Lariato said: How do you edit weights for FM26 roles in that tool? I just get empy non-editable bar
For FM24 (Legacy) roles I do get tables with attributes as expected Expand Yeah I have the same issue, I didn't even realize before because I used FM24 legacy role only. If you can apply the FM24 legacy roles to your import, then it should make no difference.
Kamas1 said: @GeorgeFloydOverdosed so for fm 26 the best way to find meta players is still the file from the first post of this topic? Expand I just tried to test the 15 pace/acc template in FM26 demo. Gave up after I realized it would take all day without the editor.
So I don't have any knowledge on this, but I would guess that the weights for the FM26 scoring system app would work about as well in FM26.
If you look at what HarvestGreen's data shows changed in FM26, the only important one missing from my template is long shots. Finishing is also boosted in FM26. So I would just change the following weights: Finishing '48', Long Shots '34'.
You can also plug these weights into Genie Scout, but there are serious issues with Genie Scout. I would say for FM26 it's these latest weights (in FM26 scoring system tool) > these latest weights in Genie Scout > FM26 first file > Blended file. For FM24 it's these latest weights (in FM26 scoring system tool) > these latest weights in Genie Scout > Blended/Performance file > FM24 first file.
BTW these are the weights for hiddens:
Pressure '49' Professionalism '49'
Injury Proneness and dirtiness '50', but I think I just set these pretty arbitrarily for now. Basically means aiming for no higher than 13 in each.
The rest matter too little for performance, but you may also want to want to give some weight to ambition, loyalty, adaptability (if buying foreign players).
BaZuKa said: So I decided to try the FM26 Player Scoring System and found this striker for just 800k. He went on to score 42 goals in League One, despite being rated only 13.4 in the app. I used the weights mentioned in post #215. Expand Not quite sure if you're praising the weights or saying they're invalid lol. But 13.4 is Doku's rating, and he got 6.9 (above average) in Premier League, so the results for that player seem about right.
Rain said: Using the Scoring System I can't seem to get the ratings on my list to change even when editing and adding my own role. Any idea how to fix? Expand I edited an existing FM24 (legacy) role. Whenever I save it I have to refresh the page in my browser due to some bug where there's a remaining overlay I can't close. Maybe try different browser, or have you missed pressing the 'calculate' button again?
Rain said: I actually bought FMRTE to try it out and you can enter weights like on genie scout for the most part, but there really isn't a good way of looking through all players who are on your list. Usually I will just look through all players willing to join me, but there didn't seem to be a way to do that there it only showed "All Players" or shortlists.
Are you using the same weights for all positions? Expand I've only been looking at the ST position, but you can use the same weights (removing the ones '10' or less) for all other positions except GK. I will look into positional differences more later, but it wouldn't be much different from what I've tested so far.
LightningFlik said: I've just bought a house and after I get myself situated I'll have access to my Windows PC again so I can port the tool I've been developing on my laptop. It'll let you read your team from in-game memory and search the player database based on a number of constraints.
No fee, no weird start-up time.
It'll just take a little bit longer. Expand That sounds great, even if it's further down the track.
I was thinking of asking for your Linux tool before, but I didn't want to put any pressure on you. And then I found 'FM26 Player Scoring System' mostly does the job anyway.
In my searches, I did come across your posts on reddit. I think you said you used the original version 'FM26 Player Scoring System' is derived from but worked out how to read direct from memory.
@LightningFlik You were right about the discrepancy between Genie Scout's ratings and your own calculated ones.
Something is really out of wack with Genie Scout. I don't know how many players it affects, but it's just unacceptable to me.
I realized that according to Genie Scout, Osimhen (91.65%) is better than Haaland (89.86%):
It is also the case with Nunez (88.54%) vs. Haaland. I couldn't find any way in Genie Scout to solve it, but looking around I did find that you can see progress rate for players in the free version by clicking 'compare with' on the player screen, which is useless nowadays but something I didn't know before.
The issue appears to be with the general & positional rating. If I click 'show potential', suddenly Haaland goes from 89% to 99%, even though he has very little room to grow. For other players, their potential rating is ~2% less than their current rating. So I think the way it's calculating things is just bugged, at least when using my custom weights.
Unfortunately there is no proper substitute for Genie Scout. I've been considering FMRTE, but the free version doesn't allow you to save weight changes, and only some people will be willing to pay money for FMRTE.
Thankfully 'FM26 Player Scoring System' is a free alternative that can get the job done. Ignore the version in the name, it can do both FM24 & FM26. The main problem with it is that it doesn't load from memory, so it's a bit less convienant (though Genie Scout takes time to load anyway) and it can't use/view hidden attributes (including CA/PA). It does at least give personality & footedness descriptions to go by, but you can't put it in the weightings. Nonetheless I find it viable and actually quite pleasant to use, and you could use it in conjunction with Genie Scout if need be.
Just follow the instructions to load it up, then input these weights into any role ('add roles' green button > 'edit weightings' small button):
acc 78 agg 42 agi 34 ant 50 bal 46 cnt 50 cmp 44 det 52 dri 50 fin 27 fir 7 hea 4 jum 53 lon 10 otb 2 pac 78 pas 4 sta 35 str 30 vis 27 wor 43
Where values are 10 or less, I added them in because they should have an effect, but I found they are non-essential attributes.
Initially pace/acc was 57, but I adjusted this according to observed data, such as these results someone on youtube got from putting the following at Man City:
goals/matches/rating
Haaland 40/38 7.57 Lewandowski 38/38 7.52 Mbappe 35/36 7.30 Messi 26/38 7.26 Ronaldo 23/33 7.19 Kane 28/38 7.12
I also observed other kinds of players. I couldn't get it 100% accurate, and Kane's rating is oddly low there (the goals scored seem more accurate), but overall I'm pleased with the result. Their scores:
Haaland 16.7 (1st) Lewandowski 15.3 (4th) Mbappe 15.3 (6th) Kane 15.0 (9th) Messi 14.5 (17th) Ronaldo 13.3
What was said about players like Doku still appears to be true. Doku is only 13.4.
Some fascinating things can occur at the edges. For instance this is the best player (7.8) at near zero wage/value:
We can see here he was a dud in Div 5, but got picked up by a Premier League team probably because he was the best young freebie option, before being dumped until eventually ending up being a dud even in Div 8.
