fact, even if you increase them to 17, you still won't win, but you'll come a close 2nd.
Now if you have perfect hidden attributes then you can have 17.. 16.. even 15 pace/acc, you can win the Premier League no problem, but we're talking about what is realistic here.
Some things to note:
- 17 jump on DC (or some other player perhaps?) is necessary, big difference between 15 and 17 here.
- Surprisingly it does not appear that Jumping Reach is an all-or-nothing attribute. There is a clear and important benefit to all players having moderate instead of low Jumping Reach.
- Agility, work rate and composure are the least crucial of the remaining attributes, hence why I have lowered them to 10-11 with no adverse consequence for my team position
- A lot of attributes are set at '8' because that's what you can roughly expect your minimums to be for a player once you're in the Premier League (even if it's a newgen you signed 3 years ago in League 2 say). This is to make the template as accurate to reality as possible, and yes '8' makes a difference to '1'. Of course players will often have a '3' in something, but you can either train this up or compensate for it in other ways and sometimes it just doesn't matter (i.e. set piece attributes).
- Some essential attributes are at '8', but only because the required amount wasn't higher than '8'. This includes vision, finishing, strength for instance.
Now I have tried using HarvestGreen, Orion, my own, etc. data to try and optimize for positional differences. But everything I tried just made the result worse, even gentle adjustments. So I am now faced with 2 options, after being satisfied I've whittled it down to the essential attributes as a whole:
a) Do dozens or hundreds of positional trial and error changes, or
b) Give up on it and just go forward with what I have so far
The more I think about it, the more I am convinced to choose (b). I can always return to positional refinement later. Keep in mind I've already tested a lot of different things, and I know from my 1 CA testing that certain attributes are simply universal (i.e. pace, acc, drib is as essential for DC as it is for AMR).
In my GS I can't filter by division since all the players belong to "Jupiter Pro League". Does anyone have this problem?
GeorgeFloydOverdosed said: A team of these (hiddens ~8-13; DC jump 17) achieves 4th-5th in the Premier League:
This is surprising to me. I looked at how effective a player like this would be using your blended weights and they don't appear to be anywhere near good enough to achieve a top-four finish.
I created a player using the same attribute values you listed (see screenshot) and calculated his rating at all outfield positions (I gave him 17 jumping reach when calculating centre-back rating). Here's what I got:
| Position | Rating |
|----|---------|
| FB | 74.4% |
| CB | 76.98% |
| WB | 73.97% |
| DM | 67.56% |
| W | 65.42% |
| MC | 68.73% |
| AM | 69.26% |
| ST | 62.37% |
Don't give too much thought to the actual values because my weights are arbitrary but here's the rating of Newcastle's best XI on May 4th, 2025 in my current save (the season is over, they've finished in 4th place):
| Player | Position | Rating |
|----|--------|--------|
| Nick Pope | GK | 75.65% |
| Ferdi Kadıoğlu | DL | 85.35% |
| Gianluca Mancini | DC | 77.97% |
| Sven Botman | DC | 78.32% |
| Tino Livramento | DR | 78.93% |
| Alexander Isak | ML | 79.31% |
| Joe Willock | MC | 74.17% |
| Sandro Tonali | MC | 76.27% |
| Joelinton | MR | 74.78% |
| Bryan Mbeumo | ST | 69.2% |
| Anthony Gordon | ST | 68.2% |
The low-CA player you posted is at least 3-4% worse off than this real team that finished 4th in my game (in 5th place was an even stronger Manchester United). If I make Newcastle play with a defensive midfielder, it picks Tonali and Willock who score 74% each.
If you holidayed with a tactic which isn't considered overpowered, I see this as a bit of an indictment of the weights we've been talking about.
Of course, if you've crunched the numbers yourself and actually find your team is considered much stronger than your closest competition (ideally without the use of Genie Scout thumbing the scale mysteriously) then I'll hold my hands up. This has left me confused though.
EDIT: It strikes me that this might be what you meant when you wrote this:
GeorgeFloydOverdosed said: Now I have tried using HarvestGreen, Orion, my own, etc. data to try and optimize for positional differences. But everything I tried just made the result worse, even gentle adjustments.
DOUBLE EDIT: Ignore the CA in the Lewdandowski screenshot. Ignore everything except the ability values and the overall rating. I took the real player and modified his attributes just to see what the formula would spit out.
LightningFlik said: This is surprising to me. I looked at how effective a player like this would be using your blended weights and they don't appear to be anywhere near good enough to achieve a top-four finish.
