A fair chunk of my earlier post on tactics is in error, as I changed position proficiency at some point for my meta team and forgot to change them back.. this only affects the statements about meta team vs. real Luton/Man City results. Everything else remains valid. I'll conduct fresh tests on the meta team to sort it out.
I've come up with my own variation of the 4-2-4 formation/Knap tactic that is quite unique I believe, based on a theory I have. Initial results are 5th, 7th, 10th, 3rd as Luton and 6th, 7th as meta team after I've fixed it back up. So possibly equal to the Knap tactic, and perhaps a tweak or two can push it higher. I figure that I would probably only release a tactic if I can get 1st as Luton.
Does anyone know if 1st as Luton has been achieved with a tactic test? Obviously just the default test run, no managing the team throughout the season to achieve it.
LightningFlik said: @GeorgeFloydOverdosed Are you in the market for an FM 24 scouting tool with customisable weights? or do you have one already? Mine is nearly done but there's been so many released recently, I can't even keep track. Expand Yes, I'm still keen to see your go at it. The more options the better.
I guess with FM27 coming soon, its 50/50 whether people will move on entirely from FM24 or not. But if FM27 turns out to be crap 2.0, then I think these new tools will help keep FM24 alive by injecting a bit of novelty into it.
I've been making my first foray into tactic creation. Unfortunately I haven't been able to top Knap's EF 424 IF HP V2 P101 AC, but it did bring out some things worth mentioning. I know most of us are waiting now for FM27, and this stuff should apply to FM27 as well.
I started testing with my meta team + Sterling that always seemed to finish a rock solid ~6th.
It didn't take me long to come up with an unusual tactic that finished 5th. However when I tested this with real Luton, they only finished 8th-16th, whereas the Knap tactic can finished as high as 2nd (2nd-10th?). My tactic won with Man City, but only with 89 pts (Knap as high as ~100-110?).
And it's not just my tactic. 424 Mountain King did better than the Knap tactic with my test team from memory, but I didn't get 2nd with it for real Luton.
I tested a bunch more Knap tactics. The results in his spreadsheet don't line up with my results, whereas by contrast it seems that FM Arena's testing is accurate in spite of the methodology critiques I would have of it. Nonetheless all the ones I tried did well, at least 9th for real Luton and often better. But nothing got 2nd like EF 424 IF HP V2 P101 AC.
Of the alternatives I tested, FM24.4HGF4231 V3X RM P107 ALL CUPS seemed the best. 3rd, 6th Luton. Man City 1st 94pts.
Curiously I could use my unusual tactic with several positions I had no player proficiency in at all across my meta team, and yet it still did well. Not quite as good, but it held up well.
Now here's the point we can take beyond FM24:
The results suggest that roles or player instructions are in fact dependent on the attributes of players in some respect.
Yet we also know from the meta team results that it's not what the roles highlight, nor combined team quality.
Injuries, differing player choices, etc. can account for the high variation in the real Luton results relative to the meta team test, but it doesn't explain why only the Knap tactic peaks far higher with real Luton.
One idea I have is that perhaps its not combined team quality, but how each player's attribute is relative to others in his team. In the official guide, they talk of 'absolute' vs 'relative' attributes, and most attributes are 'relative'; perhaps this is what they were actually talking about. There is some suggestive evidence that makes me think along these lines, such as that with finishing, you only need a player with ~7 finishing, but any figure can do the job really, and what happens is whoever in the team has the highest finishing is the one scores the most goals, even if they're a DL with 5 finishing. So we see here that a low attribute altered the tactical outcomes, because of how it was relative to others in the team.
So for instance in my meta team all my players have 8 passing and 13 dribbling, the tactic treats them as all equally likely distribute the task of passing and dribbling I suppose. Perhaps an attribute such as 'teamwork' even influences how well these tasks are evenly distributed throughout the team. But then also for each individual player, they would be more likely to dribble (13) than pass (8).
It's pure speculation, but it would explain why virtually every technical attribute matters zero percent. It would be that the tactic minimizes their influence, and that tactics which don't minimize their influence get beat anyway, whereas Knap's tactic wins almost every match. In other words it could be that just the tactics are broken, rather than that the attributes are also broken. 0 direct passes x 20 passing = effectively 0 passing. Or, 20 direct passes x 20 passing x 0% win rate = 0% win rate for 20 passing.
Torresinho said: what about playing time for young players ? Expand 18 or less = Don't need competitive matches, but do play in friendlies (u18s/reserves fine) mainly to keep their match sharpness up in order to avoid injury more than development 21+ = Need at least 15 full matches (1350 minutes) per season to develop adequately, 25-35 is ideal, so this is why loans typically become necessary 19-20 = Inbetween the two above
Don't bother bringing on young players for a few minutes at the end of a game. The game counts experience in terms of minutes not appearances, and I suspect it could also be counting it the same way it counts minutes for match sharpness, which is that it counts in precise 11 minute blocks (I think it was 11 minutes.. something near that anyway).
alexej said: Sorry if this has already been said but are these all at x2 training intensity? Expand I don't know what HarvestGreen tested them with, but he recommends 0-0-0-2x-2x, and I've played around with it myself before and found that to be best too.
So here I am trying to take into account outfield + GK + CA efficiency + intensity.
HarvestGreen didn't include the GK stats for 331 as far as I can see, so I've just inserted a high value of 1000. Even with this high value for GK, it doesn't seem as quite as good overall as 214 or 211, though it does have the highest absolute gain in its favor.
There is one more thing to add into the assessment, which is the effect on other various in-game factors:
331 has a slight edge over 214 due to the better balancing of match sharpness.
211 is much better for match sharpness, but is probably going to wear down condition too much during the competitive season, especially if you have weeks with 2 matches. The main problem with 211 is that it would be so finicky to implement that in all likelihood you're not going to implement it properly every week, and the careful handling of condition would add to this. 211 is nonetheless very good for GK and absolute gain, both of which 214 falls a bit short on.
One more thing to say about 214 is that its what I found to be best using a different method previously. So where things are roughly equal, I'm inclined to lean towards it anyway. And in HarvestGreen's assessment, 214 also scores higher on his 'quality' index than these other 2.
So overall I will still be recommending 214. I just realized that 331 was actually my 2nd recommendation before, I didn't realize until now that that was HarvestGreen's recommendation.
It might be worth setting up 211 for your u21s/u18s, especially if they have no competitive league or cup and you can set and forget it with weekly friendlies for the whole season. Reason is, u21s/u18s easily end up in match sharpness ruts they never get out of, and low match sharpness is a big influence on injury rate, and more injuries = less development.
It's splitting hairs between 214 and 331, so if you feel more comfortable going with 331, that's perfectly fine.
One thing left unaddressed is GK focus. HarvestGreen's data used 'Agility and Balance' as GK focus, so we can't be sure of the effect of any others. But it seems to be the case that Agility isn't as critical as a few others: Determination, Reflexes, Concentration. Personally I've started favoring 'GK reactions' because it focuses on two of those (Reflexes, Concentration). But there's no data on this.
211 = 130% GK, 100% Def, 90% Att 214 = 100% GK, 100% Def, 100% Att 260 = 105% GK, 105% Def, 105% Att 142 = 125% GK, 125% Def, 125% Att
..and now, as I was examining HarvestGreen's own rankings, I gradually cottoned to the fact that there's a huge standout error in my data.
214, 211, 142 and 260 are all at the top of his rankings just like mine.. yet he has 331 as the best. I figured there has to be something amiss here, and sure enough, getting ChatGPT to recalculate it gives 1202.567.
That means HarvestGreen's recommendation of 331, would in fact be the best schedule for outfield.
I could have sworn I verified it with ChatGPT as 1020.63 three times just to be sure at the time. So I went back to check, since I still had the convo open:
Fuckin' ChatGPT man..
I presume it got at least most of them right, but now I can't be certain some weren't miscalculated by ChatGPT.
Anyway, at least we know from the backing of HarvestGreen's rankings that 331, 260, 211, 214 and 142 are the top ones and that the numbers are most likely correct for at least those.
