FM22 Positional Filters – what are the best attributes for each position?

by Mark, Nov 29, 2021

Hope it'll help

The information exist on GS, but i think it's impossible to see it in columns.
It's my only request.

0

Falbravv said: Hope it'll help

The information exist on GS, but i think it's impossible to see it in columns.
It's my only request.


Yes, it is one of the fields you cant get access to. I can derive it down to about 10 out of 20. Beyond that I have nothing. If you are only interested in the players and their positions that have 10+ for positions I think I can go close. Otherwise I am sorry but I cant help

0

If you want to test something to go close, i promise you to use it :woot:

Thanks for the answer

0

Falbravv said: If you want to test something to go close, i promise you to use it :woot:

Thanks for the answer


I will have something for you in a couple of days.

1

Falbravv said: If you want to test something to go close, i promise you to use it :woot:

Thanks for the answer


Here is the formula that is very close to the mark. I only tested it against Wingers downloading data to Excel. In the formula A1 is the location of the rating you want to manipulate and B1 is the location of the position score out of 20. Down to about 10 out of 20 the derived score was within 0.1-0.2% of the general rating score. Below that the variance tends to blow out a fair bit.

=A1*(1-(20-B2)/46)/(1-(20-B1)/40)

I am sure the calculation could be simplified but I haven't had much time to work on it.

Update:

I now have it all within 0.1% and is working across all the positions.

Here is the updated formula, it seems like there are 2 cut off points, position score below 5, position score below 10:

=IF(B1<5,A1*(1-(20-B1)/55.5)/(1-(20-B1)/40),IF(B1<10,A1*(1-(20-B1)/48)/(1-(20-B1)/40),A1*(1-(20-B1)/46)/(1-(20-B1)/40)))

General Ratings calculations in Excel



Genie Scout General Ratings



0

Thanks a lot, i try it now !

0

Falbravv said: Hope it'll help

The information exist on GS, but i think it's impossible to see it in columns.
It's my only request.


I noticed you are managing Holstein Kiel. I play golf here in Australia with a mate who was born in Kiel, so I have played them a few times as well. Good fun getting them to Bundesliga and then challenging Bayern for the title. Good luck and let me know how you go.

1

Thanks a lot, i try it now !

Mark said: I noticed you are managing Holstein Kiel. I play golf here in Australia with a mate who was born in Kiel, so I have played them a few times as well. Good fun getting them to Bundesliga and then challenging Bayern for the title. Good luck and let me know how you go.

It's my actual save yes !

Promote and win B2 first year, challenge Bayern second season and finish 2nd, actually 2nd and 1/8 of CL

0

Falbravv said: Thanks a lot, i try it now !

Let me know how it goes mate. I am hoping you now have what you were after

0

In GS - when you club search - youre able to check the average rating % for the squad (first team).  But the ratings dont correspond - e.g Fylde in Vanarama North has average ratings of 62% (but the highest player for Zaz (50 minus 25) is 58.7%.

0

First of all good work! I love meta-analyses like this!
If you compare within a position, that's great! But if I want to find out which position a youth player should play in the future, the result is usually Winger or FB. For example, if I create a player with 20 in all attributes, a winger gets 120% and a FS gets 109%. Or do I just normalize the result to 100%? Is it even possible to decide in this way which position is the best for a player?

1

gazmlrs said: In GS - when you club search - youre able to check the average rating % for the squad (first team).  But the ratings dont correspond - e.g Fylde in Vanarama North has average ratings of 62% (but the highest player for Zaz (50 minus 25) is 58.7%.

If you go to the club list screen and select your division/league, and then double click on your side you will see the rating for your team using the current ratings file. It will initially be based on a 4-4-2 set up. To see the other sides in your league you can press the left or right arrow to scroll through. all the ratings will be based on the formation your have selected with the default being 4-4-2.

I tend to calculate my own ratings because the offset for non natural players is much lower in the GS calculations.

0

lordus said: First of all good work! I love meta-analyses like this!
If you compare within a position, that's great! But if I want to find out which position a youth player should play in the future, the result is usually Winger or FB. For example, if I create a player with 20 in all attributes, a winger gets 120% and a FS gets 109%. Or do I just normalize the result to 100%? Is it even possible to decide in this way which position is the best for a player?