The highest rated (12.2) ultra-low salary ($50/wk) option was in Chile and got picked up later by Nottm Forest:
Maybe I could adjust the Genie Scout values until the player lists become closely aligned, so that even though the weights will become non-sensical the results will be accurate, we'll see.
LightningFlik said: @GeorgeFloydOverdosed I've been thinking recently about an experiment to measure the importance of attributes for a given role (e.g. Striker); could you tell me if you think this is worthwhile?
Create a test league (maybe 10 or so teams) with identical coaching staffs (Assistant Manager is probably sufficient, since he'll be the one managing games), and identical players in all positions. That is, all the teams' goalkeepers are identical (copied from a real player), as are their fullbacks, their centrebacks and so on. None of the players ought to be able to play the position you're testing (e.g. Striker).
All teams are set up to play the same formation (if this can't be done in the pre-game editor, it would require ten human managers on holiday, hence identical Assistant Managers). All players have their attributes and conditions frozen.
The experiment is taking real-world strikers (e.g. Haaland, Mbappe, Vini, Goncalo Ramos) and putting multiple copies of them on to different teams, meaning all teams are identical except for their different strikers. If they're all playing the same tactics with the same coaches, the only variation will come from their strikers.
And I don't even think you necessarily need to look at the league table at the end of the experiment to determine which striker is better, since games can be won and lost by other players too. I'd look at the goals scored by the strikers, their assist count and possibly their average match ratings (or other things like key passes, xG, if you don't trust how match ratings are generated).
From here, you would perform a linear regression to assess what attributes / personality values are important in affecting the things we care about. I have to think that if was useful, someone would have done it by now but I can't see what's wrong with it. Expand My view is that doing tests in situ is better than these artificial simulations. For example, if you make every player have flat 10 in everything, then the results would show you that Mbappe does worse than Ramos because both have pace/acc far exceeding 10, but Ramos has better mentals than Mbappe. I don't know if it would actually happen that way, but it's just an example to illustrate what can go wrong. Also if you freeze morale, condition, etc. then that impairs the effect of temperament, natural fitness, and so on. I can think of 100 ways things could wrong, and that's just what we know.. what if there's things we don't know about, such as rock-paper-scissors situations that occur in standard gameplay?
It's like taking this guy and sticking him in the Ukraine war:
Yes, he's the world's best finisher under controlled conditions, but context matters. And the following isn't part of my response to you, but I just want to point out that we don't need to send this guy to Ukraine 1000+ times to have an inkling - clone and send him in 5-10 times, and we'll end up with a pretty solid idea.
If you just test for what actually works in a real league, with standard play conditions, then you can skip all that and yet know what works for certain.
What I'm currently trying to grapple with is how to weight pace/acc vs. the rest of what matters. Neither alone works. Pace/acc is more all-or-nothing, so you can't weight it at '15' like concentration '13' is weighted. Filtering pace/acc at ~14 instead of weighting it works, but then you end up with Ramos being better than Mbappe which isn't really what we want. And it doesn't help that ratings aren't a reliably solid measure of actual performance. And we also have to consider that pace/acc gains through training of +4 are achievable.
As touched on before, we also have to consider that someone like Ramos probably does great because his league (I only assume here) generally has low pace/acc to allow him to thrive. If you stuck him in the Premier league, the situation would switch where high pace/acc is required and therefore dominates over high mentals/technicals. So dealing with apples & oranges discrepancies as well as possible.
As I'm testing to find the right balance, I realize that Genie Scout seems to have something f'd up going on with how it calculates ratings again. Might try FMRTE for now to see if the calculation is different.
LightningFlik said: Does this lad get 90 minutes or come off the bench? Expand Mostly full appearances
Rain said: So what is it in this case that is dragging Doku down? I see he has the 10 temperament and good natural fitness, pace, acceleration. Is it mostly the consistency? Expand
We can see that Doku is lacking in anticipation, concentration, jumping reach, and is a bit lower on a few physicals. Rashford also has significantly better hiddens & personality, but Doku's personality/hiddens aren't particularly bad.
I've found ambition to be non-essential, and while consistency is essential, 'important matches' is curiously more of a 50/50 thing. Sometimes you finish 4th, other times you finish 12th. So I will de-prioritize important matches slightly. Natural fitness is essential, difference between 1 and 14 is 4th to relegated(!). Temperament is essential, but you should aim for ~10, not 1 or 20 as it follows a different rule to other attributes (see Orion's data on this).
Interestingly if I plug in the values I have so far, while leaving out pace/acc, the best ST isn't Haaland or Mbappe, but Goncalo Ramos.
If you go by just pace/acc, this guy is nothing special, and yet he's averaging 7.58 rating. That's less than Mbappe's 7.90 at same club, which means pace/acc still mogs, but low pace/acc shouldn't be automatically discounted.
Here is what we get if we filter for pace/acc 17+:
The 15 pace/acc template that gets ~4th scores ~58.6%, so I take that as a kind of cut-off point. We see Rashford is slightly under that at 58.29%, but then he also has 18 acc, 17 pace. He is also not a ST by default. Rashford of course does well with 7.38 average in the EPL.
Going further down the list we find Largie Ramazani who has 18 acc, 17 pace as Rashford does, and even same ST proficiency (17), but is only 52.74%.
We see here that Southampton thought enough of him to give him a go in the EPL, and then with another team in the EPL, but he just didn't quite cut the mustard in spite his high pace/acc.
Jeremy Doku plays alongside some of the best at Man City, and yet is supposedly a dud 48.12% in spite of 18 acc, 17 pace.
I think even if you spent a while looking at this guy's stats, you'd have to conclude there's something wrong with that low 48.12% rating..
..and yet Doku gets a measly 6.9 average rating before being sold off.
And finally let's take a look at Ricky-Jade Jones (46.63%), who judging by the pace/acc/drib primarily, seems like a very good pickup for a lower league club:
Even if the low rating is disingenuous, the lack of goals and assists is probably a good indication of the truth.
Rain said: Are you only doing FM24 or both? Expand I'll probably end up taking the FM24 version and just make a few guesstimate adjustments for FM26. Focusing on FM24 first in any case.