You should be aware also this mirroring of HarvestGreen's data is a pleasant unexpected surprise. I did have HarvestGreen's weights as a starting or reference point for Premier League 2.1, but as you can see with Rain's comparison test of HarvestGreen's weights vs. Premier League 2.1, things should have changed substantially in the end. So it's surprising that the favored training schedules nonetheless align.
bf3metro said: If you had to pick only one between 214 and 260 for the entire squad, which one would you use? Expand I'll get back to you in a day or two, as things are still evolving right now, as you can see from my words above
I've done most of the remaining training schedules in HarvestGreen's spreadsheet.
We have a new top one, that I've highlighted green. This one was actually suggested by tam1236 back in january. He took the HarvestGreen attribute test findings and did a calculation on it to find this as the best training schedule.
It even has pretty good efficiency, though I suspect the workload could be too high to be worth using with all those modules. Very impressive though given it also has one of the highest CA gains (37.75), so it's particularly good if you want to keep your players looking well rounded for aesthetic reasons without compromising on actual performance.
It turns out there are a lot of very high efficiency options above 1000 (so still ~85%+ of the absolute best), and the most efficient is now Recoveryx13. I wouldn't recommend it, but it's interesting there's a profound difference in training quality between rest and recovery on their own.
I've highlighted in yellow some schedules that stand out to me, but I haven't taken a close look yet.
masterjackfruit said: I personally like the training schedules with recovery the most, I try to keep my backup players sharpness as high as possible and recovery seems to preserve sharpness better than rest. Also I notice that for some reason the Attacking session yields an unusual number of injuries, more so than Defending or Overall. I could be wrong, but it's something I've noticed in my saves. Expand It's a more subjective matter, but what I found is that using home friendly matches against the worst team possible is better than using recovery periods in training.
Friendly matches improve morale, build familiarity, build up match sharpness a lot better, and I daresay even result in less injuries.
When using all recovery I noticed that it tends to result in too low condition and eventually fatigue, so it had to be half recovery half rest, which also can be tedious to set up properly each week.
I can't remember if this is the case, but recovery might also interact differently with the other training modules to produce different results than what it shown in HarvestGreen's spreadsheets.
bf3metro said: Hey, based on all the testing you’ve done here, do you have what you’d consider an optimal schedule to actually use week-to-week over a full season?
I’ve been using ZaZ’s Growth 2, but in my save it really isn’t delivering the development I expected.
What I’m mainly looking for is a schedule that pushes CA growth as fast as possible so players reach their PA quickly, rather than just maximizing CA efficiency.
Would you recommend switching fully to something like 260 / 214 / 156 depending on the players’ CA–PA gap, or do you have a specific weekly/monthly rotation you think is optimal in practice? Expand ZaZ Growth 2 would be 1064.2675-1115.573
Oddly enough the 2-match Physicalx2 Attackingx2 scores better than the Physicalx5 Attackingx5 and results in higher CA gain too.
Because CA-PA gap is a critical accelerant to growth, you actually want the lowest overall CA growth possible - while highest in the CA growth where it matters. There is a hard cap on CA per season anyway, which is ~20-25 (someone has reported ~30 i think). If your players aren't developing much, it's more likely that they're over 18 and not getting enough match experience. No match experience for older players can reduce growth to ~1 CA/season.
I've re-examined training according to what matters in the Premier League 2.1 weights, for both outfield and GK.
This time I was able to calculate everything quickly and comprehensively utilizing ChatGPT with HarvestGreen's excel sheets.
I haven't calculated every schedule exhaustively, just the ones you see here, most of which were the most promising previously.
Overall I recommend either 260 [Physical][Match Practice][Attacking]x2[Quickness Focus] for squad with high CA-PA gap or 156 [Physical][Quickness][Transition Restrict][Quickness Focus] with low CA-PA gap. Or 214 [Physical]x2[Attacking]x2[Quickness Focus] for something inbetween.
Green = The new best training schedules, in my opinion Light blue = The two schedules I have last recommended. They're both still worth considering. Deep blue = ZaZ's recommended schedule Pink = A long time popular meta Purple = HarvestGreen's last recommendation
Where no focus is listed, it's Quickness Focus.
Workload will be a little inaccurate, as I eyeballed the intensity bars rather than calculated it.
bf3metro said: So whatever attribute-count limitation you ran into doesn’t seem to reproduce on my current FMSS version. Maybe it was related to an older build/plugin or a different part of the data pipeline. Expand I figured it could be this. I'm using the latest version I believe. I'll try doing a reinstall later on.
You've got it working, and anyone who uses FMSS can use your weights.
Trying to debug code frustrates the hell out of me. I don't know how people willingly do it every day for a living. I've also found it tedious translating the FMST26 weights to GS.
I'm not quite satisfied with the GS weights, especially in light of Rain's recent test, and the 15% discrepancy rate between the FMST26 and GS weights. I want my weights to perform at least slightly better in every respect than anything else ideally. I give 8/10 now though.
I guess Rain's finding that there isn't too much difference overall between different weightings makes sense when you do the math on it.
5/11 players @ +2% = +10% for +2.75 position = +0.55 position per player @ +2% Maeda (83%) 3.0 vs. Dybala (69%) 6.4 = +14% for +3.4 position
+10% for +2.75 position = +3.63% for +1 position +14% for +3.4 position = +4.11% for +1 position
Could be just a coincidental correlation, but I reckon it's probably reflecting a consistent effect. The 80%/78%/76%/74% team test does suggest that the % differences can be simplified to addition, as there were strong distinctions in the results.
I'm putting my comparison test of keithb's biggest discrepancies with Premier League 2.1 on hold for now, but if we just take Maeda vs. Heung-Min Son for a quick picture, there's a ~8.9% discrepancy. Let's say in a worst case scenario you sign 4 of your starting 11 as these discrepancies:
8.9% x 4 = 35.6% If ~4% ≈ 1 position, then 35.6% = 8.9 position
So the most difference you could get from keithb or GS default weights say and mine, or whatever is best, is probably 9th vs 18th.
If we suppose the chance of a Maeda-like discrepancy (+ the curve leading up to Maeda) is ~8% total, then:
1/11 discrepancy ≈ 39.4% chance of -2.22 position 2/11 discrepancy ≈ 17.3% chance of -4.45 position 3/11 discrepancy ≈ 4.6% chance of -6.67 position 4/11 discrepancy ≈ 0.82% chance of -8.9 position
That is to be added on top of the common ~2% difference in ratings.
So let's say its a realistic worst case scenario where its 2x Maeda + 3x identical + 6x the typical ~2% difference:
~15% chance, with result of -7.45 position (i.e. 9 vs 16.45 position)
I examined GKs in particular, as GKs have no CA cost for it and some associated attributes. I did it months ago, so I can't recall the exact details now, but I can surmise that there was no significant difference at all.
The difference between all technicals (which includes all set pieces) except dribbling & finishing 1 vs 20 is negligible season position change. So there's no bonus hidden away in the results rather than the penalty % success rate either.
Having a rapid scroll through that thread trying to reach the bottom, I'd say his data is probably going to be at least as good as mine and you should probably believe what he concludes, which is that "the penalty taking attribute *might* have a very small impact on the outcome of a penalty".
Rain said: I mentioned this before, but forgot to add it to the post. Everyone was £1 million sale value or lower on Genie Scout. I tried to use FMST specifically for your 2.1 file but was tough to do considering you can't filter by sale value and that program just constantly freezes for me. I'm sure this test has a lot of rng, but it was more so for myself to see if there's even a big difference at this point. Made me feel better about using the blended and pure performance the most when buying players I guess. I didn't really take a close look at stats, but when I did look it seemed Zrelak was doing really well. I uploaded the save file to that google drive folder. Expand I was wrong about the transfer value theory. I realized this as I was creating my own my test afterwards, saw the overlap in players I myself was selecting, and remembered the 1mil limit you mentioned. My mind saw the high wages and went straight to transfer value.
I thought that nonetheless, wage is kind of a similar measurement.. particularly for older players, who have low value to advantaged age and short contracts.