That is an interesting question and one I have thought about a few times. I have meant to run some tests on this but have yet to get to it. Leave it with me and I will have a look in a few weeks.

0

Im finishing off a save on 2021 , do you think the ykykyky balanced will work on 2021 , or is there a disparage between the attributes needed ,, the file works on 2021 , but im guessing the attributes are based on 2022 , is there much difference i need to make it work for 2021 or should i forget using it for 2021

0

Mark said: That is an interesting question and one I have thought about a few times. I have meant to run some tests on this but have yet to get to it. Leave it with me and I will have a look in a few weeks.

You can probably normalize by multiplying the weight of all attributes for some X factor. I'll give an example using the case @lordus used:
"For example, if I create a player with 20 in all attributes, a winger gets 120% and a FS gets 109%."

In this case, X factor would be 1.0 for winger and 0.90 for fast striker (120% is 1.0, 109% is X, basic rule of three). Then you multiply all weights from winger by 0.9, and they will now have a normalized rating. You can do the same to all roles, based on the position with lowest rating (since all others will be reduced to reach similar level).

0

ZaZ said: You can probably normalize by multiplying the weight of all attributes for some X factor. I'll give an example using the case @lordus used:
"For example, if I create a player with 20 in all attributes, a winger gets 120% and a FS gets 109%."

In this case, X factor would be 1.0 for winger and 0.90 for fast striker (120% is 1.0, 109% is X, basic rule of three). Then you multiply all weights from winger by 0.9, and they will now have a normalized rating. You can do the same to all roles, based on the position with lowest rating (since all others will be reduced to reach similar level).


I am new to this... What exactly does the weight coefficient do in Genie Scout?
The values ​​there are exactly the values ​​you get when generating a player with 20 in all attributes for a specific position. For example FS is 109% and the weight in Genie Scout is 109 (based on "ykykyk balanced" ).

0

Rhumble said: Im finishing off a save on 2021 , do you think the ykykyky balanced will work on 2021 , or is there a disparage between the attributes needed ,, the file works on 2021 , but im guessing the attributes are based on 2022 , is there much difference i need to make it work for 2021 or should i forget using it for 2021

You definitely use the ratings in 2021 version. The changes each year are quite minimal. Pace and Acceleration have always had the biggest impact, but some positions more than others.

1

lordus said: I am new to this... What exactly does the weight coefficient do in Genie Scout?
The values ​​there are exactly the values ​​you get when generating a player with 20 in all attributes for a specific position. For example FS is 109% and the weight in Genie Scout is 109 (based on "ykykyk balanced" ).


The weighting eg 109% for FS is the balancing aspect that allows you to compare players across positions. Otherwise all the top players would be from one position. You can strip the weightings out if you want and just make all positions 100%. I would suggest you save that rating with a different name in case you don't like the results.

0

Mark said: The weighting eg 109% for FS is the balancing aspect that allows you to compare players across positions. Otherwise all the top players would be from one position. You can strip the weightings out if you want and just make all positions 100%. I would suggest you save that rating with a different name in case you don't like the results.

Indeed but i think there is a problem with weightings because in my save there is no one player who the best position in "General Rating" is FS or TS. The most player are better in FB, WB or W.

0

dzek said: Indeed but i think there is a problem with weightings because in my save there is no one player who the best position in "General Rating" is FS or TS. The most player are better in FB, WB or W.

You can adjust the weightings for positions to offset this. I tried around a dozen different saves and averaged it. It is not perfect, so you can adjust as you see fit

0

Mark said: You can adjust the weightings for positions to offset this. I tried around a dozen different saves and averaged it. It is not perfect, so you can adjust as you see fit

I will ask you if you can try and average it as it is in the game. I mean lets say "Ilaix Moriba" Has natural position as AMC but his attributes in GS says that it will be better in RB by far and with many more players like Haaland from striker to AMC etc.

0

dzek said: I will ask you if you can try and average it as it is in the game. I mean lets say "Ilaix Moriba" Has natural position as AMC but his attributes in GS says that it will be better in RB by far and with many more players like Haaland from striker to AMC etc.

