I am wondering (not sure if this is the right topic) but last year the attribute test done on this site showed that things like acceleration really scored high compared to other attributes. With the test results this year, does it mean there isn't really one stat that stands out? And acceleration seems weakest now? What does that mean for individual training in general? Last year I put everyone on acceleration lol (https://fm-arena.com/table/13-fm22-attributes-ratings/)
DoubleR said: I am wondering (not sure if this is the right topic) but last year the attribute test done on this site showed that things like acceleration really scored high compared to other attributes. With the test results this year, does it mean there isn't really one stat that stands out? And acceleration seems weakest now? What does that mean for individual training in general? Last year I put everyone on acceleration lol (https://fm-arena.com/table/13-fm22-attributes-ratings/) Expand
It's the other way around. Lower scores mean attributes had higher impact in performance. Acceleration and Pace are still the most important.
ZaZ said: It's the other way around. Lower scores mean attributes had higher impact in performance. Acceleration and Pace are still the most important. Expand
Looking at and testing through both ratings files from @Mark and @ZaZ I'd say if you are looking for most "bang for your buck" and maybe not a lot of money I'd use the one from Zaz as it strictly looks at the attributes table and if you have no money you can squeeze in pretty much any MFer that has important atts and they will do well. Personally as I go up and have any money to spend on players I like Mark's better as it still prioritizes the important attributes, however truly seems to differentiate the top players (assuming other stats DO matter somewhat)
Edit: no disrespect to either - both your ratings files are definitely better than the standard rating file, just my opinion on them. Your mileage may vary
saycarramrod said: Looking at and testing through both ratings files from @Mark and @ZaZ I'd say if you are looking for most "bang for your buck" and maybe not a lot of money I'd use the one from Zaz as it strictly looks at the attributes table and if you have no money you can squeeze in pretty much any MFer that has important atts and they will do well. Personally as I go up and have any money to spend on players I like Mark's better as it still prioritizes the important attributes, however truly seems to differentiate the top players (assuming other stats DO matter somewhat)
Edit: no disrespect to either - both your ratings files are definitely better than the standard rating file, just my opinion on them. Your mileage may vary Expand
Yeah, mine had exactly that mentality that the more attributes you consider important, the more expensive the player will be. Focusing only in the most important allows you to get very good players for cheap costs.
Thank you ZaZ and Mark for all your work! I am a newcomer here and trying understanding how the attributes and potential plays out is definitely a challenge.
I already downloaded Mark's filter and I would like to compare them with yours ZaZ but the WeTransfer link expired I would be very grateful if you could upload it again. I am playing mostly in lower league so it may be better to find cheap players.
I was also wondering if you guys ever tried playing with negative percentages in the Genie rating for the expensive attributes with little effects? I vaguely remember that negative percentages worked in previous versions but I am not sure.
ZaZ said: Yeah, mine had exactly that mentality that the more attributes you consider important, the more expensive the player will be. Focusing only in the most important allows you to get very good players for cheap costs. Expand
Mnj said: Thank you ZaZ and Mark for all your work! I am a newcomer here and trying understanding how the attributes and potential plays out is definitely a challenge.
I already downloaded Mark's filter and I would like to compare them with yours ZaZ but the WeTransfer link expired I would be very grateful if you could upload it again. I am playing mostly in lower league so it may be better to find cheap players.
I was also wondering if you guys ever tried playing with negative percentages in the Genie rating for the expensive attributes with little effects? I vaguely remember that negative percentages worked in previous versions but I am not sure. Expand
I cant say I have ever tried using negative percentages.
Mnj said: Thank you ZaZ and Mark for all your work! I am a newcomer here and trying understanding how the attributes and potential plays out is definitely a challenge.
I already downloaded Mark's filter and I would like to compare them with yours ZaZ but the WeTransfer link expired I would be very grateful if you could upload it again. I am playing mostly in lower league so it may be better to find cheap players.
I was also wondering if you guys ever tried playing with negative percentages in the Genie rating for the expensive attributes with little effects? I vaguely remember that negative percentages worked in previous versions but I am not sure. Expand
About negative values, I've used to a couple of attributes, like injury proneness. I was thinking about using it for undesired attributes, but then it would be a metric for player potential, and not for player general ability.
@Mark@ZaZ Have either of you looked at the Genie Role ratings while using either of your rating files? Seems that it DOES change the underlying positional ratings you have modified, while also giving additional percentage points to the critical role attributes. Just wondering if that helps to even further differentiate better players, although I understand ZaZ's point of "wasting" CA on useless attributes.
saycarramrod said: @Mark@ZaZ Have either of you looked at the Genie Role ratings while using either of your rating files? Seems that it DOES change the underlying positional ratings you have modified, while also giving additional percentage points to the critical role attributes. Just wondering if that helps to even further differentiate better players, although I understand ZaZ's point of "wasting" CA on useless attributes. Expand
I'm not sure what you mean. Our files are genie role ratings. If I remember well, Mark's version considers the standard ratings of FM Genie Scout in his calculations.
I'm sorry if this seems as a bit of a noob question. So I downloaded genie scout, uploaded the ratings and everything. My question is, how do you find players for specific tactics, for instance yours @ZaZ
venomenom said: I'm sorry if this seems as a bit of a noob question. So I downloaded genie scout, uploaded the ratings and everything. My question is, how do you find players for specific tactics, for instance yours @ZaZ Expand
Genie Scout will just show the current rating and potential rating of all players. You need to filter the ones that match what you are searching, like age, position, sale price and minimum current rating, then press enter. After that, you need to talk to them inside the game and see if they are available.