But after conducting my own test, I am getting about the same results, even with my refinements that I feel do it better. My refinements are that I select just one player for each position but duplicate 3-4 times, I get my choices first by exporting shortlist from FM24 for Luton where interest is at least 'doubtful', then I choose one of the top 4 or 5 who has the lowest value, and I play it out as a normal game (meta training, friendlies, etc.) and I'm testing one team at a time (replaced the roster at Luton).
My results so far are:
Premier League 2.1 - 10th, 9th, 11th = 10 position Keithb - 11th, 11th, 16th, 13th = 12.75 position
keithb:
GK - Edgar Badia 79.1% DL - Hamza Mendyl 69.9% DC - Kim Min Tae 72.4% DR - Jusuf Gazibegovic 72.1% DM - Gaius Makouta 73.5% AML - Michael (Al-Hilal) 70.3% AMR - Jacob Murphy 74.8% ST - Jhon Cordoba 77.4%
Premier League 2.1:
GK - Edgar Badia 79.1% DL - Jesus Gallardo 72.6% DC - Kim Min Tae 72.4% DR - Wajdi Kechrida 74.1% DM - Aliou Dieng 74.8% AML - Zakaria Aboukhlal 73.2% AMR - Jacob Murphy 74.8% ST - Jhon Cordoba 77.4%
% is my FMST26 rating. So there's a ~2% difference on 5/11 of the team.
I'm currently testing a new idea that occurred to me after this, which is what if I take keithb's top players that my rating says are rubbish - so the biggest discrepancies - and then test a team of those. So we'll see what that difference is.
Of course, ordinarily one would not happen to pick all the bad eggs in keithb's file; the difference in normal use remains ~2.75 positions according to my test, or a marginal difference according to yours.
Before when you were doing your test, I also manually had a go at wiping Luton and buying the best players I could with a £10mil budget. So basically actually playing the game to see the best I could actually sign as the lowest and impoverished Premier League team. Honestly I expected I would get 1st, but actually I only achieved 8th.
Your translation isn’t wrong at all — I think we’re just optimizing for two slightly different things.
Your method normalizes the coefficients so the strongest attribute weight becomes 20, which makes a lot of sense if the goal is to express the weights themselves on an FM-style 1–20 scale.
For mine, I used GPT-5.6’s reasoning to preserve the full FMST formula as literally as possible and translate the final FMST score onto FMSS’s 1–20 scale.
The main difference is that I use the same ÷5 conversion for both outfield and GK, because that gives:
FMSS Meta = FMST score ÷ 5
So 100 → 20, 90 → 18, 75 → 15, etc.
That also means GK and outfield remain on exactly the same scale. With the “highest coefficient = 20” approach, outfield ends up using roughly ÷4.7 while GK uses roughly ÷4.4445, so the two formulas are normalized slightly differently.
I also kept every FMST term that FMSS can access, including hidden/personality attributes, the CA penalty for GKs, and the original nonlinear Injury Proneness and Dirtiness calculations rather than simplifying them into fixed weights.
So I definitely wouldn’t call your version wrong — it’s a good normalized/simplified implementation. I just think the GPT-5.6 version is more faithful if the objective is specifically to port the complete FMST calculation into FMSS while keeping the output on a consistent 1–20 scale.
If you want, I can send/upload the exact code I’m using so you can compare the implementations directly. Expand With your formula, the problem is that it uses too many attributes. For some reason, FMSS will fail with all of them included. At least it did for me.
Suppose we have that situation where a file makes Haaland rank 300. How do we actually tell Haaland from Joe Blow from Slough, in that case? How can we say that file is 'wrong', when it actually 'performs' better?
The answer is we look at his value. This is why value matters, even for assessing pure performance. Because once we see Haaland's £300mil asking price, compare it to our £18,582 transfer budget, we quickly move on to the player below him - who is Joe Blow from Slough.
Keep in mind it's not at all saying that higher value = strictly better performance. What I'm instead saying is that if a player has a significantly higher value, it's highly unlikely that he is going to be worse. It's like if I buy a $200 toaster. Is it going to be worse than the old cheap one? Probably not. Is it going to toast better? Not necessarily.
The fundamental basis for this is CA vs attribute imbalance. Very old players have the worst imbalances (i.e. Dybala - ~9 physical, ~18 mental/technical) and often high rep, and therefore generally the worst over-valuations. Dybala is valued at £10mil. Now there are a few players who are significantly better for the same price or less, such as Maeda, but it's a minority. Now try the inverse, find me a player who is low value but very highly rated. Again, it's a minority.
Now you might say, well if value is such a reliable correlation, why not rely on that alone? The answer is that those players with £300mil valuations, even though they make up the majority of the top players, are useless to us. We are looking for the diamonds in the rough, even if we 'just want the best performing players right now', because you need to pay to get them! Hence valuations are an important/useful measurement for telling Haaland from Joe Blow, but NOT for ranking utility, and even if you did have the billions to spend, valuation does an inadequate job of distinguishing exceptions to the rule. It is an assessment and therefore indicator of quality itself rather than a contributor to quality, and so it can be used to critique a ranking but not determine it. In fact we want our ranking system to find as many exceptions to the valuation ranking as possible, so long as they are valid.
Rain said: If anyone is interested in doing more I can upload the save file Expand If you don't mind, I would like that save file, to have a closer look at it.
I was anxious about this test because I haven't done it myself since comparing Pure Performance & Premier League 1.0 with Orion's weights, and I thought it's 50/50 how it goes.
Now that it's done I'll make my comments and critiques, and thankfully at least my Pure Performance file topped out by a decent margin, so hopefully less chance I'll be seen as just being a sour puss.
I guess I'll start by saying now it can be seen that I wasn't just making up the numbers in my own comparison test, where the results were Pure Performance 6.5, Premier League 1.0 7.333, and Orion's weights 13.166. You can see the results are similar, particularly if you look at the individual samples (I didn't do as many samples).
Now, my critiques:
I'll start with an intuitive example. In keithb's file Maeda is rank 102 for FS, and in Premier League 2.1 he is rank 17. Regardless of which file you assume is better, if you were assessing based on this player alone there would be obviously be a great difference. The thing is, these exceptions to the rule constitute a minority of players, and so if you're picking a full squad, half are going to be identical players, and many of the rest are going to be minor differences like Mbappe vs. Haaland. In fact, it's a surprise there's any significant difference at all between HarvestGreen's/Orion's weights and Pure Performance for this reason.
Now if you're on board with that, then let's take it to the next matter. Because now I'm going to show that those exceptions matter, and I would argue you can't really just conclude 'well, if they're all about the same once you buy a full squad, who cares'.
Suppose I want a top striker for my team. I load up keithb's ratings, sort by FS, and find that the cheapest of the top 20 is Victor Gyokeres for a GS estimated sale value of £100mil. If I use Premier League 1.0, in the top 20 I have Furuhashi (£33.8mil), Maeda (£22.4mil), Jhon Cordoba (£8.5mil) and even a few more under the cheapest in keithb's ratings. If I load up FM24 now as Fulham and make an offer for Gyokeres, they want £41.5mil, whereas for Maeda it's £12.5mil, so it's not just GS. Pure Performance also has this problem, although to a lesser extent than keithb's file - the cheapest in the top 20 FS is Umar Sadiq for £40mil sale value (~£20mil in game).
Now who would argue money is no matter in FM?
So one thing I'm interested in seeing in your save, is the relative cost/value of each squad. I don't actually know if Premier League 2.1 would be significantly less expensive to purchase, but I suspect it would be.
Now the matter of performance. I have high confidence that Premier League 2.1 is going to give more accurate performance estimations than Pure Performance. I am confident because I have tested a player like Maeda, and I know with certainty he gets the team to 3rd, so he's definitely up there with the best. So how then does it finish worse here? Because these files should be comparing by performance alone, not cost - the low cost would just be a symptom of the improved efficiency.
I can think of a few possibilities here, but I'll just say the main two.
You 'froze' and 'healed' the players. Two of the biggest factors in Premier League 2.1 are dirtiness and injury proneness. If you prevent red cards, injuries, and so on, then you'd be disrupting a key part of the natural process. But I'm not sure exactly what the freeze/healing extends to.