Cant say I have ever worried about the General Rating. It takes too long to retrain people from scratch. It could just be minor errors in the balancing I did that throws that up. It is not exact as you cannot use partial percentages and attribute weightings are limited to whole numbers and cant go above 100. So there are major limitations on getting it exact. I rely more on the Positional Ratings. At least there they have some existing ability that can be built upon. I doubt the attributes are identical but in my game Moriba would be best as a central midfielder, MC, DM, AMC and maybe left winger. I wouldn't try outside of that. He could fill in as a right winger but I would do that sparingly.

1

Mark said: Cant say I have ever worried about the General Rating. It takes too long to retrain people from scratch. It could just be minor errors in the balancing I did that throws that up. It is not exact as you cannot use partial percentages and attribute weightings are limited to whole numbers and cant go above 100. So there are major limitations on getting it exact. I rely more on the Positional Ratings. At least there they have some existing ability that can be built upon. I doubt the attributes are identical but in my game Moriba would be best as a central midfielder, MC, DM, AMC and maybe left winger. I wouldn't try outside of that. He could fill in as a right winger but I would do that sparingly.

I would like to add that, in my experiments, having a player natural in two positions will make it use the attribute weights of the most expensive position (you can test that on Editor checking CA). That means training players in other positions will hinder the potential of the player, lowering his maximum potential attributes. As an example, training a fast defender as winger will make his CA increase, since acceleration and pace is more "expensive" for wingers than for defenders.

1

Mark said: Cant say I have ever worried about the General Rating. It takes too long to retrain people from scratch. It could just be minor errors in the balancing I did that throws that up. It is not exact as you cannot use partial percentages and attribute weightings are limited to whole numbers and cant go above 100. So there are major limitations on getting it exact. I rely more on the Positional Ratings. At least there they have some existing ability that can be built upon. I doubt the attributes are identical but in my game Moriba would be best as a central midfielder, MC, DM, AMC and maybe left winger. I wouldn't try outside of that. He could fill in as a right winger but I would do that sparingly.

So if i have a player lets say who has natural position as AMC and his positional rating is 85% and in General Rating, which is like the positional rating without the position familiarity included, has 95% in lets say CB you still stick with the real natural position and not try to retrain him in other position to get the best of him?

Btw im a player of FM since the first game and i use GS from 2013. I speak with many guys creating tactics, training schedules and in general FM essentials and i want to create a new program like GS because i think Eugene dont update it anymore. He just update the program to read the new patch of the game and fix some bugs. There is a lot potential in his program but i think he doesnt care anymore, i dont know the reasons. If anyone can help me with this i will be glad to start this project. My biggest problem is that i dont know how to create a program to read the encrypted save game of the game.

Thank you!

0

Not sure if its been mentioned before , what does Genie Scout take into account when it gives its percentage ratings for positions , does it calculate whether a player is natural , accomplished etc... and what footed a player is , or is it purely on the attributes and weighting.

Ive tried adding up the main attributes and getting an average of ykykyk Balanced and i also look at pace and acceleration ,, the ones i work out on my own who i think will be a better fit for a position, never seem to be the players with the highest percentage for that position on Genie Scout,  is it better to use Genie scouts position percentage for finding a player for a position , or just use quickness and the main attributes that are weighted 100 - 60

0

Rhumble said: Not sure if its been mentioned before , what does Genie Scout take into account when it gives its percentage ratings for positions , does it calculate whether a player is natural , accomplished etc... and what footed a player is , or is it purely on the attributes and weighting.

Ive tried adding up the main attributes and getting an average of ykykyk Balanced and i also look at pace and acceleration ,, the ones i work out on my own who i think will be a better fit for a position, never seem to be the players with the highest percentage for that position on Genie Scout,  is it better to use Genie scouts position percentage for finding a player for a position , or just use quickness and the main attributes that are weighted 100 - 60


Look at post #185 for positional offsets. I correct this back to normal and then deduct 2.5% offset for each point below 20 the player is in that position.

I do my own calculations after using either ykykyky or my own ratings. Both seem to work really well both for selection purposes in my squad and analysing the players my scouts find. You cant go wrong using players with high acceleration, and using the higher rated attributes from the machine learning ratings will always get you someone worthwhile. You can do as much work or as little as you feel comfortable with.

Also take into account injury proneness, professionalism and how players will fit into the club dynamics.  The main thing is that you enjoy yourself.