Busy with life at the moment. Don't want to stress myself out with a deadline, so I'll just say it'd probably take 4 days if I dedicated all my time to it. So hopefully I'll put out a new file about a week from now, but depends how many breaks and tests I need.
Here are some new test results:
Luton 15 pace/acc - 14th, 10th, 8th Luton 15 pace/acc alt tactic - 8th Luton pace/acc max - 14th, 9th, 9th, 9th Luton pace/acc max alt tactic - 13th, 12th
15 pace/acc = template I showed before, with £200k/player budget Alt tactic = different 4-2-4 knap tactic I used (TON 424 HP HUB MC P105 EFC) Pace/acc max = picking for just highest pace/acc players with £3mil/player budget
I found this fruitful. I think the implications are pretty obvious. I'll be doing more alternate tactic testing to get more data on that aspect.
Exellent said: Your Genie Scout ratings seem to be broken or inaccurate for the DM position in your FM24 Blended file. I noticed that a player with 124 CA has a 3% higher GS rating than a player with 168 CA, even though the 168 CA player has strictly better attributes across the board. Furthermore, in the Orion ratings, the second player's rating is 10% higher.
Spoiler Expand I agree that doesn't look right
Rather than fixing the Blended file, I'll be fixing this up in the redo
I had another go with Luton, doing the player selections more seriously this time and no age limit (I used 26 age limit last time). With a £200k budget per player, I got 14th this time
Still couldn't survive relegation with just free transfers only though
I used filters for pace, acc and jump. This seems important, the only wiggle room is probably 14-15 pace/acc, no less, and jump probably needs to be a minimum of ~10 on every player.
LightningFlik said: My favourite tactic. Is there a concern that your low-CA teams actually aren't good enough to finish fourth on their merits but through the strength of the tactics? I don't want to tell you or any of the other hardworking testers your business but it might be more instructive not to use a tactic that could see even the worst team finish in the top half.
Now it makes sense why that team wasn't comparable in terms of rating to the team who finished fourth in my save. Expand Without a doubt, if I didn't use that tactic or something akin to it, the team would have most likely ended up relegated.
Tactic is actually somewhat more important than attributes. If you take default Luton, no buying high pace/acc players or anything, then using that Knap tactic will get them to ~9th.
This isn't something that bothers me though (aside from gameplay perspective), because if you have a park the bus tactic or whatnot you lose almost every game regardless of what players you have. The best defensive strategy in FM is to go on the offensive. "You can have any colour you want, as long as it's black".
debelizec19 said: Dear @GeorgeFloydOverdosed first thank you for dedicated research work. My question is for inaccurate result in GS. When you say it is inaccurate, do you think that potential attributes are incorrect when allocating values to certain attributes? What I'm trying to ask, lets say we have player with current value for pace 13 and acc 13 and in GS we can see that his max value for those attribures are 17. Does it mean that with proper trainig we will achive those max values for pace and acc or it will be even 20 if we will use training for maximize pace and acc. Hope I'm clear with my question Expand Yes, that's precisely what I have in mind. And also it doesn't tell you the chance of reaching their PA, if you don't have paid version.
So it's inaccuracy (progress rate) on top of inaccuracy (potential rating) on top of inaccuracy (current rating).
CA-PA gap is both a very important factor in progress rate and an indicator of how much your player has room to grow. Because pace/acc is most important, and you can redistribute mental/technical into it if you really need to, high PA becomes less important than the starting distribution (current rating) of your attributes in my opinion.
Out of interest, do you and others run these tests using the same tactics? If you're getting conflicting reports of what attributes matter, it might come down to player roles, which is what the game tells us. Expand No, different tactics are used, and I had forgot in my own mind this factor when I was writing all that.
HarvestGreen has given results for different tactics, even comparing and contrasting the top AI tactics. There were quite a few significant differences, but overall I think it was like ~20% difference max and pace/acc were still king.
I doubt that tactics explain even half the difference between all these attribute tests.
I myself always use Knap's EF 424 IF HP V2 P101 AC tactic. My reasoning is that's popular and it's one of the most successful tactics, so why bother tailoring to say a more defensive tactic. I have been considering trying an 'underdog' tactic out for a change for my low pace/acc testing though, to see if it makes it more viable. I'll report later on how that goes.
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.
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.
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.
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).
I want to say a few things about 'cities' and whatnot before I forget them.
Singapore has no set local region, yet it produced newgens born in 'Singapore(city)'. No newgens had birth cities other than 'Singapore', even though there are 54 in the database. All other cities had 0 attraction and 'not set' inhabitants range, suggesting that one or both of these settings are required to make it active. Based on other results, I believe only 'inhabitants range' matters here, and that actually 'not set' doesn't inactivate, it only makes those cities very low probability (this part remains questionable). Note that Singapore clubs have all sorts of different Singaporean cities set as their city.
In Slovenia, no cities have a local region set, however attraction and inhabitants range are. Results showed a diversity of birth cities, in frequency corresponding roughly to inhabitants range. Again there is some doubt over the role of 'attraction'. I ended up being content to rule it out as a factor, but further testing may reveal this is mistaken. I actually suspect that why 'attraction' appears potentially correlated at times, is simply because (a) high inhabitants cities tend to be more attractive, and (b) a large city surrounded by towns is obviously likely to be designated as more 'attractive' than a large city without surrounding towns.
Some tests:
Normal (England) - 49 players born in London. Bristol City 122 PA median.
London '1-1000' inhabitants, all others 'not set' - Bristol City 119.5 PA median, 2nd sample 126 PA. Resulted in much greater diversity of city of birth, but did not prevent 'not set' cities from producing players.
London '20,000,001+' inhabitants, all others '1-1000' - 71 players born in London. Bristol city 121 PA median.
London high pop, no local regions - 64 players born in London. Bristol City 127.5 pa median.
Poland-England switch - Bristol City 101.5 PA median. Man City 120 PA (142.5 normal). Lech 133 PA.
France-England switch - Paris SG 159.5 median.
So you can see here that cities & local regions don't affect PA, and also that the number of players being 'born in' certain cities can be influenced but probably not as much or easily as you'd think. The nation ID, or unique division ID, by contrast does matter significantly.