I created a player using the same attribute values you listed (see screenshot) and calculated his rating at all outfield positions (I gave him 17 jumping reach when calculating centre-back rating). Here's what I got:
| Position | Rating |
|----|---------|
| FB | 74.4% |
| CB | 76.98% |
| WB | 73.97% |
| DM | 67.56% |
| W | 65.42% |
| MC | 68.73% |
| AM | 69.26% |
| ST | 62.37% |
Don't give too much thought to the actual values because my weights are arbitrary but here's the rating of Newcastle's best XI on May 4th, 2025 in my current save (the season is over, they've finished in 4th place):
| Player | Position | Rating |
|----|--------|--------|
| Nick Pope | GK | 75.65% |
| Ferdi Kadıoğlu | DL | 85.35% |
| Gianluca Mancini | DC | 77.97% |
| Sven Botman | DC | 78.32% |
| Tino Livramento | DR | 78.93% |
| Alexander Isak | ML | 79.31% |
| Joe Willock | MC | 74.17% |
| Sandro Tonali | MC | 76.27% |
| Joelinton | MR | 74.78% |
| Bryan Mbeumo | ST | 69.2% |
| Anthony Gordon | ST | 68.2% |
The low-CA player you posted is at least 3-4% worse off than this real team that finished 4th in my game (in 5th place was an even stronger Manchester United). If I make Newcastle play with a defensive midfielder, it picks Tonali and Willock who score 74% each.
If you holidayed with a tactic which isn't considered overpowered, I see this as a bit of an indictment of the weights we've been talking about.
Of course, if you've crunched the numbers yourself and actually find your team is considered much stronger than your closest competition (ideally without the use of Genie Scout thumbing the scale mysteriously) then I'll hold my hands up. This has left me confused though.
EDIT: It strikes me that this might be what you meant when you wrote this:
DOUBLE EDIT: Ignore the CA in the Lewdandowski screenshot. Ignore everything except the ability values and the overall rating. I took the real player and modified his attributes just to see what the formula would spit out.
What is this screenshot from?
Rain said: What is this screenshot from?
An analysis tool I'm working on. It's like those websites that ask you to export all your players as a webpage for analysis, except this reads data straight from the game (because who's got time for that?)
This makes it more like Genie Scout, but without the ludicrous startup time. Since then I've added some editing tools that make it similar to FMRTE too, except mine is free.
It's Linux only at the moment though (because that's all I use and the only machine I have to develop on, so I can't test the memory reading code on anything else).
captain3 said: In my GS I can't filter by division since all the players belong to "Jupiter Pro League". Does anyone have this problem?
Yeah, I don't how many FM versions it's had this problem, but it's like that now for the FM24 version. This and the long load time and lack of full weighting customization are making me consider FMRTE again.
LightningFlik said: This is surprising to me. I looked at how effective a player like this would be using your blended weights and they don't appear to be anywhere near good enough to achieve a top-four finish.
I created a player using the same attribute values you listed (see screenshot) and calculated his rating at all outfield positions (I gave him 17 jumping reach when calculating centre-back rating). Here's what I got:
| Position | Rating |
|----|---------|
| FB | 74.4% |
| CB | 76.98% |
| WB | 73.97% |
| DM | 67.56% |
| W | 65.42% |
| MC | 68.73% |
| AM | 69.26% |
| ST | 62.37% |
Don't give too much thought to the actual values because my weights are arbitrary but here's the rating of Newcastle's best XI on May 4th, 2025 in my current save (the season is over, they've finished in 4th place):
| Player | Position | Rating |
|----|--------|--------|
| Nick Pope | GK | 75.65% |
| Ferdi Kadıoğlu | DL | 85.35% |
| Gianluca Mancini | DC | 77.97% |
| Sven Botman | DC | 78.32% |
| Tino Livramento | DR | 78.93% |
| Alexander Isak | ML | 79.31% |
| Joe Willock | MC | 74.17% |
| Sandro Tonali | MC | 76.27% |
| Joelinton | MR | 74.78% |
| Bryan Mbeumo | ST | 69.2% |
| Anthony Gordon | ST | 68.2% |
The low-CA player you posted is at least 3-4% worse off than this real team that finished 4th in my game (in 5th place was an even stronger Manchester United). If I make Newcastle play with a defensive midfielder, it picks Tonali and Willock who score 74% each.
If you holidayed with a tactic which isn't considered overpowered, I see this as a bit of an indictment of the weights we've been talking about.
Of course, if you've crunched the numbers yourself and actually find your team is considered much stronger than your closest competition (ideally without the use of Genie Scout thumbing the scale mysteriously) then I'll hold my hands up. This has left me confused though.
EDIT: It strikes me that this might be what you meant when you wrote this:
DOUBLE EDIT: Ignore the CA in the Lewdandowski screenshot. Ignore everything except the ability values and the overall rating. I took the real player and modified his attributes just to see what the formula would spit out.