By the way, you need to change the rating from default to the ones we provided, clicking in the start button in the top left.
venomenom said: I'm sorry if this seems as a bit of a noob question. So I downloaded genie scout, uploaded the ratings and everything. My question is, how do you find players for specific tactics, for instance yours @ZaZ Expand
I use this view, you can then see the positional ability and positional ratings. You need positional ability above 13 out of 20 and the best positional rating you can afford.
Below is the Google translation of the text from the site but I will also include at the top a summary of their data against the ZaZ Blue DM tactic by @ZaZ.
I now intend to produce a Genie Scout rating based on the summary data above. I will share it in the next few days for anyone that want s to test it.
This does appear to loosely align with but build on some of the work already discussed in this forum. Here is the text from the PlayGM web site.
[Player experience] Share player attribute utility calculated based on deep neural network
This post was last edited by C·Gunners at 2021-12-30 14:36
Basic introduction: 1. We are a basketball manager game development team. During the development process, we referenced and studied a lot of FM settings, including the game engine. I learned a lot of game knowledge from the bursting room. The purpose of this post is to give back to the community.
2. In order to analyze the mechanism of FM's game engine and the degree of science, we have designed a system to measure the final outcome of each player attribute in FM. The degree of influence.
3. fm-arena.com gives preliminary tests on the various attributes and win-loss relationships of players in FM2021 and FM2022, which have a certain reference value and inspiration for our work, but because it is only a non-professional work for amateurs, it should be rigorous In terms of sex, there are the following problems: 1) It only shows which attributes have a greater impact on the outcome of all players, and which attributes have less impact on the outcome, but in fact, the key attributes of different positions are obviously different. Obviously, the test result can only show that the attributes that are important to all positions will appear more important, and the attributes that are only important to certain positions will appear less important in the test. For example, the test results have almost no impact on the outcome of shooting, but explosive power has a great impact on the outcome. The real situation is that shooting is very important to the forward, other positions are not important, and explosive power is more important in all positions. 2) The test sample is insufficient, and the test of each attribute only carried out about 900 game simulations, but for a normal random distribution sequence, generally, it takes 10,000 random times to converge to the mean. 3) The influence of attribute on the outcome is non-linear. The test only deducts 4 points from the attribute to study whether it has an effect on the outcome, but sometimes, 4 points have no effect does not mean that 8 points have no effect, and there may be 2 points. It is close to 4 o'clock. 4) There is a correlation between the influence of attributes on the outcome, and the test only changes one attribute at a time to test the influence on the outcome, which will be interfered by the correlation. For example, a breakthrough can speed up when lying on the ball, relying on physical fitness instead of dribbling, or relying on dribbling. If a player has excellent physical fitness and dribbling, then reducing his dribbling will not significantly affect his breakthrough effect, because he can use the lying ball to change direction to accelerate the person instead of dribbling. 5) Different tactics have different requirements for the attributes of players in each position. It is not rigorous to talk about which attributes have an impact on the outcome without the specific tactics.
4. We use artificial intelligence to study the importance of attributes. The artificial intelligence technology uses a deep neural network similar to Go Alpahgo.
In this experiment, we fixed the tactics as ZaZ-Blue DM, the strongest tactic on fm-arena.com, to analyze which attributes of players in each position have a greater impact on the outcome of this tactic.
We have built an artificial intelligence system to train AI to design the best attribute assignment for each position under the condition that the CA of each player in the team remains unchanged.
The general process is as follows. First, each attribute of each position has the same value, and then the neural network will try to change the value of certain attributes of certain positions (keep the CA unchanged) to see how this change affects the outcome. Positive or negative, in order to continue to iterate on the multi-layer network, knowing that iteratively produces the best 11-person combination that AI thinks.
We trained about 140 machines for 3 weeks, simulated 40 million games, and got convergence results.
5. In order to verify whether the best 11-person attribute assignment given by AI is really the best, we conducted a result verification test to design three groups of teams, one team’s player attributes are evenly distributed, the other group It is the best 11-person attribute allocation scheme given manually by senior players based on their game experience. The last group is given by AI. It was put into a test league for 100,000 matches. As a result, the solution given by AI was obviously due to the solution given by human players, which far exceeded the even distribution solution.
in conclusion: For ZaZ-Blue DM tactics, the degree of importance of the attributes we get for each position is as follows (in order to facilitate the player to view, we have processed the value to 5, for the game, this accuracy should be sufficient):
GK: 40 strikes, talent 20 Explosive power 70 passes 45 defensive positions 40 bounce 45 kick-off 35 work commitment 10 flexibility 100 reaction 80 concentration 65 endurance 10 interception passes 40 decision 50 balance 20 boxing 0 leadership 10 strong 70 hand control 50 aggression 40 Speed 50 Throwing the ball 30 Field of view 40 Physique 10Stopping 30 Teamwork 20 Unused feet 10 One-on-one 45 Running without the ball 0 Unexpectedness 20 Willpower 20 Command and Defense 30 Brave 30 Air Dominance 60 Pre-judgment 40 Composure 40
Usage Significance: 1. Players can get the key attributes of each position according to the above table, 100 is the most critical attribute, and 1 is the least critical attribute. So as to guide the selection of materials.
2. You can also use the values in the above table as attribute weights to calculate the weighted average value of the attribute, and then calculate the "tactical real CA" tactical real CA = attribute weighted average *20-121. Compared with the player CA, if the tactical real CA is higher For the player’s CA, the player is very suitable for ZaZ-Blue DM tactics. The higher the tactical real CA, the more suitable it is for player selection.