The second thing is that while the goal of Premier League 2.1 is to assess performance rather than value, this does nonetheless loop back around.. let me demonstrate what I mean by this:
Your selections for ST are as follows:
Premier League 2.1 - Adam Zrelak, 113 CA, £2k/week, Warta Poznań (Poland Div 1) Keithb - Alexis Sanchez, 137 CA, £78k/wk, Inter Milan Pure Performance - Jair Silva, 112 CA, £1.2k/week, Leiria (Portugal Div 2) Orion - Choupo-Moting, 137 CA, £123k/wk, FC Bayern HarvestGreen - Dickson Abiama, 110 CA, £10k/wk, FC Kaiserslautern
If I've thought about this right, what is going on here is a problem with the methodology where because the better players are actually getting pushed up in Premier League 2.1 (i.e. Maeda 102 -> 16) this leaves what's left down at player 102 slightly weaker and yet still adequate, while also being more populated by even lower CA players who have been pushed up (i.e. Zrelak 300 > 100 say).
To complicate matters, this would only affect a minority of players, because people like Maeda and Zrelak are the exceptions not the rule, and for that majority remainder most of them are going to be the same selections (overlap) while one or two (differences) may even be better than Premier League 2.1. So for example if you selected players at rank 102, guess what, keithb gets Maeda and Premier League 2.1 gets someone like Zrelak. But actually Alexis Sanchez is higher rated in keithb's file than in Premier League 2.1. So it's overall its like I got Maeda/Kuruhashi/Cordoba right and Sanchez wrong, while keithb gets Maeda/Kuruhashi/Cordoba wrong, but gets Sanchez right. But you tell me what you think of this theory.
To put it another way, the file that would finish 1st in your test would have Haaland, Mbappe, etc. at the rank of 300 or whatever it is that you are selecting at. However it does nonetheless have use, because it tells us how many duds are pulling the team down. For example, keithb's team has an ST from Inter Milan, yet why does it still get beaten by Pure Performance on average? I reckon if we have a look at the file, Sanchez will get 7.3+ rating, but a few duds will get 6.5. Just my guess.
I couldn't include every attribute, there's some limit occurring in FMSS, but it'll only be a few minor ones missing. I had to take off professionalism from outfield though. You could swap 'Balance: 3.03' for 'Professionalism: 3.19' if you like.
I've simplified dirtiness & injury proneness. I also can't verify the accuracy at all because it's FM26 only.
Rain said: Currently running sims it will probably take a while to get a good sample size. Took me forever just setting it up for 5 teams. There's definitely dupes (Especially pure performance team), but should be different enough. Had to avoid the big 3 FA GKs in the original db as they would have been in every team (De Gea, Leno, Neto) Although ended up with a few Ochoa anyway. Expand So what files are you comparing, and are you using FM24 or FM26?
Make sure you're using full detail, can be overlooked
I've come up with my own variation of the 4-2-4 formation/Knap tactic that is quite unique I believe, based on a theory I have. Initial results are 5th, 7th, 10th, 3rd as Luton and 6th, 7th as meta team after I've fixed it back up. So possibly equal to the Knap tactic, and perhaps a tweak or two can push it higher. I figure that I would probably only release a tactic if I can get 1st as Luton.
Does anyone know if 1st as Luton has been achieved with a tactic test? Obviously just the default test run, no managing the team throughout the season to achieve it.
Yes, I'm still keen to see your go at it. The more options the better.
I guess with FM27 coming soon, its 50/50 whether people will move on entirely from FM24 or not. But if FM27 turns out to be crap 2.0, then I think these new tools will help keep FM24 alive by injecting a bit of novelty into it.
I started testing with my meta team + Sterling that always seemed to finish a rock solid ~6th.
It didn't take me long to come up with an unusual tactic that finished 5th. However when I tested this with real Luton, they only finished 8th-16th, whereas the Knap tactic can finished as high as 2nd (2nd-10th?). My tactic won with Man City, but only with 89 pts (Knap as high as ~100-110?).
And it's not just my tactic. 424 Mountain King did better than the Knap tactic with my test team from memory, but I didn't get 2nd with it for real Luton.
I tested a bunch more Knap tactics. The results in his spreadsheet don't line up with my results, whereas by contrast it seems that FM Arena's testing is accurate in spite of the methodology critiques I would have of it. Nonetheless all the ones I tried did well, at least 9th for real Luton and often better. But nothing got 2nd like EF 424 IF HP V2 P101 AC.
Of the alternatives I tested, FM24.4HGF4231 V3X RM P107 ALL CUPS seemed the best. 3rd, 6th Luton. Man City 1st 94pts.
Curiously I could use my unusual tactic with several positions I had no player proficiency in at all across my meta team, and yet it still did well. Not quite as good, but it held up well.
Now here's the point we can take beyond FM24:
The results suggest that roles or player instructions are in fact dependent on the attributes of players in some respect.
Yet we also know from the meta team results that it's not what the roles highlight, nor combined team quality.
Injuries, differing player choices, etc. can account for the high variation in the real Luton results relative to the meta team test, but it doesn't explain why only the Knap tactic peaks far higher with real Luton.
One idea I have is that perhaps its not combined team quality, but how each player's attribute is relative to others in his team. In the official guide, they talk of 'absolute' vs 'relative' attributes, and most attributes are 'relative'; perhaps this is what they were actually talking about. There is some suggestive evidence that makes me think along these lines, such as that with finishing, you only need a player with ~7 finishing, but any figure can do the job really, and what happens is whoever in the team has the highest finishing is the one scores the most goals, even if they're a DL with 5 finishing. So we see here that a low attribute altered the tactical outcomes, because of how it was relative to others in the team.
So for instance in my meta team all my players have 8 passing and 13 dribbling, the tactic treats them as all equally likely distribute the task of passing and dribbling I suppose. Perhaps an attribute such as 'teamwork' even influences how well these tasks are evenly distributed throughout the team. But then also for each individual player, they would be more likely to dribble (13) than pass (8).
It's pure speculation, but it would explain why virtually every technical attribute matters zero percent. It would be that the tactic minimizes their influence, and that tactics which don't minimize their influence get beat anyway, whereas Knap's tactic wins almost every match. In other words it could be that just the tactics are broken, rather than that the attributes are also broken. 0 direct passes x 20 passing = effectively 0 passing. Or, 20 direct passes x 20 passing x 0% win rate = 0% win rate for 20 passing.
That's my thinking so far.
18 or less = Don't need competitive matches, but do play in friendlies (u18s/reserves fine) mainly to keep their match sharpness up in order to avoid injury more than development
21+ = Need at least 15 full matches (1350 minutes) per season to develop adequately, 25-35 is ideal, so this is why loans typically become necessary
19-20 = Inbetween the two above
Don't bother bringing on young players for a few minutes at the end of a game. The game counts experience in terms of minutes not appearances, and I suspect it could also be counting it the same way it counts minutes for match sharpness, which is that it counts in precise 11 minute blocks (I think it was 11 minutes.. something near that anyway).
I don't know what HarvestGreen tested them with, but he recommends 0-0-0-2x-2x, and I've played around with it myself before and found that to be best too.
211 = (((10x1182.2185) + 1000.59) / ((10x37.75) + 43.25)) / 0.977272 = 31.1847
331 = (((10x1202.567) + 1000) / ((10x39.62) + 43)) / 1.05 = 28.2454
260 = (((10x1156.35) + 965.23) / ((10x38.98) + 43.00)) / 1.05 = 27.5696
142 = (((10x1135.2835) + 1048.42) / ((10x36.43) + 42.33)) / 1.25 = 24.3981
So here I am trying to take into account outfield + GK + CA efficiency + intensity.
HarvestGreen didn't include the GK stats for 331 as far as I can see, so I've just inserted a high value of 1000. Even with this high value for GK, it doesn't seem as quite as good overall as 214 or 211, though it does have the highest absolute gain in its favor.