2

Mark said: Look at post #185 for positional offsets. I correct this back to normal and then deduct 2.5% offset for each point below 20 the player is in that position.

I do my own calculations after using either ykykyky or my own ratings. Both seem to work really well both for selection purposes in my squad and analysing the players my scouts find. You cant go wrong using players with high acceleration, and using the higher rated attributes from the machine learning ratings will always get you someone worthwhile. You can do as much work or as little as you feel comfortable with.

Also take into account injury proneness, professionalism and how players will fit into the club dynamics.  The main thing is that you enjoy yourself.


Since people keep coming back and forth here, you should probably add the frequently added questions to the opening post.

1

Mark said: Look at post #185 for positional offsets. I correct this back to normal and then deduct 2.5% offset for each point below 20 the player is in that position.

I do my own calculations after using either ykykyky or my own ratings. Both seem to work really well both for selection purposes in my squad and analysing the players my scouts find. You cant go wrong using players with high acceleration, and using the higher rated attributes from the machine learning ratings will always get you someone worthwhile. You can do as much work or as little as you feel comfortable with.

Also take into account injury proneness, professionalism and how players will fit into the club dynamics.  The main thing is that you enjoy yourself.


Thanks bud , ive been trying to read the forum on my phone away from the TV  , which ive now realised is useless to try and take it all in , reading it from the start on a tablet ive bought i can see ive missed bits , you essentially answered some of my question in the first post so thats my bad , i will have a proper read through now as i can see better on a tablet.

Im sort of understanding it better , i dare say i might have other questions as this fascinates me ,, i can play around with finding players using this and putting the squad together with a lot more enthusiasm and enjoyment than i can with whats in the game, i honestly think SI should start adding things like this to the game as it should be up to the player how they want to play it , i find their system is getting a bit old fashioned now , they give attributes for a role , then say these are only a guide , so essentially the game gives you attributes that may or may not be needed for a role that has not nearly enough significance on the ME , its all about position , i remember Knap saying it on a discord 3 years ago , so i never use the in game system to find players.

I have to thank you and everyone who works hard to make the game more interesting , some would say this is "cheating" like they do with Genie scout ,,, but all we have to judge a player are attributes , how can we use the in game system when SI fool us into believing the roles mean something and they dont give you the right attributes anyway , they are almost pulling the wool over our eyes.

As much as i love playing Football manager, i think SI are flogging a very outdated game that really should have been brought up to date years ago , they really need to start reading forums to understand how we like to play the game and start adapting it better, thats just my opinion though.

A huge thanks to you and Zaz and everyone who worked on all these formulas etc... i think i would have stopped after 2020 ,, im still finishing FM21 as i got to the threads late , i will start FM 22 soon and im looking forward to it more than i would have been.

3

Mark said: You have 4 strikers in your side and Huanca would be your 5th. I have run the ratings against all 5 strikers using each of the ratings files. I would be tempted to buy him. He doesn't rate well in my MDW ratings file but goes well in the Y50 and even the ykykyk files.

I think from the analysis I have, done that the machine learning works. I personally will be using the full ykykyk ratings file as I think it will be more stable longer term. I will now balance that ratings file and derive the unrated positions as best I can.

@ZaZ  prefers the chopped down versions because you get better value players. I prefer to try and get players you can keep as you progress. And I certainly move between the approaches and value what ZaZ does immensely.

My focus beyond changing my ratings file is to upgrade my approach to individual training focus. There were some notes in the machine learning example relating to that. I am exploring how I improve on that aspect in my games. At some stage I will provide my learnings on that front.


@Mark Why did you come to the choice of using the ykykyk file? Didn't the ZaZ under minus 25 score higher when you did your test? I understand that ZaZs gives more value but why do you think ykykyk is more stable longer term?

0

Middleweight165 said: @Mark Why did you come to the choice of using the ykykyk file? Didn't the ZaZ under minus 25 score higher when you did your test? I understand that ZaZs gives more value but why do you think ykykyk is more stable longer term?

Testing wasn't a huge sample. I think if you believe in the machine learning work then you go with their total results. I had a lot of faith in my original ratings and these seem to be slightly better. So I continue to use the ykykyky one successfully.

1
Create an account or log in to leave a comment