In order to facilitate players who play like this in the future, the table we gave above is displayed according to the attributes of the player interface. Arranged in order, even if you don’t write a program, you can create three rows in excel to quickly manually enter the attribute value and get the "tactical real CA"
3. The guiding significance for training, the attribute is occupied by the CA, in order to cultivate the tactical real CA For the highest player, we can cultivate the most cost-effective attribute. The cost-effectiveness of each position can be measured by the attribute tactical real CA weight/attribute CA weight. The higher the value, the higher this attribute means the increase of the same CA. Will make the player's attribute tactical real CA higher.
Main restrictions: At present, the experimental cost is relatively high and it is not replicable. The engine version is changed and the tactics are changed. If there is no transfer learning, the AI needs to be retrained. Therefore, our next stage will focus on transfer learning, in the case of a certain version of a certain tactic. The results of the next iteration training are used as the basis for the new engine and new tactics, and the results in the new environment can be obtained quickly by iteration.
wow lots of info, hard to understand everything but in excel sheet but as i understand it shows that for example composure is most important for CD(80) and DM(60) and for rest of positions are not so important?
Wigo said: wow lots of info, hard to understand everything but in excel sheet but as i understand it shows that for example composure is most important for CD(80) and DM(60) and for rest of positions are not so important? Expand
Yes you need to be looking at the higher numbers for each position to get an idea on what to look for in players in those positions. Interesting work I think
This is the FM Genie Scout filter based on the study of ykykyk05251 for best attributes for each position of Blue 3.0 RPM (aka Blue DM). The study used machine learning to get the weights.
Keep in mind that two attributes of the table have ranges from -100 to 0 in FM Genie Scout instead of 0 to 100: eccentricity and punching tendency. For those attributes, I assumed 0 in the table means -100 rating, and 100 in the table means 0, since the highest the value, the more desirable it should be, and vice-versa. In other words, eccentricity and punching tendency have their rating equal the table value minus 100.
ZaZ said: This is the FM Genie Scout filter based on the study of ykykyk05251 for best attributes for each position of Blue 3.0 RPM (aka Blue DM). The study used machine learning to get the weights.
Keep in mind that two attributes of the table have ranges from -100 to 0 in FM Genie Scout instead of 0 to 100: eccentricity and punching tendency. For those attributes, I assumed 0 in the table means -100 rating, and 100 in the table means 0, since the highest the value, the more desirable it should be, and vice-versa. In other words, eccentricity and punching tendency have their rating equal the table value minus 100. Expand
I think this needs a little work to balance out the ratings. As it sits it favours Strikers and AMs too much and there are no GKs anywhere to be seen. The weightings have only been developed by position and not across all positions so need adjusting in terms of weight for each position. I would also like to derive the positions not covered.
I will have a play and update it next week. I have a very busy weekend unfortunately.
Mark said: I think this needs a little work to balance out the ratings. As it sits it favours Strikers and AMs too much and there are no GKs anywhere to be seen. The weightings have only been developed by position and not across all positions so need adjusting in terms of weight for each position. I would also like to derive the positions not covered.
I will have a play and update it next week. I have a very busy weekend unfortunately. Expand
I'll do my poor man's version too, with only the most important attributes, so it puts more importance on those attributes and gives good cheaper players.
@Mark, I made three poor man's version and I would like to know your opinion in which one gives the best results. Can you test the three and tell me which one you like the most? Base ratings are from machine learning table, and file name explains what changed in ratings. All of them try to focus on most important attributes in some way.
Version 1: all ratings minus 25 (removes lower ratings and increases % gap of remaining) Version 2: all ratings under 50 minuts 25 (removes lower ratings and keeps gap of top) Version 3: all ratings minus difference to 100 (removes lower ratings and increases absolute gap of remaining) * Gap means the difference in rating between two attributes. For example, 25 to 40 have a gap of 15.
ZaZ said: @Mark, I made three poor man's version and I would like to know your opinion in which one gives the best results. Can you test the three and tell me which one you like the most? Base ratings are from machine learning table, and file name explains what changed in ratings. All of them try to focus on most important attributes in some way.
Version 1: all ratings minus 25 (removes lower ratings and increases % gap of remaining) Version 2: all ratings under 50 minuts 25 (removes lower ratings and keeps gap of top) Version 3: all ratings minus difference to 100 (removes lower ratings and increases absolute gap of remaining) * Gap means the difference in rating between two attributes. For example, 25 to 40 have a gap of 15.
Anyone else's opinion would also be appreciated! Expand
OK so I had a quick play with the ratings files. The first observation is that you need to balance the ratings files. I learned this from last years attempt. What do I mean by this, well have a look at the screenshot below. There should not be major discrepancies between positions ie you should be getting most positions covered in the top 50 or so top players.
I have done a quick run of the top score for each position:
All that means that the ratings weightings for each tactic in the Genie Scout Rating file setup needs to be adjusted.
Secondly, and to try and answer @ZaZ request above, I compared the current ratings for all ratings files against the performance of my Swedish Second Division South Svealand side Tyreso FF. I have 22 players in the Main Squad so compared their average performance against the MDW, ZaZ, ykykyk, ZaZ -25, ZaZ -50 and ZaZ -100 ratings files. I have only used ratings for the position selected.
Here are the results:
Disclaimer: I have used players out of the positions identified in the results, but in the main they have played the positions indicated. In addition I have used 2 main tactics, Dragons Breath that uses an MC instead of AMC, and ZaZ Blue DM, but given the ratings files don't all have MC ratings I just used AMC.
To answer @ZaZ question the ZaZ-100 ended up equal with MDW in picking 8 of the best 11 positional players. It does need balancing to make it more usable though. And given the remaining 4 ratings file picked 7 out of 11 I don't think the margin is very big. All are definitely usable.
Now if we look at average % difference from the best rated team the order is MDW, ykykyk, ZaZ -50, ZaZ -25, ZaZ, ZaZ -100.