I only semi-understand what you are saying, so forgive me if I'm wrong in assuming that what you are essentially pointing out is that the team of players I said achieve 4th, are not in line with the 'Blended' ratings file I posted, and that your own Newcastle result of 4th (same position) yet with players with significantly higher % rating is further evidence that the Blended file is out of whack.
If that is what you mean, then you are correct. The 'Blended' ratings file is only at best an a guesstimate of what will win you the Premier league. The 15 pace/acc team template is concrete and precise about what achieves ~4th-5th.
Here's the problems though if one is to simply ditch the guesstimates and go with just what is found through brute force trial and error:
1) Attributes have to be realistically achievable. No newly promoted team is going to have a full roster of 18+ pace/acc players, in fact they'll likely have none. Optimals for certain positions won't even exist. And of what is available, you have to set a purchase value limit on it (say £1mil/player for newly promoted team?).
2) The most optimal configuration is the most inflexible. If one uses a 17 pace/acc template as the basis, a single player dipping to 16 could be the difference between 5th and relegation. If one uses my 15 pace/acc template, dips anywhere have the same consequences. I've actually been doing a realistic test of Luton, giving them the best players according to that new 15 pace/acc template that cost less than £500k and they survived relegation but not by much (in my view this is actually pretty good as it shows it's at least viable realistically). When I tried doing it with free transfers only before that, they were relegated.
3) Aiming for avoiding relegation, 4th/5th, or 1st. Long term success or short term best. With the 15 pace/acc template, the idea is that 1st season you come 4th-5th (or at least survive relegation), and then you train up that 15 pace/acc to 18 pace/acc to come 1st through the high pace/acc alone. If you just go with 16 pace/acc players regardless of mentals/technicals, you will get relegated. If you go with old players with better stats and cheap but no future, you will finish mid-high in 1st season but won't win your 2nd season.
4) Genie Scout is particularly bad in dealing with inflexibility. If your players absolutely need at least 14 pace/acc, even 100% pace/acc weight isn't going to guarantee that, only extra filtering for that will. But if you're going to filter anyway, then you may as well set pace/acc to its actual weight rather than just trying to ensure 14+ pace/acc bubbles to the top. But then also sometimes we just won't bother with filters.. so realistically we actually really do need flexibility.
Now let's distill this to the main dilemma. My Genie Scout rating files and HarvestGreen's results (and training) have a heavy physical focus. The idea is to get to that threshold of ~17-18 pace/acc ASAP so no one can catch you and you win everything. But my thinking is, given you need to survive at least the first season or two without that, you have to first work out a way of getting lower pace/acc players to work for you. And it turns out it is possible, but it may be unrealistic or too inflexible. And you kind of have to go one way or the other, because if you pacemaxx (especially with full rest training) you will never get those mentals/technicals back, and more importantly it also turns out that simply taking my template and upping the pace/acc to 16 doesn't make it do much better.
I'm still mulling over this. The Luton result shows that low pace/acc is viable, but not as much I'd hoped. I've got a few ideas, and hopefully at least one of them will sure things up.
GeorgeFloydOverdosed said: you are essentially pointing out is that the team of players I said achieve 4th, are not in line with the 'Blended' ratings file I posted, and that your own Newcastle result of 4th (same position) yet with players with significantly higher % rating is further evidence that the Blended file is out of whack.
Yeah, precisely. It would have been lovely if the weighted formula produced similar ratings for your artificial player and a real-world example but the real team is clearly regarded as much better.
If you're seeing a certain synergy between some attributes or significant performance gains/drop-off after crossing attribute thresholds, it would suggest that the actual formula is more complicated than just summing up weighted values. It sounds obvious but I wouldn't put it past Sports Interactive given how many corners they've cut.
I think Harvest Green in one of his spreadsheets had a formula which changed weights depending on how high or low the values were; I might try to implement that and compare the players that way.
LightningFlik said: Yeah, precisely. It would have been lovely if the weighted formula produced similar ratings for your artificial player and a real-world example but the real team is clearly regarded as much better.
If you're seeing a certain synergy between some attributes or significant performance gains/drop-off after crossing attribute thresholds, it would suggest that the actual formula is more complicated than just summing up weighted values. It sounds obvious but I wouldn't put it past Sports Interactive given how many corners they've cut.
I think Harvest Green in one of his spreadsheets had a formula which changed weights depending on how high or low the values were; I might try to implement that and compare the players that way.
I think what you possibly have in mind is the precise formula for CA which FM Scout did an article on, and EBFM tried to work out but even he couldn't fully figure it out. It is true that essentially the more all over the place your attributes are, the more/less CA it costs.
You may also have in mind, or at least I'll point out, HarvestGreen's synergies he found between certain attributes (invented example: say finishing + composure to be worth more than either separately). He also hypothesizes that for many attributes, the number is proficiency as well as tendency and that that may explain a few things.