There is one more thing to add into the assessment, which is the effect on other various in-game factors:
214 [Physical]x2[Attacking]x2[Quickness Focus]
331 [Physical][Match Practice][Attacking][Defending][Quickness Focus]
211 [Handling][Shot Stopping][Attacking][Physical][Chance Conversion][Aerial Defence][Ground Defence] [Distribution][Quickness Focus]
331 has a slight edge over 214 due to the better balancing of match sharpness.
211 is much better for match sharpness, but is probably going to wear down condition too much during the competitive season, especially if you have weeks with 2 matches. The main problem with 211 is that it would be so finicky to implement that in all likelihood you're not going to implement it properly every week, and the careful handling of condition would add to this. 211 is nonetheless very good for GK and absolute gain, both of which 214 falls a bit short on.
One more thing to say about 214 is that its what I found to be best using a different method previously. So where things are roughly equal, I'm inclined to lean towards it anyway. And in HarvestGreen's assessment, 214 also scores higher on his 'quality' index than these other 2.
So overall I will still be recommending 214. I just realized that 331 was actually my 2nd recommendation before, I didn't realize until now that that was HarvestGreen's recommendation.
It might be worth setting up 211 for your u21s/u18s, especially if they have no competitive league or cup and you can set and forget it with weekly friendlies for the whole season. Reason is, u21s/u18s easily end up in match sharpness ruts they never get out of, and low match sharpness is a big influence on injury rate, and more injuries = less development.
It's splitting hairs between 214 and 331, so if you feel more comfortable going with 331, that's perfectly fine.
One thing left unaddressed is GK focus. HarvestGreen's data used 'Agility and Balance' as GK focus, so we can't be sure of the effect of any others. But it seems to be the case that Agility isn't as critical as a few others: Determination, Reflexes, Concentration. Personally I've started favoring 'GK reactions' because it focuses on two of those (Reflexes, Concentration). But there's no data on this.
142 GK - 1048.42 / 42.33 = 24.76643
211 GK - 1000.59 / 43.25 = 23.14023
260 GK - 965.23 / 43.00 = 22.44721
212 GK - 868.76 / 41.00 = 21.18927
214 GK - 805.20 / 39.33 = 20.45888
211 = 130% GK, 100% Def, 90% Att
214 = 100% GK, 100% Def, 100% Att
260 = 105% GK, 105% Def, 105% Att
142 = 125% GK, 125% Def, 125% Att
..and now, as I was examining HarvestGreen's own rankings, I gradually cottoned to the fact that there's a huge standout error in my data.
214, 211, 142 and 260 are all at the top of his rankings just like mine.. yet he has 331 as the best. I figured there has to be something amiss here, and sure enough, getting ChatGPT to recalculate it gives 1202.567.
That means HarvestGreen's recommendation of 331, would in fact be the best schedule for outfield.
I could have sworn I verified it with ChatGPT as 1020.63 three times just to be sure at the time. So I went back to check, since I still had the convo open:
Fuckin' ChatGPT man..
I presume it got at least most of them right, but now I can't be certain some weren't miscalculated by ChatGPT.
Anyway, at least we know from the backing of HarvestGreen's rankings that 331, 260, 211, 214 and 142 are the top ones and that the numbers are most likely correct for at least those.
You should be aware also this mirroring of HarvestGreen's data is a pleasant unexpected surprise. I did have HarvestGreen's weights as a starting or reference point for Premier League 2.1, but as you can see with Rain's comparison test of HarvestGreen's weights vs. Premier League 2.1, things should have changed substantially in the end. So it's surprising that the favored training schedules nonetheless align.
bf3metro said: If you had to pick only one between 214 and 260 for the entire squad, which one would you use?
I'll get back to you in a day or two, as things are still evolving right now, as you can see from my words above
We have a new top one, that I've highlighted green. This one was actually suggested by tam1236 back in january. He took the HarvestGreen attribute test findings and did a calculation on it to find this as the best training schedule.
It even has pretty good efficiency, though I suspect the workload could be too high to be worth using with all those modules. Very impressive though given it also has one of the highest CA gains (37.75), so it's particularly good if you want to keep your players looking well rounded for aesthetic reasons without compromising on actual performance.
It turns out there are a lot of very high efficiency options above 1000 (so still ~85%+ of the absolute best), and the most efficient is now Recoveryx13. I wouldn't recommend it, but it's interesting there's a profound difference in training quality between rest and recovery on their own.
I've highlighted in yellow some schedules that stand out to me, but I haven't taken a close look yet.
104 [Recovery]x13 = 1008.64 / 17.32 = 58.23557
85 [Match Review] = 1008.62 / 17.40 = 57.96667
100 [Match Review]x2 = 1005.48 / 17.72 = 56.74831
105 [Quickness][Attacking] = 1014.63 / 17.84 = 56.87276
52 [Match Tactics] = 1033.08 / 19.54 = 52.87001
124 [Physical][Quickness][Aerial Defence][Recovery]x10 = 1001.86 / 21.70 = 46.16820
127 [Physical][Quickness][Aerial Defence][Match Review] = 1002.89 / 22.92 = 43.75742
125 [Physical]x2[Aerial Defence] = 1032.49 / 24.20 = 42.66488
126 [Quickness]x2[Aerial Defence] = 1013.15 / 23.92 = 42.35535
122 [Physical][Quickness][Ground Defence] = 1020.02 / 24.26 = 42.04700
116 [Physical]x2[Chance Conversion] = 1022.72 / 24.57 = 41.61579
94 [Physical][Tactical][Recovery]x7 = 1053.14 / 25.88 = 40.69397
118 [Physical]x3[Tactical] = 1042.33 / 25.51 = 40.85927
138 [Physical][Quickness][Resistance][Tactical] = 1046.47 / 26.48 = 39.51926
93 [Physical][Tactical] = 1061.43 / 27.12 = 39.13827
215 [Physical]x3[Attacking] = 1018.42 / 25.91 = 39.30606
213 [Physical]x2[Attacking] = 1043.1425 / 27.79 = 37.53665
95 [Match Practice][Recovery]x7 = 1052.22 / 28.04 = 37.52568
106 [Quickness][Attacking][Overall] = 1051.52 / 28.96 = 36.31077
325 [Physical][Quickness][Resistance][Defending][Defending Engaged] = 1126.288 / 31.62 = 35.61379
321 [Physical][Quickness][Aerial Defence][Attacking] = 1081.229 / 31.46 = 34.36834
340 [Physical]x2[Attacking Wings][Attacking][All in Attack Group] = 1085.2745 / 31.27 = 34.73855
334 [Physical]x2[Aerial Defence][Attacking][All in Defend Group] = 1078.23 / 31.21 = 34.54758
317 [Physical]x2[Chance Conversion][Attacking][All in Attack Group] = 1070.2455 / 30.71 = 34.84811
186 [Physical][Quickness][Aerial Defence][Attacking] = 1074.0845 / 31.24 = 34.38555
53 [Match Practice] = 1038.74 / 30.02 = 34.60293
335 [Physical][Quickness][Ground Defence][Attacking][All in Defend Group] = 1103.957 / 32.56 = 33.90531
96 [Physical][Tactical][Match Practice][Recovery]x7 = 1112.92 / 32.98 = 33.74530
98 [Defending]x3[Attacking]x2[Physical]x3[Match Practice]x2[Ground Defence] = 1112.56 / 33.72 = 32.99525
91 [Attacking]x5[Defending]x5[Match Practice]x2[GoalKeeping][No focus] = 1126.31 / 34.78 = 32.38384
338 [Physical][Match Practice][Attacking][Defending][Recovery]x7 = 1117.508 / 34.99 = 31.93792
112 [Physical][Match Practice][Chance Conversion] = 1117.34 / 35.60 = 31.38596
212 2x Physical + 2x Match Practice + Attack + Defend + Set Pieces = 1126.966 / 35.78 = 31.49709
211 Handling, Shot Stopping, Attacking, Physical, Chance Conversion, Aerial Defence, Ground Defence and Distribution = 1182.2185 / 37.75 = 31.317
149 [Match Practice][Attacking][Match Tactics]x2[Tactical]x2 = 1100.35 / 35.50 = 30.99577
205 Game preset training 1 'Preset-Training Style-Balance' [Overall][Defending][Attacking][Set Piece Routines][Outfield][Recovery]x2 = 1125.04 / 36.50 = 30.82301
97 [Chance creation][Attacking][Aerial Defense][Handling][Defending from the front][Quickness] = 1122.54 / 36.48 = 30.77138
It's a more subjective matter, but what I found is that using home friendly matches against the worst team possible is better than using recovery periods in training.