Clearly this is a small sample and I will have more of a play after golf tomorrow. But I think if you wanted to use one of the ratings files above, I would go with the ZaZ -50.
Mark said: OK so I had a quick play with the ratings files. The first observation is that you need to balance the ratings files. I learned this from last years attempt. What do I mean by this, well have a look at the screenshot below. There should not be major discrepancies between positions ie you should be getting most positions covered in the top 50 or so top players.
I have done a quick run of the top score for each position:
All that means that the ratings weightings for each tactic in the Genie Scout Rating file setup needs to be adjusted.
Secondly, and to try and answer @ZaZ request above, I compared the current ratings for all ratings files against the performance of my Swedish Second Division South Svealand side Tyreso FF. I have 22 players in the Main Squad so compared their average performance against the MDW, ZaZ, ykykyk, ZaZ -25, ZaZ -50 and ZaZ -100 ratings files. I have only used ratings for the position selected.
Here are the results:
Disclaimer: I have used players out of the positions identified in the results, but in the main they have played the positions indicated. In addition I have used 2 main tactics, Dragons Breath that uses an MC instead of AMC, and ZaZ Blue DM, but given the ratings files don't all have MC ratings I just used AMC.
To answer @ZaZ question the ZaZ-100 ended up equal with MDW in picking 8 of the best 11 positional players. It does need balancing to make it more usable though. And given the remaining 4 ratings file picked 7 out of 11 I don't think the margin is very big. All are definitely usable.
Now if we look at average % difference from the best rated team the order is MDW, ykykyk, ZaZ -50, ZaZ -25, ZaZ, ZaZ -100.
Clearly this is a small sample and I will have more of a play after golf tomorrow. But I think if you wanted to use one of the ratings files above, I would go with the ZaZ -50. Expand
Thanks for the thorough analysis, Mark. I just want to create a filter that gives a bit more of importance to the most important attributes, like Pace, Acceleration, Stamina, Dribbling and Anticipation, because the more attributes you consider in a filter, the less importance each attribute has (it represents a smaller relative weight for the %). Doing that, it allows you to find good players for cheaper prince, since cost of players is directly linked to CA and reputation.
guys what you think of that way? i made filters for every position... first i add all skill attributes of every position what is required... for example for advanced forvard i add all skills what is marked in game as important atributes... then i look at "test excel sheet" and then adjust those skills how it shows like finishing 8, dribbling 8,acceleration 10 and so on... you can add + or - in filter search depending on your strong of team... i am testing now and it works incredibble! i mean i am finding cheap players who are performing amazing! for example check shane lavery in blackpool! he can be a beast according to those attributes for very cheap... i can upload my filters for you to check as a fm files... hope you get it... feel still kind off dizzy after new years celebration
Wigo said: guys what you think of that way? i made filters for every position... first i add all skill attributes of every position what is required... for example for advanced forvard i add all skills what is marked in game as important atributes... then i look at "test excel sheet" and then adjust those skills how it shows like finishing 8, dribbling 8,acceleration 10 and so on... you can add + or - in filter search depending on your strong of team... i am testing now and it works incredibble! i mean i am finding cheap players who are performing amazing! for example check shane lavery in blackpool! he can be a beast according to those attributes for very cheap... i can upload my filters for you to check as a fm files... hope you get it... feel still kind off dizzy after new years celebration Expand
JohnSariman said: is this on fmscout?
That image is available in FM Scout, based on weights from FM Editor.
ZaZ said: That image is available in FM Scout, based on weights from FM Editor.
thank you @ZaZ !
SantosFC said: Download copy. Please
Here is my copy of the original default ratings file mate
Default Ratings File
I am wondering (not sure if this is the right topic) but last year the attribute test done on this site showed that things like acceleration really scored high compared to other attributes. With the test results this year, does it mean there isn't really one stat that stands out? And acceleration seems weakest now? What does that mean for individual training in general? Last year I put everyone on acceleration lol (https://fm-arena.com/table/13-fm22-attributes-ratings/)
DoubleR said: I am wondering (not sure if this is the right topic) but last year the attribute test done on this site showed that things like acceleration really scored high compared to other attributes. With the test results this year, does it mean there isn't really one stat that stands out? And acceleration seems weakest now? What does that mean for individual training in general? Last year I put everyone on acceleration lol (https://fm-arena.com/table/13-fm22-attributes-ratings/)
It's the other way around. Lower scores mean attributes had higher impact in performance. Acceleration and Pace are still the most important.
ZaZ said: It's the other way around. Lower scores mean attributes had higher impact in performance. Acceleration and Pace are still the most important.
Omg i'm so stupid.. ill cry in a corner now
Looking at and testing through both ratings files from @Mark and @ZaZ I'd say if you are looking for most "bang for your buck" and maybe not a lot of money I'd use the one from Zaz as it strictly looks at the attributes table and if you have no money you can squeeze in pretty much any MFer that has important atts and they will do well. Personally as I go up and have any money to spend on players I like Mark's better as it still prioritizes the important attributes, however truly seems to differentiate the top players (assuming other stats DO matter somewhat)
Edit: no disrespect to either - both your ratings files are definitely better than the standard rating file, just my opinion on them. Your mileage may vary
saycarramrod said: Looking at and testing through both ratings files from @Mark and @ZaZ I'd say if you are looking for most "bang for your buck" and maybe not a lot of money I'd use the one from Zaz as it strictly looks at the attributes table and if you have no money you can squeeze in pretty much any MFer that has important atts and they will do well. Personally as I go up and have any money to spend on players I like Mark's better as it still prioritizes the important attributes, however truly seems to differentiate the top players (assuming other stats DO matter somewhat)
Edit: no disrespect to either - both your ratings files are definitely better than the standard rating file, just my opinion on them. Your mileage may vary
Yeah, mine had exactly that mentality that the more attributes you consider important, the more expensive the player will be. Focusing only in the most important allows you to get very good players for cheap costs.