For the moment, I'm putting CA aside and just considering what attributes and their numbers are actually required to succeed - an attribute floor to then build on.
Most simply put, yes the way attributes are being calculated is complex. Messy even. I've dug deep into this, finding everything I could about what SI staff have said in regards to how attributes work under the hood and analyzing how things worked in the old CM games. Turned out most of the stuff the SI staff said was rubbish, so I ended up triangulating between the data from HarvestGreen, Orion, and my own 1 CA testing. But even Orion's data I ended up factoring out, and now I'm seeing that what remains isn't the whole truth either.
To communicate this, I have to step back and present the bigger picture to you. Either take my word for it that these results are all substantially different, or go and take a look at the data yourself:
1) ykykyk05251's FM21/FM22 tests that purportedly used mass testing using AI doing knockout tournaments over 40 million matches and then pitting it against human picks in 100,000 matches to determine even ideal positional attributes. Supposedly dribbling is more important than pace for ST, stamina is most important for DL/DR, and decisions is 40% weight for every position.
2) Zippo's tests which seem to have been ongoing for a few years and claims that composure is pointless above 8 and concentration + determination combined is less important than stamina.
3) HarvestGreen's data which says that decisions is ~0% benefit and that concentration is more important than stamina (at both 10 > 18 and 10 > 1).
4) Orion's data that different positions favor different things, but stamina doesn't feature in the top 8 for DL/DR.
5) My 1 CA testing which shows you can win the Premier League with a team of 1 CA players with it mostly just being about having high pace/acc/drib/jump, but also some positional quirks like high concentration for DC/DL/DR but unnecessary on other positions.
6) My 15 pace/acc template, where having a bunch of mental/technical stats at ~13 is strictly necessary to come ~4-5th instead of relegated. And I find that Orion's positional adjustments, or even my own ones like concentration on only DC/DL/DR doesn't work here.
Now is everyone lying, or making serious errors here? I don't think it's either. I think that the process changes at different levels of attributes, in ways that are too difficult to predict and require brute force testing to deduce.
A hypothetical example as how this plays out in the match engine:
AML with 18 pace/acc > exceeds DC's 16 acc (speed rate) and 15 agility (direction change speed) > no further calculations, AML 'escapes' and ends up scoring against GK's reflexes
AML with 16 pace/acc > matched by DC's 16 acc and 15 agil > AML 14 balance challenged by DC 12 strength > DC outmuscle failed > AML decision (8) continue dribble (13 proficiency/tendency) > DC decision (13) slide in with both feet (tackling 13, bravery 14, aggression 15, dirtiness 15) > AML escape (14 balance, 18 agility) or AML deposed of ball (14 balance, 9 agility)
From this hypothetical we can see how when pace/acc 18+, a lot of those technicals/mentals like tackling and bravery become pointless, but if you lower the pace/acc to 16 then the reverse starts becoming true. If you lower pace/acc to 10, probably another process plays out. And yet of course exceedingly high pace/acc is always your best bet.
Then of course there are things like 'pooled' and compensatory attributes to further complicate the process, where it seems to be ok if your ST has 1 finishing, so long as someone else in your team has 7 finishing. Stamina seems necessary to preserve low pace/acc, but with 20 pace/acc you have so much to spare that you can get away with a team of 1 stamina (that's how I'm reading stamina's results anyway).
So returning to Genie Scout ratings, for newly promoted team 15 pace/acc with good mentals/technicals is what works, but what also would work if you could get it is 18+ pace/acc team with limited mentals/technicals. So how do you value mentals/technicals? Something inbetween? Then no one wins, as the 16 pace/acc thing I explain. Different set of values for each? Too much work and I think everyone likes just one simple rating set, plus I think one set is feasible. Like a good knife, it needs to be sharp but have some flex to it, so that even if you *can* have those 18 pace/acc players but miscalculate things and end up with a team of duds/injuries/etc then you'll still survive relegation at least.
@GeorgeFloydOverdosed So if you are starting a save with a small club in a small league with little reputation I'm wondering what the best play is. Do you just go for the best players you can get according to current GS rating at the start or would you look for the best potential ratings?
Rain said: @GeorgeFloydOverdosed So if you are starting a save with a small club in a small league with little reputation I'm wondering what the best play is. Do you just go for the best players you can get according to current GS rating at the start or would you look for the best potential ratings?
Use current rating + take into account or filter CA-PA gap. Genie Scout potential rating will give you a rough idea how good they can become, but it will be inaccurate.
You need to first establish your club to the point that you can actually attract high potential players. Plus you aren't really developing anyone in Latvian 2nd division with bad facilities, bad coaches and likely semipro status