Friendly matches improve morale, build familiarity, build up match sharpness a lot better, and I daresay even result in less injuries.
When using all recovery I noticed that it tends to result in too low condition and eventually fatigue, so it had to be half recovery half rest, which also can be tedious to set up properly each week.
I can't remember if this is the case, but recovery might also interact differently with the other training modules to produce different results than what it shown in HarvestGreen's spreadsheets.
I’ve been using ZaZ’s Growth 2, but in my save it really isn’t delivering the development I expected.
What I’m mainly looking for is a schedule that pushes CA growth as fast as possible so players reach their PA quickly, rather than just maximizing CA efficiency.
Would you recommend switching fully to something like 260 / 214 / 156 depending on the players’ CA–PA gap, or do you have a specific weekly/monthly rotation you think is optimal in practice?
ZaZ Growth 2 would be 1064.2675-1115.573
Oddly enough the 2-match Physicalx2 Attackingx2 scores better than the Physicalx5 Attackingx5 and results in higher CA gain too.
[Physical]x2[Attacking]x2 = 1115.573 / 33.83 = 32.976
[Physical]x5[Attacking]x5 = 1064.2675 / 27.85 = 38.214
Both of these are very close to the best.
Because CA-PA gap is a critical accelerant to growth, you actually want the lowest overall CA growth possible - while highest in the CA growth where it matters. There is a hard cap on CA per season anyway, which is ~20-25 (someone has reported ~30 i think). If your players aren't developing much, it's more likely that they're over 18 and not getting enough match experience. No match experience for older players can reduce growth to ~1 CA/season.
This time I was able to calculate everything quickly and comprehensively utilizing ChatGPT with HarvestGreen's excel sheets.
I haven't calculated every schedule exhaustively, just the ones you see here, most of which were the most promising previously.
Overall I recommend either 260 [Physical][Match Practice][Attacking]x2[Quickness Focus] for squad with high CA-PA gap or 156 [Physical][Quickness][Transition Restrict][Quickness Focus] with low CA-PA gap. Or 214 [Physical]x2[Attacking]x2[Quickness Focus] for something inbetween.
Weighted attribute improvement:
260: [Physical][Match Practice][Attacking]x2 = 1156.35
142: [Physical][Quickness][Attacking]x3 = 1135.2835
212: 2x Physical + 2x Match Practice + Attack + Defend + Set Pieces = 1126.97
97: [Chance creation][Attacking][Aerial Defense][Handling][Defending from front][Quickness][Quick focus] = 1122.54
338: [Physical][Match Practice][Attacking][Defending][Recovery]x7[Quickness Focus] = 1117.508
214: [Physical]x2[Attacking]x2[Quickness Focus] = 1115.573
149: [Match Practice][Attacking][Match Tactics]x2[Tactical]x2 = 1101.13
139: [Physical][Quickness][Resistance][Defending] = 1091.794
339: [Quickness][Attacking][Match Practice] = 1086.43
243: [Quickness][Match Practice[Attackingx2][Quickness focus] = 1079.689
150: [Attackingx6][Quickness focus] = 1079.6675
334: [Physical]x2[Aerial Defence][Attacking] = 1078.23
90: [Attacking]x3 = 1061.16
466: [Quickness][Physical][Attacking] Additional focus: Ball Control = 1056.47
147: [Attacking][Match Tactics]x3 = 1046.77
164: [Physical][Quickness][Aerial Defence]x2[Match Review] = 1039.84
166: [Resistance][Quickness][Chance Conversion] = 1028.64
156: [Physical][Quickness][Transition Restrict] = 1026.7215
331: [Physical][Match Practice][Attacking][Defending] = 1020.63
160: [Physical][Quickness][Resistance][Transition Restrict]x2 = 1018.89
148: [Attacking][Match Tactics]x3[Tactical]x2 = 1011.46
327: [Quickness][Match Practice][Attacking]x2[Strength Focus] = 1005.93
116: [Physical]x2[Chance Conversion] = 1001.58
65: [Chance Creation] = 1000.90
53: [Match Practice] = 992.97
170: [Endurance][Quickness][Chance Conversion] = 962.78
123: [Quickness][Attacking][Transition Restrict] = 958.12
106: [Quickness][Attacking][Overall] = 948.62
83: [Community Outreach] = 940.47
113: [Quickness][Match Practice][Chance Conversion][Quickness focus] = 927.185
161: [Physical][Quickness][Aerial Defence][Recovery]x7 = 916.81
75: [One on Ones] = 910.01
54: [Attacking Wings] = 907.79
55: [Attacking Patient] = 891.86
99: [Chance Conversion][Match Review] = 848.14
168: [Quickness]x2[Chance Conversion] = 838.56
86: [Rest]Quickness Focus] = 792.26
43: [Rest][No Focus] = 735.1305
Weighted attribute improvement divided by CA cost:
86: [Rest]Quickness Focus] = 792.26 / 17.53 = 45.194
43: [Rest][No Focus] = 735.1305 / 16.85 = 43.628
156: [Physical][Quickness][Transition Restrict] = 1026.7215 / 24.20 = 42.42
166: [Resistance][Quickness][Chance Conversion] = 1028.64 / 25.71 = 40.010
65: [Chance Creation] = 1000.90 / 25.80 = 38.794
139: [Physical][Quickness][Resistance][Defending] = 1091.794 / 28.71 = 38.027
466: [Quickness][Physical][Attacking] Additional focus: Ball Control = 1056.47 / 29.00 = 36.430
334: [Physical]x2[Aerial Defence][Attacking] = 1078.23 / 31.21 = 34.545
214: [Physical]x2[Attacking]x2[Quickness Focus] = 1115.573 / 33.83 = 32.973
338: [Physical][Match Practice][Attacking][Defending][Recovery]x7[Quickness Focus] = 1117.508 / 34.99 = 31.939
212: 2x Physical + 2x Match Practice + Attack + Defend + Set Pieces = 1126.97 / 35.78 = 31.495
142: [Physical][Quickness][Attacking]x3 = 1135.2835 / 36.43 = 31.164
149: [Match Practice][Attacking][Match Tactics]x2[Tactical]x2 = 1101.13 / 35.50 = 31.017
97: [Chance creation][Attacking][Aerial Defense][Handling][Defending from front][Quickness][Quick focus] = 1122.54 / 36.48 = 30.771
339: [Quickness][Attacking][Match Practice] = 1086.43 / 35.72 = 30.416
90: [Attacking]x3 = 1061.16 / 34.98 = 30.335
150: [Attackingx6][Quickness focus] = 1079.6675 / 36.20 = 29.825
260: [Physical][Match Practice][Attacking]x2 = 1156.35 / 38.98 = 29.664
243: [Quickness][Match Practice[Attackingx2][Quickness focus] = 1079.689 / 37.93 = 28.460
331: [Physical][Match Practice][Attacking][Defending] = 1020.63 / 39.62 = 25.759
GK:
142: [Physical][Quickness][Attacking]x3 = 1048.42
260: [Physical][Match Practice][Attacking]x2 = 965.23
214: [Physical]x2[Attacking]x2[Quickness Focus] = 805.20
156: [Physical][Quickness][Transition Restrict] = 717.03
139: [Physical][Quickness][Resistance][Defending] = 662.44
243: [Quickness][Match Practice[Attackingx2][Quickness focus] = 614.57
86: [Rest][Quickness Focus] = 310.616
43: [Rest][No Focus] = 282.70
10x outfield + 1x GK:
260: [Physical][Match Practice][Attacking]x2 = 1138.975
142: [Physical][Quickness][Attacking]x3 = 1127.386
214: [Physical]x2[Attacking]x2[Quickness Focus] = 1087.357
139: [Physical][Quickness][Resistance][Defending] = 1052.762
243: [Quickness][Match Practice[Attackingx2][Quickness focus] = 1037.405
156: [Physical][Quickness][Transition Restrict] = 998.567
86: [Rest][Quickness Focus] = 748.474
43: [Rest][No Focus] = 694.000
Workload:
156: [Physical][Quickness][Transition Restrict] = 95%
260: [Physical][Match Practice][Attacking]x2 = 100%
214: [Physical]x2[Attacking]x2 = 105%
243: [Quickness][Match Practice[Attackingx2][Quickness focus] = 120%
142: [Physical][Quickness][Attacking]x3 = 125%
139: [Physical][Quickness][Resistance][Defending] = 135%
Notes:
Green = The new best training schedules, in my opinion
Light blue = The two schedules I have last recommended. They're both still worth considering.