Further implications:
Since newgen production is probabilistic based on just 'inhabitants range' and is initiated by the club (club creates 16 players, not drawn from some regional 'pool' ), then there is no competing for players going on.
Once you start thinking about it, it becomes obvious why you would expect this all to be the case:
- If newgens were 'drafted' or 'poached' from a regional pool, no doubt the game would at least on rare occasion come up with errors, where some clubs wouldn't get their 16th player while another would end up with 17 and shenanigans of that nature. Yet I have never heard of since an error in all the years of FM. - If the above was happening, it sounds like it would be computationally expensive. It would be the first corner to be cut for game optimization. - The best player production cities in the real world also tend to have the greatest club numbers: Buenos Aires, Argentina; Montevideo, Uruguay; London, England. Aside from making club competition for players within cities a difficult thing to pull off in the game correctly for the programmer, the fact is that in Buenos Aires in the game you have 8 clubs in the Argentine Premier Division including Boca and River Plate. - If 'inhabitants range' and/or 'attraction' were used as an absolute measure for newgen quality and/or number, then Mumbai India (20 attraction, 20000001+ inhabitants) would be a goldmine for newgens, yet it isn't (in the game, and in real life). - Local regions serve a purpose in the editor for setting division boundaries where coordinates do not suffice; i.e. a team located in Wales, may actually play in England in real life. Coords fail here, but 'local region' can put the club's city in the right division.
The other part of this is, where exactly does this leave Youth Recruitment?
Youth recruitment still works as a factor, as a pecking order within a division. In fact, youth recruitment makes more sense now, given that I had known before that there is overlap between divisions (i.e. Premier League 140 > 100 median say, Championship 115 > 80 median) instead of Premier League 140 > 100, Championship 99 > 80.
There are other hints in the database that youth recruitment is just relative (i.e. pecking order) rather than absolute. In Turkey players in the top league have YR of 2 to 8. In Jamaica, 8 clubs in Div 1 have YR 20. Tiny nations can have YR 2-3 (San Marino) to YR 20 (Turks & Caicos Islands). This is neither absolute in terms of quality or number of production nor geographic distance.
The fact that it bears no relation to geography in my view is further evidence that there is no underlying 'draft' or 'pool' system going on, as you would expect 'youth recruitment' to be the extent of the club's tendrils into surrounding cities or coordinates of distance.
So there is nothing so sophisticated going on with Youth Recruitment, but it is drawing the better players to drop into unique divisions from somewhere, and that appears to be the national pool, by which I mean a number of hidden factors & values rather than a hidden pool of players within the game. It's clear that each nation or unique division has at least one hidden quality setting that is very significant (see 'Poland-England switch' above for example).
Kamas1 said: What do you thinks about use FM24 Blended (recommended) in FM26? Expand It would be kind of ok because pace/acc dominate and that didn't change, but a few attributes have changed up to 3x and some of them are moderately important not just minor. It wouldn't be as simple as making the attribute 21 > 63 (3x), as that's not how my calculations work and why I haven't done a FM26 version yet.
So I would stick with the old FM26 file I posted for now.
Alternatively you could look at HarvestGreen's differences for FM26, and give a +10 bonus for each +1x for attributes that have 10 weighting or more. I.e. 3x would be 12 > 32, 1.5x would be 25 > 30. That would make things roughly more appropriate for FM26.
I will eventually create a FM26 version, but it will be with my new method that I'm currently working on.
Rain said: Are there certain positions that are more important than others and if so what are they? Thanks Expand Although I haven't tested this, I think the tactic is more important than the positions. Or put another way, the best tactic determines what positions you require.
But there is definitely also a difference in CA required for positions, or what positions are easier to fill.
Starting with the easiest/cheapest: DM, ST, DL/DR, GK, DC, AML/AMR
In terms of what position can have the largest variation in impact, I would say it's DL/DR (has a lot of cheap yet highly impactful attributes, which you should try and take advantage of), then DC (expensive and rarer attributes but crucial, things like pace/acc, dribbling and jumping reach), then AML/AMR (mainly just needs high pace/acc which is very costly, but also has some cheap beneficial extras such as jumping reach and defensive attributes but these are also hard to find).
The DM requires very little CA, and ST can be whittled down to pace/acc with jump reach, dribbling semi-optional, and finishing almost completely optional. GK doesn't have freebies, is only 20% impact of outfield player according to HarvestGreen, and either already has the required agility/reflexes/aerial reach or they don't.
Is this 20% scale still the case? If so then I think it's not a coincidence that mental/physical/technical/goalkeeper attributes are internally stored as values 1-100 and personality attributes are 1-20.
If Genie Scout doesn't normalise these values then it's multiplying the personality weights against values that are one-fifth those of M/P/T/G attributes. Expand If you clear everything and make 'professionalism' 100 weight, players with 20 professionalism will have 20% rating. Here's someone else who discovered this and he lists the affected attributes. If you don't use it alone in your weights, you'll find it is a negative modifier on the rating.
But if you take the 'blended' ratings and do -21 to professionalism, +21 to pace, the rating will only change from 72.4% to 70.8% say. If you do -14 dribbling, +14 professionalism, it might go from 65.3% to 66.0%. So overall it seems to be working as its supposed to, but imperfectly.
I thought of the feet thing, and position proficiency. The Mbappe-Salah case rules this out specifically, as they are both the same, apart from a slight foot difference that should be a mere fraction of 1%. I would guess that what is going on, if you haven't made a mistake/change in your own calculations (i.e. perhaps you left out something like -27 for injury proneness), is that the personality attributes, injury proneness, natural fitness, etc. are exacerbating each other's minor errors in being combined, creating a large overall error. That's just a guess though.
You may be on to something in saying that it could be to do with attributes being 100 vs personality being 20. Could be why Genie Scout is doing it this complicated way in the first place.
LightningFlik said: Apologies if I've missed this somewhere @GeorgeFloydOverdosed, but at what point (if any) do we start to penalise a player for having reduced competency in a position for which they're being rated (e.g. I think Vini has 16/20 for Striker but is considered to be the second best in the world by your blended metric)?
I will have more dumb questions later as soon as I organise them.