Thank you ZaZ and Mark for all your work! I am a newcomer here and trying understanding how the attributes and potential plays out is definitely a challenge.
I already downloaded Mark's filter and I would like to compare them with yours ZaZ but the WeTransfer link expired I would be very grateful if you could upload it again. I am playing mostly in lower league so it may be better to find cheap players.
I was also wondering if you guys ever tried playing with negative percentages in the Genie rating for the expensive attributes with little effects? I vaguely remember that negative percentages worked in previous versions but I am not sure.
ZaZ said: Yeah, mine had exactly that mentality that the more attributes you consider important, the more expensive the player will be. Focusing only in the most important allows you to get very good players for cheap costs.
Mnj said: Thank you ZaZ and Mark for all your work! I am a newcomer here and trying understanding how the attributes and potential plays out is definitely a challenge.
I already downloaded Mark's filter and I would like to compare them with yours ZaZ but the WeTransfer link expired I would be very grateful if you could upload it again. I am playing mostly in lower league so it may be better to find cheap players.
I was also wondering if you guys ever tried playing with negative percentages in the Genie rating for the expensive attributes with little effects? I vaguely remember that negative percentages worked in previous versions but I am not sure.
I cant say I have ever tried using negative percentages.
Mnj said: Thank you ZaZ and Mark for all your work! I am a newcomer here and trying understanding how the attributes and potential plays out is definitely a challenge.
I already downloaded Mark's filter and I would like to compare them with yours ZaZ but the WeTransfer link expired I would be very grateful if you could upload it again. I am playing mostly in lower league so it may be better to find cheap players.
I was also wondering if you guys ever tried playing with negative percentages in the Genie rating for the expensive attributes with little effects? I vaguely remember that negative percentages worked in previous versions but I am not sure.
New link here.
About negative values, I've used to a couple of attributes, like injury proneness. I was thinking about using it for undesired attributes, but then it would be a metric for player potential, and not for player general ability.
nice
@Mark @ZaZ Have either of you looked at the Genie Role ratings while using either of your rating files? Seems that it DOES change the underlying positional ratings you have modified, while also giving additional percentage points to the critical role attributes. Just wondering if that helps to even further differentiate better players, although I understand ZaZ's point of "wasting" CA on useless attributes.
saycarramrod said: @Mark @ZaZ Have either of you looked at the Genie Role ratings while using either of your rating files? Seems that it DOES change the underlying positional ratings you have modified, while also giving additional percentage points to the critical role attributes. Just wondering if that helps to even further differentiate better players, although I understand ZaZ's point of "wasting" CA on useless attributes.
I'm not sure what you mean. Our files are genie role ratings. If I remember well, Mark's version considers the standard ratings of FM Genie Scout in his calculations.
I'm sorry if this seems as a bit of a noob question. So I downloaded genie scout, uploaded the ratings and everything. My question is, how do you find players for specific tactics, for instance yours @ZaZ
venomenom said: I'm sorry if this seems as a bit of a noob question. So I downloaded genie scout, uploaded the ratings and everything. My question is, how do you find players for specific tactics, for instance yours @ZaZ
Genie Scout will just show the current rating and potential rating of all players. You need to filter the ones that match what you are searching, like age, position, sale price and minimum current rating, then press enter. After that, you need to talk to them inside the game and see if they are available.
By the way, you need to change the rating from default to the ones we provided, clicking in the start button in the top left.
venomenom said: I'm sorry if this seems as a bit of a noob question. So I downloaded genie scout, uploaded the ratings and everything. My question is, how do you find players for specific tactics, for instance yours @ZaZ
I use this view, you can then see the positional ability and positional ratings. You need positional ability above 13 out of 20 and the best positional rating you can afford.
We had the linked post below from @showmake which linked to a Chinese site where some players had done analysis of positional attributes.
Post by @showmake
Below is the Google translation of the text from the site but I will also include at the top a summary of their data against the ZaZ Blue DM tactic by @ZaZ.
I now intend to produce a Genie Scout rating based on the summary data above. I will share it in the next few days for anyone that want s to test it.
This does appear to loosely align with but build on some of the work already discussed in this forum.
Here is the text from the PlayGM web site.
[Player experience] Share player attribute utility calculated based on deep neural network
This post was last edited by C·Gunners at 2021-12-30 14:36
Basic introduction:
1. We are a basketball manager game development team. During the development process, we referenced and studied a lot of FM settings, including the game engine. I learned a lot of game knowledge from the bursting room. The purpose of this post is to give back to the community.
2. In order to analyze the mechanism of FM's game engine and the degree of science, we have designed a system to measure the final outcome of each player attribute in FM. The degree of influence.
3. fm-arena.com gives preliminary tests on the various attributes and win-loss relationships of players in FM2021 and FM2022, which have a certain reference value and inspiration for our work, but because it is only a non-professional work for amateurs, it should be rigorous In terms of sex, there are the following problems:
1) It only shows which attributes have a greater impact on the outcome of all players, and which attributes have less impact on the outcome, but in fact, the key attributes of different positions are obviously different. Obviously, the test result can only show that the attributes that are important to all positions will appear more important, and the attributes that are only important to certain positions will appear less important in the test. For example, the test results have almost no impact on the outcome of shooting, but explosive power has a great impact on the outcome. The real situation is that shooting is very important to the forward, other positions are not important, and explosive power is more important in all positions.
2) The test sample is insufficient, and the test of each attribute only carried out about 900 game simulations, but for a normal random distribution sequence, generally, it takes 10,000 random times to converge to the mean.