Deep blue = ZaZ's recommended schedule
Pink = A long time popular meta
Purple = HarvestGreen's last recommendation
Where no focus is listed, it's Quickness Focus.
Workload will be a little inaccurate, as I eyeballed the intensity bars rather than calculated it.
I figured it could be this. I'm using the latest version I believe. I'll try doing a reinstall later on.
You've got it working, and anyone who uses FMSS can use your weights.
Trying to debug code frustrates the hell out of me. I don't know how people willingly do it every day for a living. I've also found it tedious translating the FMST26 weights to GS.
I'm not quite satisfied with the GS weights, especially in light of Rain's recent test, and the 15% discrepancy rate between the FMST26 and GS weights. I want my weights to perform at least slightly better in every respect than anything else ideally. I give 8/10 now though.
I guess Rain's finding that there isn't too much difference overall between different weightings makes sense when you do the math on it.
5/11 players @ +2% = +10% for +2.75 position = +0.55 position per player @ +2%
Maeda (83%) 3.0 vs. Dybala (69%) 6.4 = +14% for +3.4 position
+10% for +2.75 position = +3.63% for +1 position
+14% for +3.4 position = +4.11% for +1 position
Could be just a coincidental correlation, but I reckon it's probably reflecting a consistent effect. The 80%/78%/76%/74% team test does suggest that the % differences can be simplified to addition, as there were strong distinctions in the results.
I'm putting my comparison test of keithb's biggest discrepancies with Premier League 2.1 on hold for now, but if we just take Maeda vs. Heung-Min Son for a quick picture, there's a ~8.9% discrepancy. Let's say in a worst case scenario you sign 4 of your starting 11 as these discrepancies:
8.9% x 4 = 35.6%
If ~4% ≈ 1 position, then 35.6% = 8.9 position
So the most difference you could get from keithb or GS default weights say and mine, or whatever is best, is probably 9th vs 18th.
If we suppose the chance of a Maeda-like discrepancy (+ the curve leading up to Maeda) is ~8% total, then:
1/11 discrepancy ≈ 39.4% chance of -2.22 position
2/11 discrepancy ≈ 17.3% chance of -4.45 position
3/11 discrepancy ≈ 4.6% chance of -6.67 position
4/11 discrepancy ≈ 0.82% chance of -8.9 position
That is to be added on top of the common ~2% difference in ratings.
So let's say its a realistic worst case scenario where its 2x Maeda + 3x identical + 6x the typical ~2% difference:
~15% chance, with result of -7.45 position (i.e. 9 vs 16.45 position)
I examined GKs in particular, as GKs have no CA cost for it and some associated attributes. I did it months ago, so I can't recall the exact details now, but I can surmise that there was no significant difference at all.
The difference between all technicals (which includes all set pieces) except dribbling & finishing 1 vs 20 is negligible season position change. So there's no bonus hidden away in the results rather than the penalty % success rate either.
Having a rapid scroll through that thread trying to reach the bottom, I'd say his data is probably going to be at least as good as mine and you should probably believe what he concludes, which is that "the penalty taking attribute *might* have a very small impact on the outcome of a penalty".
I was wrong about the transfer value theory. I realized this as I was creating my own my test afterwards, saw the overlap in players I myself was selecting, and remembered the 1mil limit you mentioned. My mind saw the high wages and went straight to transfer value.
I thought that nonetheless, wage is kind of a similar measurement.. particularly for older players, who have low value to advantaged age and short contracts.
But after conducting my own test, I am getting about the same results, even with my refinements that I feel do it better. My refinements are that I select just one player for each position but duplicate 3-4 times, I get my choices first by exporting shortlist from FM24 for Luton where interest is at least 'doubtful', then I choose one of the top 4 or 5 who has the lowest value, and I play it out as a normal game (meta training, friendlies, etc.) and I'm testing one team at a time (replaced the roster at Luton).
My results so far are:
Premier League 2.1 - 10th, 9th, 11th = 10 position
Keithb - 11th, 11th, 16th, 13th = 12.75 position
keithb:
GK - Edgar Badia 79.1%
DL - Hamza Mendyl 69.9%
DC - Kim Min Tae 72.4%
DR - Jusuf Gazibegovic 72.1%
DM - Gaius Makouta 73.5%
AML - Michael (Al-Hilal) 70.3%
AMR - Jacob Murphy 74.8%
ST - Jhon Cordoba 77.4%
Premier League 2.1:
GK - Edgar Badia 79.1%
DL - Jesus Gallardo 72.6%
DC - Kim Min Tae 72.4%
DR - Wajdi Kechrida 74.1%
DM - Aliou Dieng 74.8%
AML - Zakaria Aboukhlal 73.2%
AMR - Jacob Murphy 74.8%
ST - Jhon Cordoba 77.4%
% is my FMST26 rating. So there's a ~2% difference on 5/11 of the team.
I'm currently testing a new idea that occurred to me after this, which is what if I take keithb's top players that my rating says are rubbish - so the biggest discrepancies - and then test a team of those. So we'll see what that difference is.
Of course, ordinarily one would not happen to pick all the bad eggs in keithb's file; the difference in normal use remains ~2.75 positions according to my test, or a marginal difference according to yours.
Before when you were doing your test, I also manually had a go at wiping Luton and buying the best players I could with a £10mil budget. So basically actually playing the game to see the best I could actually sign as the lowest and impoverished Premier League team. Honestly I expected I would get 1st, but actually I only achieved 8th.
Your translation isn’t wrong at all — I think we’re just optimizing for two slightly different things.
Your method normalizes the coefficients so the strongest attribute weight becomes 20, which makes a lot of sense if the goal is to express the weights themselves on an FM-style 1–20 scale.
For mine, I used GPT-5.6’s reasoning to preserve the full FMST formula as literally as possible and translate the final FMST score onto FMSS’s 1–20 scale.
The main difference is that I use the same ÷5 conversion for both outfield and GK, because that gives:
FMSS Meta = FMST score ÷ 5
So 100 → 20, 90 → 18, 75 → 15, etc.
That also means GK and outfield remain on exactly the same scale. With the “highest coefficient = 20” approach, outfield ends up using roughly ÷4.7 while GK uses roughly ÷4.4445, so the two formulas are normalized slightly differently.
I also kept every FMST term that FMSS can access, including hidden/personality attributes, the CA penalty for GKs, and the original nonlinear Injury Proneness and Dirtiness calculations rather than simplifying them into fixed weights.
So I definitely wouldn’t call your version wrong — it’s a good normalized/simplified implementation. I just think the GPT-5.6 version is more faithful if the objective is specifically to port the complete FMST calculation into FMSS while keeping the output on a consistent 1–20 scale.
If you want, I can send/upload the exact code I’m using so you can compare the implementations directly.
With your formula, the problem is that it uses too many attributes. For some reason, FMSS will fail with all of them included. At least it did for me.
Normalizing them however you like is fine though
Suppose we have that situation where a file makes Haaland rank 300. How do we actually tell Haaland from Joe Blow from Slough, in that case? How can we say that file is 'wrong', when it actually 'performs' better?