EDIT: I might as well ask this one now, while it's fresh. When you rated players according to various weights and formulae in this post, where did the player data come from?
I ask because I tried recreating the calculations manually (using, for example, the Genie weights you attached via screenshot earlier) and got a completely different order of players when looking for the best strikers in the world.
I created a new game in FM 24, July 3rd 2023, and got the following when calculating using your Fast Striker weights:
1. Erling Haaland 2. Mohamed Salah 3. Kylian Mbappé 4. Lautaro Martínez 5. Marcus Rashford 6. Victor Osimhen 7. Robert Lewandowski 8. Lionel Messi 9. Randal Kolo Muani 10. Romelu Lukaku
I did get it working when using HarvestGreen's new GK weights but either we're using different data for strikers or I'm incorrectly guessing how Genie Scout calculates ratings via weights (it is the sum of all attributes multiplied by their respective weights, right?) Expand So it turned out the Genie Scout thing was mostly a false alarm
If you isolate personality/negative attributes, it does weird things, but in standard use +21 professionalism is in fact equivalent to +21 pace or +21 tackling. It's not 100% exact though, it can change the results ~2-3%, probably because it's doing this unusual non-linear thing.
It is concerning though that you find Salah to be ahead of Mbappe. That's a 7.5% difference. Not sure what's going on with that.
For FM24 (Legacy) roles I do get tables with attributes as expected
Yeah I have the same issue, I didn't even realize before because I used FM24 legacy role only. If you can apply the FM24 legacy roles to your import, then it should make no difference.
I just tried to test the 15 pace/acc template in FM26 demo. Gave up after I realized it would take all day without the editor.
So I don't have any knowledge on this, but I would guess that the weights for the FM26 scoring system app would work about as well in FM26.
If you look at what HarvestGreen's data shows changed in FM26, the only important one missing from my template is long shots. Finishing is also boosted in FM26. So I would just change the following weights: Finishing '48', Long Shots '34'.
You can also plug these weights into Genie Scout, but there are serious issues with Genie Scout. I would say for FM26 it's these latest weights (in FM26 scoring system tool) > these latest weights in Genie Scout > FM26 first file > Blended file. For FM24 it's these latest weights (in FM26 scoring system tool) > these latest weights in Genie Scout > Blended/Performance file > FM24 first file.
BTW these are the weights for hiddens:
Pressure '49'
Professionalism '49'
Injury Proneness and dirtiness '50', but I think I just set these pretty arbitrarily for now. Basically means aiming for no higher than 13 in each.
The rest matter too little for performance, but you may also want to want to give some weight to ambition, loyalty, adaptability (if buying foreign players).
BaZuKa said: So I decided to try the FM26 Player Scoring System and found this striker for just 800k. He went on to score 42 goals in League One, despite being rated only 13.4 in the app. I used the weights mentioned in post #215.
Not quite sure if you're praising the weights or saying they're invalid lol. But 13.4 is Doku's rating, and he got 6.9 (above average) in Premier League, so the results for that player seem about right.
Rain said: Using the Scoring System I can't seem to get the ratings on my list to change even when editing and adding my own role. Any idea how to fix?
I edited an existing FM24 (legacy) role. Whenever I save it I have to refresh the page in my browser due to some bug where there's a remaining overlay I can't close. Maybe try different browser, or have you missed pressing the 'calculate' button again?
Also this calculator that has been posted here is similar
https://fm-arena.com/thread/15561-player-attribute-calculator-aggregator/
Are you using the same weights for all positions?
I've only been looking at the ST position, but you can use the same weights (removing the ones '10' or less) for all other positions except GK. I will look into positional differences more later, but it wouldn't be much different from what I've tested so far.
LightningFlik said: I've just bought a house and after I get myself situated I'll have access to my Windows PC again so I can port the tool I've been developing on my laptop. It'll let you read your team from in-game memory and search the player database based on a number of constraints.
No fee, no weird start-up time.
It'll just take a little bit longer.
That sounds great, even if it's further down the track.
I was thinking of asking for your Linux tool before, but I didn't want to put any pressure on you. And then I found 'FM26 Player Scoring System' mostly does the job anyway.
In my searches, I did come across your posts on reddit. I think you said you used the original version 'FM26 Player Scoring System' is derived from but worked out how to read direct from memory.
Something is really out of wack with Genie Scout. I don't know how many players it affects, but it's just unacceptable to me.
I realized that according to Genie Scout, Osimhen (91.65%) is better than Haaland (89.86%):
It is also the case with Nunez (88.54%) vs. Haaland. I couldn't find any way in Genie Scout to solve it, but looking around I did find that you can see progress rate for players in the free version by clicking 'compare with' on the player screen, which is useless nowadays but something I didn't know before.
The issue appears to be with the general & positional rating. If I click 'show potential', suddenly Haaland goes from 89% to 99%, even though he has very little room to grow. For other players, their potential rating is ~2% less than their current rating. So I think the way it's calculating things is just bugged, at least when using my custom weights.
Unfortunately there is no proper substitute for Genie Scout. I've been considering FMRTE, but the free version doesn't allow you to save weight changes, and only some people will be willing to pay money for FMRTE.
Thankfully 'FM26 Player Scoring System' is a free alternative that can get the job done. Ignore the version in the name, it can do both FM24 & FM26. The main problem with it is that it doesn't load from memory, so it's a bit less convienant (though Genie Scout takes time to load anyway) and it can't use/view hidden attributes (including CA/PA). It does at least give personality & footedness descriptions to go by, but you can't put it in the weightings. Nonetheless I find it viable and actually quite pleasant to use, and you could use it in conjunction with Genie Scout if need be.
Just follow the instructions to load it up, then input these weights into any role ('add roles' green button > 'edit weightings' small button):
acc 78
agg 42
agi 34
ant 50
bal 46
cnt 50
cmp 44
det 52
dri 50
fin 27
fir 7
hea 4
jum 53
lon 10
otb 2
pac 78
pas 4
sta 35
str 30
vis 27
wor 43
Where values are 10 or less, I added them in because they should have an effect, but I found they are non-essential attributes.