3) The influence of attribute on the outcome is non-linear. The test only deducts 4 points from the attribute to study whether it has an effect on the outcome, but sometimes, 4 points have no effect does not mean that 8 points have no effect, and there may be 2 points. It is close to 4 o'clock.
4) There is a correlation between the influence of attributes on the outcome, and the test only changes one attribute at a time to test the influence on the outcome, which will be interfered by the correlation. For example, a breakthrough can speed up when lying on the ball, relying on physical fitness instead of dribbling, or relying on dribbling. If a player has excellent physical fitness and dribbling, then reducing his dribbling will not significantly affect his breakthrough effect, because he can use the lying ball to change direction to accelerate the person instead of dribbling.
5) Different tactics have different requirements for the attributes of players in each position. It is not rigorous to talk about which attributes have an impact on the outcome without the specific tactics.
4. We use artificial intelligence to study the importance of attributes. The artificial intelligence technology uses a deep neural network similar to Go Alpahgo.
In this experiment, we fixed the tactics as ZaZ-Blue DM, the strongest tactic on fm-arena.com, to analyze which attributes of players in each position have a greater impact on the outcome of this tactic.
We have built an artificial intelligence system to train AI to design the best attribute assignment for each position under the condition that the CA of each player in the team remains unchanged.
The general process is as follows. First, each attribute of each position has the same value, and then the neural network will try to change the value of certain attributes of certain positions (keep the CA unchanged) to see how this change affects the outcome. Positive or negative, in order to continue to iterate on the multi-layer network, knowing that iteratively produces the best 11-person combination that AI thinks.
We trained about 140 machines for 3 weeks, simulated 40 million games, and got convergence results.
5. In order to verify whether the best 11-person attribute assignment given by AI is really the best, we conducted a result verification test to design three groups of teams, one team’s player attributes are evenly distributed, the other group It is the best 11-person attribute allocation scheme given manually by senior players based on their game experience. The last group is given by AI. It was put into a test league for 100,000 matches. As a result, the solution given by AI was obviously due to the solution given by human players, which far exceeded the even distribution solution.
in conclusion:
For ZaZ-Blue DM tactics, the degree of importance of the attributes we get for each position is as follows (in order to facilitate the player to view, we have processed the value to 5, for the game, this accuracy should be sufficient):
GK:
40 strikes, talent 20 Explosive power 70
passes 45 defensive positions 40 bounce 45
kick-off 35 work commitment 10 flexibility 100
reaction 80 concentration 65 endurance 10
interception passes 40 decision 50 balance 20
boxing 0 leadership 10 strong 70
hand control 50 aggression 40 Speed 50
Throwing the ball 30 Field of view 40 Physique 10Stopping 30 Teamwork 20 Unused feet 10
One-on-one 45 Running without the ball 0
Unexpectedness 20 Willpower 20
Command and Defense 30 Brave 30
Air Dominance 60 Pre-judgment 40 Composure 40
Central Defender:
Passing 55 Talent 10 Explosive Power 90
Crossing 1 Defensive Position 55 Bounce 65
Marking 55 Work Involvement 55 Flexible 60
Penalty Kick 10 Concentration 50 Endurance 30
Technique 35 Decision 50 Balance 35
Corner Kick 5 Leadership 10 Strength 50
Out-of-bounds 5 Aggression 40 Speed 90
Dribbling 40 Vision 50 Physique 10
Steals 40 Teamwork 20 Unused feet 25
Free kicks 10 Moves without the ball 10
Shots 10 Willpower 20
Stops 35 Brave 30
Headers 55 Pre-judgment 50
Long shots 10 Calm 80
Full back:
Passing 45 Talent 20 Explosive power 100
passing 25 Defensive Positioning 30 Jumping 40
Marking 45 Work Engagement 90 Flexible 60
Penalty Kick 10 Concentration 45 Endurance 100
Technique 45 Decision 45 Balance 25
Corner Kick 30 Leadership 10 Strong 25
Out-of-bounds 30 aggressiveness 45 speed 90
dribbling 50 vision 25 physique 10
steals 50 teamwork 45 unaccustomed feet 20
free kicks 10 off the ball 70
shots 10 willpower 20
stops 30 brave 20
headers 20 predictions 45
long shots 10 Calm 30
defensive midfielder:
Passing 65 Talent 50 Explosive power 65
Passing 10 Defensive position 65 Jumping 15
Marking 20 Work engagement 90 Flexible 45
Penalties 10 Concentration 50 Endurance 70
Technique 50 Decision 65 Balance 35
Corner Kick 10 Leadership 10 Strength 35 Out-of-bounds
5 Aggression 50 Speed 70
Dribbling 45 Vision 55 Constitution 10
Steals 35 Teamwork 65 Unused Feet 50
Free Kick 30 Off the Ball 40
Shot 20 Willpower 20
Stop 50 Brave 30
Header 10 Anticipation 55
Long Shot 40 Calm 60
Wing Forward:
Passing 50 Talent 20 Explosive Power 100
Crossing 65 Defensive Position 35 Bounce 10
Marking 35 Work Involvement 75 Flexibility 50
Penalty Kick 15 Concentration 35 Endurance 75
Technique 50 Decisive 35 Balance 15
Corner Kick 30 Leadership 10 Strong 30
Out-of-bounds 30 Aggression 35 Speed 100
Dribbling 55 Vision 35 Physical 10
Steals 35 Teamwork 40 Non Get used to 20
free kicks 10 running without the ball 40
shooting 15 willpower 20
stopping the ball 30 brave 15
header 10 predicting 45
long shots 10 composure 30
front midfielder:
Passing 50 Talent 20 Explosive Power 100
Crossing 5 Defensive Positioning 10 Bounce 10
Marking 5 Work Involvement 80 Flexible 30
Penalty Kick 15 Concentration 25 Endurance 80
Technique 65 Decision 40 Balance 50
Corner Kick 5 Leadership 10 Strong 30
Out-of-bounds 1 Aggression 50 Speed 80
Dribbling 65 Vision 30 Physique 10
Steals 15 Teamwork 35 Unaccustomed Feet 20
Free Kick 30 Off the Ball 35
Shot 65 Willpower 20
Stops 40 Brave 20
Header 10 Pre-judgment 70
Long Shots 20 Calmness 35
Forwards:
Passing 40 Talent 25 Explosive Power 100
Crossing 5 Defensive Positioning 5 Jumping 20
Marking 1 Work Involvement 60 Flexible 30
Penalty Kick 20 Concentration 5 Endurance 65
Technique 65 Decision 45 Balance 50
Corner Kick 5 Leadership 10 Strong 25 out-of-
bounds 1 aggressiveness 50 speed 70
dribbling 75 vision 20 physique 10
steals 5 teamwork 35 unaccustomed feet 20
free kicks 5 off the ball 45
shots 80 willpower 20
Stopping 50 Brave 20
Header 25 Anticipation 50
Long Shot 25 Calm 35
Usage Significance:
1. Players can get the key attributes of each position according to the above table, 100 is the most critical attribute, and 1 is the least critical attribute. So as to guide the selection of materials.