The answer is we look at his value. This is why value matters, even for assessing pure performance. Because once we see Haaland's £300mil asking price, compare it to our £18,582 transfer budget, we quickly move on to the player below him - who is Joe Blow from Slough.
Keep in mind it's not at all saying that higher value = strictly better performance. What I'm instead saying is that if a player has a significantly higher value, it's highly unlikely that he is going to be worse. It's like if I buy a $200 toaster. Is it going to be worse than the old cheap one? Probably not. Is it going to toast better? Not necessarily.
The fundamental basis for this is CA vs attribute imbalance. Very old players have the worst imbalances (i.e. Dybala - ~9 physical, ~18 mental/technical) and often high rep, and therefore generally the worst over-valuations. Dybala is valued at £10mil. Now there are a few players who are significantly better for the same price or less, such as Maeda, but it's a minority. Now try the inverse, find me a player who is low value but very highly rated. Again, it's a minority.
Now you might say, well if value is such a reliable correlation, why not rely on that alone? The answer is that those players with £300mil valuations, even though they make up the majority of the top players, are useless to us. We are looking for the diamonds in the rough, even if we 'just want the best performing players right now', because you need to pay to get them! Hence valuations are an important/useful measurement for telling Haaland from Joe Blow, but NOT for ranking utility, and even if you did have the billions to spend, valuation does an inadequate job of distinguishing exceptions to the rule. It is an assessment and therefore indicator of quality itself rather than a contributor to quality, and so it can be used to critique a ranking but not determine it. In fact we want our ranking system to find as many exceptions to the valuation ranking as possible, so long as they are valid.
If you don't mind, I would like that save file, to have a closer look at it.
I was anxious about this test because I haven't done it myself since comparing Pure Performance & Premier League 1.0 with Orion's weights, and I thought it's 50/50 how it goes.
Now that it's done I'll make my comments and critiques, and thankfully at least my Pure Performance file topped out by a decent margin, so hopefully less chance I'll be seen as just being a sour puss.
I guess I'll start by saying now it can be seen that I wasn't just making up the numbers in my own comparison test, where the results were Pure Performance 6.5, Premier League 1.0 7.333, and Orion's weights 13.166. You can see the results are similar, particularly if you look at the individual samples (I didn't do as many samples).
Now, my critiques:
I'll start with an intuitive example. In keithb's file Maeda is rank 102 for FS, and in Premier League 2.1 he is rank 17. Regardless of which file you assume is better, if you were assessing based on this player alone there would be obviously be a great difference. The thing is, these exceptions to the rule constitute a minority of players, and so if you're picking a full squad, half are going to be identical players, and many of the rest are going to be minor differences like Mbappe vs. Haaland. In fact, it's a surprise there's any significant difference at all between HarvestGreen's/Orion's weights and Pure Performance for this reason.
Now if you're on board with that, then let's take it to the next matter. Because now I'm going to show that those exceptions matter, and I would argue you can't really just conclude 'well, if they're all about the same once you buy a full squad, who cares'.
Suppose I want a top striker for my team. I load up keithb's ratings, sort by FS, and find that the cheapest of the top 20 is Victor Gyokeres for a GS estimated sale value of £100mil. If I use Premier League 1.0, in the top 20 I have Furuhashi (£33.8mil), Maeda (£22.4mil), Jhon Cordoba (£8.5mil) and even a few more under the cheapest in keithb's ratings. If I load up FM24 now as Fulham and make an offer for Gyokeres, they want £41.5mil, whereas for Maeda it's £12.5mil, so it's not just GS. Pure Performance also has this problem, although to a lesser extent than keithb's file - the cheapest in the top 20 FS is Umar Sadiq for £40mil sale value (~£20mil in game).
Now who would argue money is no matter in FM?
So one thing I'm interested in seeing in your save, is the relative cost/value of each squad. I don't actually know if Premier League 2.1 would be significantly less expensive to purchase, but I suspect it would be.
Now the matter of performance. I have high confidence that Premier League 2.1 is going to give more accurate performance estimations than Pure Performance. I am confident because I have tested a player like Maeda, and I know with certainty he gets the team to 3rd, so he's definitely up there with the best. So how then does it finish worse here? Because these files should be comparing by performance alone, not cost - the low cost would just be a symptom of the improved efficiency.
I can think of a few possibilities here, but I'll just say the main two.
You 'froze' and 'healed' the players. Two of the biggest factors in Premier League 2.1 are dirtiness and injury proneness. If you prevent red cards, injuries, and so on, then you'd be disrupting a key part of the natural process. But I'm not sure exactly what the freeze/healing extends to.
The second thing is that while the goal of Premier League 2.1 is to assess performance rather than value, this does nonetheless loop back around.. let me demonstrate what I mean by this:
Your selections for ST are as follows:
Premier League 2.1 - Adam Zrelak, 113 CA, £2k/week, Warta Poznań (Poland Div 1)
Keithb - Alexis Sanchez, 137 CA, £78k/wk, Inter Milan
Pure Performance - Jair Silva, 112 CA, £1.2k/week, Leiria (Portugal Div 2)
Orion - Choupo-Moting, 137 CA, £123k/wk, FC Bayern
HarvestGreen - Dickson Abiama, 110 CA, £10k/wk, FC Kaiserslautern
If I've thought about this right, what is going on here is a problem with the methodology where because the better players are actually getting pushed up in Premier League 2.1 (i.e. Maeda 102 -> 16) this leaves what's left down at player 102 slightly weaker and yet still adequate, while also being more populated by even lower CA players who have been pushed up (i.e. Zrelak 300 > 100 say).
To complicate matters, this would only affect a minority of players, because people like Maeda and Zrelak are the exceptions not the rule, and for that majority remainder most of them are going to be the same selections (overlap) while one or two (differences) may even be better than Premier League 2.1. So for example if you selected players at rank 102, guess what, keithb gets Maeda and Premier League 2.1 gets someone like Zrelak. But actually Alexis Sanchez is higher rated in keithb's file than in Premier League 2.1. So it's overall its like I got Maeda/Kuruhashi/Cordoba right and Sanchez wrong, while keithb gets Maeda/Kuruhashi/Cordoba wrong, but gets Sanchez right. But you tell me what you think of this theory.
To put it another way, the file that would finish 1st in your test would have Haaland, Mbappe, etc. at the rank of 300 or whatever it is that you are selecting at. However it does nonetheless have use, because it tells us how many duds are pulling the team down. For example, keithb's team has an ST from Inter Milan, yet why does it still get beaten by Pure Performance on average? I reckon if we have a look at the file, Sanchez will get 7.3+ rating, but a few duds will get 6.5. Just my guess.
1. FMST26
2. FMSS
3. FM Genie Scout
Yes
FMSS Premier League 2.1 weights
Outfield:
const META_W = {
Pace: 20, Acceleration: 19.26, JumpingReach: 9.15, Dribbling: 3.04, Balance: 3.03,
Concentration: 9.57, Anticipation: 8.51, Determination: 9.25, Agility: 3.40, Stamina: 10.64,
Dirtiness: 6.8, Composure: 5.53, WorkRate: 14.15, Pressure: 3.62, InjuryProneness: -4.8,
};
GK:
const META_GK_W = { Determination: 20, Concentration: 18.53, Reflexes: 19.35, AerialReach: 6.45, Agility: 7.35, Pace: 6.37, Acceleration: 3.38, InjuryProneness: -11.03, Dirtiness: -3.68, FirstTouch: 3.83, WorkRate: 8.7, JumpingReach: 3.22, Stamina: 7.35, Technique: 3.53, Flair: 11.03, CommandOfArea: 3.53, Pressure: 3.38, Professionalism: 2.2, NaturalFitness: 2.35, };
I couldn't include every attribute, there's some limit occurring in FMSS, but it'll only be a few minor ones missing. I had to take off professionalism from outfield though. You could swap 'Balance: 3.03' for 'Professionalism: 3.19' if you like.
I've simplified dirtiness & injury proneness. I also can't verify the accuracy at all because it's FM26 only.
So what files are you comparing, and are you using FM24 or FM26?
Make sure you're using full detail, can be overlooked