Initially pace/acc was 57, but I adjusted this according to observed data, such as these results someone on youtube got from putting the following at Man City:
goals/matches/rating
Haaland 40/38 7.57
Lewandowski 38/38 7.52
Mbappe 35/36 7.30
Messi 26/38 7.26
Ronaldo 23/33 7.19
Kane 28/38 7.12
I also observed other kinds of players. I couldn't get it 100% accurate, and Kane's rating is oddly low there (the goals scored seem more accurate), but overall I'm pleased with the result. Their scores:
Haaland 16.7 (1st)
Lewandowski 15.3 (4th)
Mbappe 15.3 (6th)
Kane 15.0 (9th)
Messi 14.5 (17th)
Ronaldo 13.3
What was said about players like Doku still appears to be true. Doku is only 13.4.
Some fascinating things can occur at the edges. For instance this is the best player (7.8) at near zero wage/value:
We can see here he was a dud in Div 5, but got picked up by a Premier League team probably because he was the best young freebie option, before being dumped until eventually ending up being a dud even in Div 8.
The highest rated (12.2) ultra-low salary ($50/wk) option was in Chile and got picked up later by Nottm Forest:
Maybe I could adjust the Genie Scout values until the player lists become closely aligned, so that even though the weights will become non-sensical the results will be accurate, we'll see.
Create a test league (maybe 10 or so teams) with identical coaching staffs (Assistant Manager is probably sufficient, since he'll be the one managing games), and identical players in all positions. That is, all the teams' goalkeepers are identical (copied from a real player), as are their fullbacks, their centrebacks and so on. None of the players ought to be able to play the position you're testing (e.g. Striker).
All teams are set up to play the same formation (if this can't be done in the pre-game editor, it would require ten human managers on holiday, hence identical Assistant Managers). All players have their attributes and conditions frozen.
The experiment is taking real-world strikers (e.g. Haaland, Mbappe, Vini, Goncalo Ramos) and putting multiple copies of them on to different teams, meaning all teams are identical except for their different strikers. If they're all playing the same tactics with the same coaches, the only variation will come from their strikers.
And I don't even think you necessarily need to look at the league table at the end of the experiment to determine which striker is better, since games can be won and lost by other players too. I'd look at the goals scored by the strikers, their assist count and possibly their average match ratings (or other things like key passes, xG, if you don't trust how match ratings are generated).
From here, you would perform a linear regression to assess what attributes / personality values are important in affecting the things we care about. I have to think that if was useful, someone would have done it by now but I can't see what's wrong with it.
My view is that doing tests in situ is better than these artificial simulations. For example, if you make every player have flat 10 in everything, then the results would show you that Mbappe does worse than Ramos because both have pace/acc far exceeding 10, but Ramos has better mentals than Mbappe. I don't know if it would actually happen that way, but it's just an example to illustrate what can go wrong. Also if you freeze morale, condition, etc. then that impairs the effect of temperament, natural fitness, and so on. I can think of 100 ways things could wrong, and that's just what we know.. what if there's things we don't know about, such as rock-paper-scissors situations that occur in standard gameplay?
It's like taking this guy and sticking him in the Ukraine war:
Yes, he's the world's best finisher under controlled conditions, but context matters. And the following isn't part of my response to you, but I just want to point out that we don't need to send this guy to Ukraine 1000+ times to have an inkling - clone and send him in 5-10 times, and we'll end up with a pretty solid idea.
If you just test for what actually works in a real league, with standard play conditions, then you can skip all that and yet know what works for certain.
What I'm currently trying to grapple with is how to weight pace/acc vs. the rest of what matters. Neither alone works. Pace/acc is more all-or-nothing, so you can't weight it at '15' like concentration '13' is weighted. Filtering pace/acc at ~14 instead of weighting it works, but then you end up with Ramos being better than Mbappe which isn't really what we want. And it doesn't help that ratings aren't a reliably solid measure of actual performance. And we also have to consider that pace/acc gains through training of +4 are achievable.
As touched on before, we also have to consider that someone like Ramos probably does great because his league (I only assume here) generally has low pace/acc to allow him to thrive. If you stuck him in the Premier league, the situation would switch where high pace/acc is required and therefore dominates over high mentals/technicals. So dealing with apples & oranges discrepancies as well as possible.
As I'm testing to find the right balance, I realize that Genie Scout seems to have something f'd up going on with how it calculates ratings again. Might try FMRTE for now to see if the calculation is different.
Mostly full appearances
Rain said: So what is it in this case that is dragging Doku down? I see he has the 10 temperament and good natural fitness, pace, acceleration. Is it mostly the consistency?
We can see that Doku is lacking in anticipation, concentration, jumping reach, and is a bit lower on a few physicals. Rashford also has significantly better hiddens & personality, but Doku's personality/hiddens aren't particularly bad.
Interestingly if I plug in the values I have so far, while leaving out pace/acc, the best ST isn't Haaland or Mbappe, but Goncalo Ramos.
If you go by just pace/acc, this guy is nothing special, and yet he's averaging 7.58 rating. That's less than Mbappe's 7.90 at same club, which means pace/acc still mogs, but low pace/acc shouldn't be automatically discounted.
Here is what we get if we filter for pace/acc 17+:
The 15 pace/acc template that gets ~4th scores ~58.6%, so I take that as a kind of cut-off point. We see Rashford is slightly under that at 58.29%, but then he also has 18 acc, 17 pace. He is also not a ST by default. Rashford of course does well with 7.38 average in the EPL.
Going further down the list we find Largie Ramazani who has 18 acc, 17 pace as Rashford does, and even same ST proficiency (17), but is only 52.74%.
We see here that Southampton thought enough of him to give him a go in the EPL, and then with another team in the EPL, but he just didn't quite cut the mustard in spite his high pace/acc.
Jeremy Doku plays alongside some of the best at Man City, and yet is supposedly a dud 48.12% in spite of 18 acc, 17 pace.
I think even if you spent a while looking at this guy's stats, you'd have to conclude there's something wrong with that low 48.12% rating..
..and yet Doku gets a measly 6.9 average rating before being sold off.
And finally let's take a look at Ricky-Jade Jones (46.63%), who judging by the pace/acc/drib primarily, seems like a very good pickup for a lower league club:
Even if the low rating is disingenuous, the lack of goals and assists is probably a good indication of the truth.