2. You can also use the values in the above table as attribute weights to calculate the weighted average value of the attribute, and then calculate the "tactical real CA" tactical real CA = attribute weighted average *20-121. Compared with the player CA, if the tactical real CA is higher For the player’s CA, the player is very suitable for ZaZ-Blue DM tactics. The higher the tactical real CA, the more suitable it is for player selection.
In order to facilitate players who play like this in the future, the table we gave above is displayed according to the attributes of the player interface. Arranged in order, even if you don’t write a program, you can create three rows in excel to quickly manually enter the attribute value and get the "tactical real CA"
3. The guiding significance for training, the attribute is occupied by the CA, in order to cultivate the tactical real CA For the highest player, we can cultivate the most cost-effective attribute. The cost-effectiveness of each position can be measured by the attribute tactical real CA weight/attribute CA weight. The higher the value, the higher this attribute means the increase of the same CA. Will make the player's attribute tactical real CA higher.
Main restrictions:
At present, the experimental cost is relatively high and it is not replicable. The engine version is changed and the tactics are changed. If there is no transfer learning, the AI needs to be retrained. Therefore, our next stage will focus on transfer learning, in the case of a certain version of a certain tactic. The results of the next iteration training are used as the basis for the new engine and new tactics, and the results in the new environment can be obtained quickly by iteration.
wow lots of info, hard to understand everything but in excel sheet but as i understand it shows that for example composure is most important for CD(80) and DM(60) and for rest of positions are not so important?
Wigo said: wow lots of info, hard to understand everything but in excel sheet but as i understand it shows that for example composure is most important for CD(80) and DM(60) and for rest of positions are not so important?
Yes you need to be looking at the higher numbers for each position to get an idea on what to look for in players in those positions. Interesting work I think
very interesting! absolutly different for fm-arena filter test! gonna give it a go! maybe you can a new thread for this? could be a game changer!
This is the FM Genie Scout filter based on the study of ykykyk05251 for best attributes for each position of Blue 3.0 RPM (aka Blue DM). The study used machine learning to get the weights.
Keep in mind that two attributes of the table have ranges from -100 to 0 in FM Genie Scout instead of 0 to 100: eccentricity and punching tendency. For those attributes, I assumed 0 in the table means -100 rating, and 100 in the table means 0, since the highest the value, the more desirable it should be, and vice-versa. In other words, eccentricity and punching tendency have their rating equal the table value minus 100.
ZaZ said: This is the FM Genie Scout filter based on the study of ykykyk05251 for best attributes for each position of Blue 3.0 RPM (aka Blue DM). The study used machine learning to get the weights.
Keep in mind that two attributes of the table have ranges from -100 to 0 in FM Genie Scout instead of 0 to 100: eccentricity and punching tendency. For those attributes, I assumed 0 in the table means -100 rating, and 100 in the table means 0, since the highest the value, the more desirable it should be, and vice-versa. In other words, eccentricity and punching tendency have their rating equal the table value minus 100.
I think this needs a little work to balance out the ratings. As it sits it favours Strikers and AMs too much and there are no GKs anywhere to be seen. The weightings have only been developed by position and not across all positions so need adjusting in terms of weight for each position. I would also like to derive the positions not covered.
I will have a play and update it next week. I have a very busy weekend unfortunately.
Mark said: I think this needs a little work to balance out the ratings. As it sits it favours Strikers and AMs too much and there are no GKs anywhere to be seen. The weightings have only been developed by position and not across all positions so need adjusting in terms of weight for each position. I would also like to derive the positions not covered.
I will have a play and update it next week. I have a very busy weekend unfortunately.
I'll do my poor man's version too, with only the most important attributes, so it puts more importance on those attributes and gives good cheaper players.
@Mark, I made three poor man's version and I would like to know your opinion in which one gives the best results. Can you test the three and tell me which one you like the most? Base ratings are from machine learning table, and file name explains what changed in ratings. All of them try to focus on most important attributes in some way.
Version 1: all ratings minus 25 (removes lower ratings and increases % gap of remaining)
Version 2: all ratings under 50 minuts 25 (removes lower ratings and keeps gap of top)
Version 3: all ratings minus difference to 100 (removes lower ratings and increases absolute gap of remaining)
* Gap means the difference in rating between two attributes. For example, 25 to 40 have a gap of 15.