Rain said: Are you only doing FM24 or both?
I'll probably end up taking the FM24 version and just make a few guesstimate adjustments for FM26. Focusing on FM24 first in any case.
Here are some new test results:
Luton 15 pace/acc - 14th, 10th, 8th
Luton 15 pace/acc alt tactic - 8th
Luton pace/acc max - 14th, 9th, 9th, 9th
Luton pace/acc max alt tactic - 13th, 12th
15 pace/acc = template I showed before, with £200k/player budget
Alt tactic = different 4-2-4 knap tactic I used (TON 424 HP HUB MC P105 EFC)
Pace/acc max = picking for just highest pace/acc players with £3mil/player budget
I found this fruitful. I think the implications are pretty obvious. I'll be doing more alternate tactic testing to get more data on that aspect.
I agree that doesn't look right
Rather than fixing the Blended file, I'll be fixing this up in the redo
Still couldn't survive relegation with just free transfers only though
I used filters for pace, acc and jump. This seems important, the only wiggle room is probably 14-15 pace/acc, no less, and jump probably needs to be a minimum of ~10 on every player.
Example of those that achieved 14th:
Now it makes sense why that team wasn't comparable in terms of rating to the team who finished fourth in my save.
Without a doubt, if I didn't use that tactic or something akin to it, the team would have most likely ended up relegated.
Tactic is actually somewhat more important than attributes. If you take default Luton, no buying high pace/acc players or anything, then using that Knap tactic will get them to ~9th.
This isn't something that bothers me though (aside from gameplay perspective), because if you have a park the bus tactic or whatnot you lose almost every game regardless of what players you have. The best defensive strategy in FM is to go on the offensive. "You can have any colour you want, as long as it's black".
My question is for inaccurate result in GS. When you say it is inaccurate, do you think that potential attributes are incorrect when allocating values to certain attributes? What I'm trying to ask, lets say we have player with current value for pace 13 and acc 13 and in GS we can see that his max value for those attribures are 17. Does it mean that with proper trainig we will achive those max values for pace and acc or it will be even 20 if we will use training for maximize pace and acc. Hope I'm clear with my question
Yes, that's precisely what I have in mind. And also it doesn't tell you the chance of reaching their PA, if you don't have paid version.
So it's inaccuracy (progress rate) on top of inaccuracy (potential rating) on top of inaccuracy (current rating).
CA-PA gap is both a very important factor in progress rate and an indicator of how much your player has room to grow. Because pace/acc is most important, and you can redistribute mental/technical into it if you really need to, high PA becomes less important than the starting distribution (current rating) of your attributes in my opinion.
LightningFlik said: Messy, messy, messy.
Out of interest, do you and others run these tests using the same tactics? If you're getting conflicting reports of what attributes matter, it might come down to player roles, which is what the game tells us.
No, different tactics are used, and I had forgot in my own mind this factor when I was writing all that.
HarvestGreen has given results for different tactics, even comparing and contrasting the top AI tactics. There were quite a few significant differences, but overall I think it was like ~20% difference max and pace/acc were still king.
I doubt that tactics explain even half the difference between all these attribute tests.
I myself always use Knap's EF 424 IF HP V2 P101 AC tactic. My reasoning is that's popular and it's one of the most successful tactics, so why bother tailoring to say a more defensive tactic. I have been considering trying an 'underdog' tactic out for a change for my low pace/acc testing though, to see if it makes it more viable. I'll report later on how that goes.
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.
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.
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.
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).
Singapore has no set local region, yet it produced newgens born in 'Singapore(city)'. No newgens had birth cities other than 'Singapore', even though there are 54 in the database. All other cities had 0 attraction and 'not set' inhabitants range, suggesting that one or both of these settings are required to make it active. Based on other results, I believe only 'inhabitants range' matters here, and that actually 'not set' doesn't inactivate, it only makes those cities very low probability (this part remains questionable). Note that Singapore clubs have all sorts of different Singaporean cities set as their city.
In Slovenia, no cities have a local region set, however attraction and inhabitants range are. Results showed a diversity of birth cities, in frequency corresponding roughly to inhabitants range. Again there is some doubt over the role of 'attraction'. I ended up being content to rule it out as a factor, but further testing may reveal this is mistaken. I actually suspect that why 'attraction' appears potentially correlated at times, is simply because (a) high inhabitants cities tend to be more attractive, and (b) a large city surrounded by towns is obviously likely to be designated as more 'attractive' than a large city without surrounding towns.
Some tests:
Normal (England) - 49 players born in London. Bristol City 122 PA median.
London '1-1000' inhabitants, all others 'not set' - Bristol City 119.5 PA median, 2nd sample 126 PA. Resulted in much greater diversity of city of birth, but did not prevent 'not set' cities from producing players.
London '20,000,001+' inhabitants, all others '1-1000' - 71 players born in London. Bristol city 121 PA median.
London high pop, no local regions - 64 players born in London. Bristol City 127.5 pa median.
Poland-England switch - Bristol City 101.5 PA median. Man City 120 PA (142.5 normal). Lech 133 PA.
France-England switch - Paris SG 159.5 median.
So you can see here that cities & local regions don't affect PA, and also that the number of players being 'born in' certain cities can be influenced but probably not as much or easily as you'd think. The nation ID, or unique division ID, by contrast does matter significantly.
Further implications:
Since newgen production is probabilistic based on just 'inhabitants range' and is initiated by the club (club creates 16 players, not drawn from some regional 'pool' ), then there is no competing for players going on.
Once you start thinking about it, it becomes obvious why you would expect this all to be the case:
- If newgens were 'drafted' or 'poached' from a regional pool, no doubt the game would at least on rare occasion come up with errors, where some clubs wouldn't get their 16th player while another would end up with 17 and shenanigans of that nature. Yet I have never heard of since an error in all the years of FM.
- If the above was happening, it sounds like it would be computationally expensive. It would be the first corner to be cut for game optimization.
- The best player production cities in the real world also tend to have the greatest club numbers: Buenos Aires, Argentina; Montevideo, Uruguay; London, England. Aside from making club competition for players within cities a difficult thing to pull off in the game correctly for the programmer, the fact is that in Buenos Aires in the game you have 8 clubs in the Argentine Premier Division including Boca and River Plate.
- If 'inhabitants range' and/or 'attraction' were used as an absolute measure for newgen quality and/or number, then Mumbai India (20 attraction, 20000001+ inhabitants) would be a goldmine for newgens, yet it isn't (in the game, and in real life).
- Local regions serve a purpose in the editor for setting division boundaries where coordinates do not suffice; i.e. a team located in Wales, may actually play in England in real life. Coords fail here, but 'local region' can put the club's city in the right division.
The other part of this is, where exactly does this leave Youth Recruitment?
Youth recruitment still works as a factor, as a pecking order within a division. In fact, youth recruitment makes more sense now, given that I had known before that there is overlap between divisions (i.e. Premier League 140 > 100 median say, Championship 115 > 80 median) instead of Premier League 140 > 100, Championship 99 > 80.
There are other hints in the database that youth recruitment is just relative (i.e. pecking order) rather than absolute. In Turkey players in the top league have YR of 2 to 8. In Jamaica, 8 clubs in Div 1 have YR 20. Tiny nations can have YR 2-3 (San Marino) to YR 20 (Turks & Caicos Islands). This is neither absolute in terms of quality or number of production nor geographic distance.
The fact that it bears no relation to geography in my view is further evidence that there is no underlying 'draft' or 'pool' system going on, as you would expect 'youth recruitment' to be the extent of the club's tendrils into surrounding cities or coordinates of distance.
So there is nothing so sophisticated going on with Youth Recruitment, but it is drawing the better players to drop into unique divisions from somewhere, and that appears to be the national pool, by which I mean a number of hidden factors & values rather than a hidden pool of players within the game. It's clear that each nation or unique division has at least one hidden quality setting that is very significant (see 'Poland-England switch' above for example).
It would be kind of ok because pace/acc dominate and that didn't change, but a few attributes have changed up to 3x and some of them are moderately important not just minor. It wouldn't be as simple as making the attribute 21 > 63 (3x), as that's not how my calculations work and why I haven't done a FM26 version yet.
So I would stick with the old FM26 file I posted for now.
Alternatively you could look at HarvestGreen's differences for FM26, and give a +10 bonus for each +1x for attributes that have 10 weighting or more. I.e. 3x would be 12 > 32, 1.5x would be 25 > 30. That would make things roughly more appropriate for FM26.
I will eventually create a FM26 version, but it will be with my new method that I'm currently working on.
Rain said: Are there certain positions that are more important than others and if so what are they? Thanks
Although I haven't tested this, I think the tactic is more important than the positions. Or put another way, the best tactic determines what positions you require.
But there is definitely also a difference in CA required for positions, or what positions are easier to fill.
Starting with the easiest/cheapest: DM, ST, DL/DR, GK, DC, AML/AMR
In terms of what position can have the largest variation in impact, I would say it's DL/DR (has a lot of cheap yet highly impactful attributes, which you should try and take advantage of), then DC (expensive and rarer attributes but crucial, things like pace/acc, dribbling and jumping reach), then AML/AMR (mainly just needs high pace/acc which is very costly, but also has some cheap beneficial extras such as jumping reach and defensive attributes but these are also hard to find).
The DM requires very little CA, and ST can be whittled down to pace/acc with jump reach, dribbling semi-optional, and finishing almost completely optional. GK doesn't have freebies, is only 20% impact of outfield player according to HarvestGreen, and either already has the required agility/reflexes/aerial reach or they don't.
Is this 20% scale still the case? If so then I think it's not a coincidence that mental/physical/technical/goalkeeper attributes are internally stored as values 1-100 and personality attributes are 1-20.
If Genie Scout doesn't normalise these values then it's multiplying the personality weights against values that are one-fifth those of M/P/T/G attributes.
If you clear everything and make 'professionalism' 100 weight, players with 20 professionalism will have 20% rating. Here's someone else who discovered this and he lists the affected attributes. If you don't use it alone in your weights, you'll find it is a negative modifier on the rating.
But if you take the 'blended' ratings and do -21 to professionalism, +21 to pace, the rating will only change from 72.4% to 70.8% say. If you do -14 dribbling, +14 professionalism, it might go from 65.3% to 66.0%. So overall it seems to be working as its supposed to, but imperfectly.
I thought of the feet thing, and position proficiency. The Mbappe-Salah case rules this out specifically, as they are both the same, apart from a slight foot difference that should be a mere fraction of 1%. I would guess that what is going on, if you haven't made a mistake/change in your own calculations (i.e. perhaps you left out something like -27 for injury proneness), is that the personality attributes, injury proneness, natural fitness, etc. are exacerbating each other's minor errors in being combined, creating a large overall error. That's just a guess though.
You may be on to something in saying that it could be to do with attributes being 100 vs personality being 20. Could be why Genie Scout is doing it this complicated way in the first place.
I will have more dumb questions later as soon as I organise them.
EDIT: I might as well ask this one now, while it's fresh. When you rated players according to various weights and formulae in this post, where did the player data come from?
I ask because I tried recreating the calculations manually (using, for example, the Genie weights you attached via screenshot earlier) and got a completely different order of players when looking for the best strikers in the world.
I created a new game in FM 24, July 3rd 2023, and got the following when calculating using your Fast Striker weights:
1. Erling Haaland
2. Mohamed Salah
3. Kylian Mbappé
4. Lautaro Martínez
5. Marcus Rashford
6. Victor Osimhen
7. Robert Lewandowski
8. Lionel Messi
9. Randal Kolo Muani
10. Romelu Lukaku
I did get it working when using HarvestGreen's new GK weights but either we're using different data for strikers or I'm incorrectly guessing how Genie Scout calculates ratings via weights (it is the sum of all attributes multiplied by their respective weights, right?)
So it turned out the Genie Scout thing was mostly a false alarm
If you isolate personality/negative attributes, it does weird things, but in standard use +21 professionalism is in fact equivalent to +21 pace or +21 tackling. It's not 100% exact though, it can change the results ~2-3%, probably because it's doing this unusual non-linear thing.
It is concerning though that you find Salah to be ahead of Mbappe. That's a 7.5% difference. Not sure what's going on with that.