Anyone else's opinion would also be appreciated!
ZaZ said: @Mark, I made three poor man's version and I would like to know your opinion in which one gives the best results. Can you test the three and tell me which one you like the most? Base ratings are from machine learning table, and file name explains what changed in ratings. All of them try to focus on most important attributes in some way.
Version 1: all ratings minus 25 (removes lower ratings and increases % gap of remaining)
Version 2: all ratings under 50 minuts 25 (removes lower ratings and keeps gap of top)
Version 3: all ratings minus difference to 100 (removes lower ratings and increases absolute gap of remaining)
* Gap means the difference in rating between two attributes. For example, 25 to 40 have a gap of 15.
Anyone else's opinion would also be appreciated!
OK so I had a quick play with the ratings files. The first observation is that you need to balance the ratings files. I learned this from last years attempt. What do I mean by this, well have a look at the screenshot below. There should not be major discrepancies between positions ie you should be getting most positions covered in the top 50 or so top players.
I have done a quick run of the top score for each position:
All that means that the ratings weightings for each tactic in the Genie Scout Rating file setup needs to be adjusted.
Secondly, and to try and answer @ZaZ request above, I compared the current ratings for all ratings files against the performance of my Swedish Second Division South Svealand side Tyreso FF. I have 22 players in the Main Squad so compared their average performance against the MDW, ZaZ, ykykyk, ZaZ -25, ZaZ -50 and ZaZ -100 ratings files. I have only used ratings for the position selected.
Here are the results:
Disclaimer: I have used players out of the positions identified in the results, but in the main they have played the positions indicated. In addition I have used 2 main tactics, Dragons Breath that uses an MC instead of AMC, and ZaZ Blue DM, but given the ratings files don't all have MC ratings I just used AMC.
To answer @ZaZ question the ZaZ-100 ended up equal with MDW in picking 8 of the best 11 positional players. It does need balancing to make it more usable though. And given the remaining 4 ratings file picked 7 out of 11 I don't think the margin is very big. All are definitely usable.
Now if we look at average % difference from the best rated team the order is MDW, ykykyk, ZaZ -50, ZaZ -25, ZaZ, ZaZ -100.
Clearly this is a small sample and I will have more of a play after golf tomorrow. But I think if you wanted to use one of the ratings files above, I would go with the ZaZ -50.
Mark said: OK so I had a quick play with the ratings files. The first observation is that you need to balance the ratings files. I learned this from last years attempt. What do I mean by this, well have a look at the screenshot below. There should not be major discrepancies between positions ie you should be getting most positions covered in the top 50 or so top players.
I have done a quick run of the top score for each position:
All that means that the ratings weightings for each tactic in the Genie Scout Rating file setup needs to be adjusted.
Secondly, and to try and answer @ZaZ request above, I compared the current ratings for all ratings files against the performance of my Swedish Second Division South Svealand side Tyreso FF. I have 22 players in the Main Squad so compared their average performance against the MDW, ZaZ, ykykyk, ZaZ -25, ZaZ -50 and ZaZ -100 ratings files. I have only used ratings for the position selected.
Here are the results:
Disclaimer: I have used players out of the positions identified in the results, but in the main they have played the positions indicated. In addition I have used 2 main tactics, Dragons Breath that uses an MC instead of AMC, and ZaZ Blue DM, but given the ratings files don't all have MC ratings I just used AMC.
To answer @ZaZ question the ZaZ-100 ended up equal with MDW in picking 8 of the best 11 positional players. It does need balancing to make it more usable though. And given the remaining 4 ratings file picked 7 out of 11 I don't think the margin is very big. All are definitely usable.
Now if we look at average % difference from the best rated team the order is MDW, ykykyk, ZaZ -50, ZaZ -25, ZaZ, ZaZ -100.
Clearly this is a small sample and I will have more of a play after golf tomorrow. But I think if you wanted to use one of the ratings files above, I would go with the ZaZ -50.
Thanks for the thorough analysis, Mark. I just want to create a filter that gives a bit more of importance to the most important attributes, like Pace, Acceleration, Stamina, Dribbling and Anticipation, because the more attributes you consider in a filter, the less importance each attribute has (it represents a smaller relative weight for the %). Doing that, it allows you to find good players for cheaper prince, since cost of players is directly linked to CA and reputation.
guys what you think of that way? i made filters for every position... first i add all skill attributes of every position what is required... for example for advanced forvard i add all skills what is marked in game as important atributes... then i look at "test excel sheet" and then adjust those skills how it shows like finishing 8, dribbling 8,acceleration 10 and so on... you can add + or - in filter search depending on your strong of team... i am testing now and it works incredibble! i mean i am finding cheap players who are performing amazing! for example check shane lavery in blackpool! he can be a beast according to those attributes for very cheap... i can upload my filters for you to check as a fm files...
hope you get it... feel still kind off dizzy after new years celebration 
Wigo said: guys what you think of that way? i made filters for every position... first i add all skill attributes of every position what is required... for example for advanced forvard i add all skills what is marked in game as important atributes... then i look at "test excel sheet" and then adjust those skills how it shows like finishing 8, dribbling 8,acceleration 10 and so on... you can add + or - in filter search depending on your strong of team... i am testing now and it works incredibble! i mean i am finding cheap players who are performing amazing! for example check shane lavery in blackpool! he can be a beast according to those attributes for very cheap... i can upload my filters for you to check as a fm files...
hope you get it... feel still kind off dizzy after new years celebration 
Upload it please
here you go mate... https://www.mediafire.com/file/tdd07q0z4zh29me/Power+attributes+2.rar/file