TotalFootballFan9
Member
- Joined
- Dec 10, 2013
- Messages
- 4,113
- Reaction score
- 201
- Points
- 63
Heh, I see that tactics testing is booming here again 
so I want to help people to make their testing systems more productive because there are many things that people don't pay attention to but these things ruin testing systems making their result totally useless for the real game.
First of all, I want that people understand that no matter how good your testing system is and no matter do you use FMRTE to keep the morale and fitness of your players locked or not during your testing, anyway, if you test only for 60 matches or less then the only what you can get is a sneak-peak about tactics' performance and if you want to get a really decent knowledge about tactics' performance then you should test no less than 200-300 matches.
The second, if you are going to catch any small tweaks like changing "D-line" by one notch or adding "Dribble More/Shot More Often" PIs by such testing then I strongly recommend you that you leave this idea because this game is too random that you can catch such small tweaks by this kind of testing but it can tell you do you move in the right direction or not.
The third, any testing with the real players in the testing team is just waste of your time because there are no real players that equally good and it inevitably that some tactics will get "better/more suited" players and some tactics will get "worse/less suited" players and that will have a huge impact on the testing result, recently I made a thread(you can search for it) where I showed that if you change only 2 players in the starting eleven for "more suited/better" players then it might improve the result up to 40%. You might ask that based on what I said above you can't test tactic during normal gameplay? How come? Of course, you also can test tactics during normal gameplay and the best way to do that is pick the weakest in the most competitive league such English Premier League and try to win the league with that team 2 TIMES IN A ROW and when you achieve that you'll get full knowledge about the performance of the tactic that you used.
And the last one, usually we have only about 60 matches to get a sneak-peak about tactics' performance to decided if it's decent enough so we test it for more matches and in total we have about 200-300 matches to test so we need to remove any randomness as much as possible to get the most accurate knowledge about tactics’ performance as possible and below in thread I'll tell what needs to be done to achieve that.
It happens every time and it might happen at the begging/middle/end of tests that some of AI teams starts underachieve and its manager gets fired and it brings great randomness to the testing because this team hires a new manger and starts to play with a different formation and approach and it might with many team in the testing league.
So it has crucial importance to keep AI managers in their teams until the end of tests and also it's very important that AI teams have something to fight for until the end of test.
Here's how to do that, you need to create 2 divisions: top division and division below it.
You can add any teams to the Top Division because we don't need it for testing and it might have "None" Detail Level during tests.
We'll be testing in the lower division because it can be configured such way that AI teams have something to fight until the end of the testing and AI managers don't get sacked until the end of tests.
For example, here's how you can configure a testing league that consist of 8 teams : the first 2 place get a straight promotion then 4 places can get promotion though play-offs
In such testing league no one gets fired until the end of the season because as you can see at the screenshot Juventus and Bayern had chances to get a promotion through the play-offs until at the end of the test and that prevented their manager from sacking. Also, the team from the middle had chances to fight for the straight promotion until the end of the season so no one stopped playing until the end of the season because nothing was known until the very last match of the season.
"Dirtiness" is extremely important attribute that has a huge impact on the result of tests so you shouldn't leave to random.
For example, some positions in the testing team have very high "Dirtiness" and some position very low "Dirtiness" and when you test tactics that have more positions that use the players with very high "Dirtiness" then these tactics get a huge disadvantage because they'll be getting sent off more often.
So in order to eliminate this randomness I recommend setting "Dirtiness" to "11" for all defenders, defensive midfielders and central midfielders in your testing team and setting "Dirtiness" to "8" for all attacking midfielders and strikers in your testing team, it puts all tactics in equal conditions and simulates the real game environment very well.
Also, if you don’t use FMRTE and you use Quick Pick to pick the team then you don't rotate your team because Quick Pick doesn't rotate the team it just picks the fit players for the starting eleven so all other players expect players from the starting won't have Match Sharpness and if something happens to the players from starting eleven then you'll be using players with very low Match Sharpness and that strongly affects their performance and adds randomness to the testing.
So in order to avoid the issue when you have to use players with low Match Sharpness due to suspension of the players from the starting eleven you should remove any suspension for the next match/matches in the league because of high amount of yellow cards or red card.
It can be done by switching to the advanced rules in the editor and changing the disciplinary rules of the league
Please notice that players still will be getting sent off in matches due to red card or second yellow card but they won't miss the next matches because of it and for those people who don't use FMRTE it's a great opportunity to reduce randomness because they don't have to use players with very low Match Sharpness for the next matches.
MATCH PREPARATION
NEVER LEAVE IT TO YOUR ASSISTANT MANAGER because the match preparation is very important, it gives a buff to your players for the next match but the assistant manger very often makes terrible choices the same as he makes terrible choices when you use Quick Pick to pick the team.
I recommend setting it to "Match Tactic" because it doesn't give any buff at all, it just speed up the rate at which tactics get "Familiarity" so all tactics will be in equally conditions all the time during testing, you won't have any buff at all but anyway it's much better than if some tactic get more useful buff and other tactic get less useful buff, it'll greatly reduce randomness in tests.
This is one of the most important things because how valuable your test results will be for the real game depends on how you configured the testing league.
First of all, It's important to understand that you don’t have all time in the world and you have only 40-60 matches to test in order to get a sneak-peek about tactics' performance in order to decide if it's enough decent and does it deserve that you test it for more matches to get better understand about its performance.
So you have only 40-60 matches to test in order to get a sneak-peek about tactics' performance and it means that you want to give a real "Stress Test" during that distance in order to unveil any weaknesses in the tactic as soon as possible and it means that you can't allow situations when the testing team wins matches because it's is stronger than its opponents so the testing team should be weaker or equal strong with the teams in the testing league.
Someone might ask if the testing team is the weakest team in the testing league then you only simulate situation when you manage "underdog" club and don't simulate situation when you manage "favorite" club but tactics might perform different when you use it with "underdog" club and when you use it with "favorite" club because opponents play different against you in both cases.
That would be a good question but let's see what different when you play with "underdog" club and when you play with "favorite" club:
1) When you play with "underdog" club then AI teams use against you attacking approaches.
2) When you play with "favorite" club then AI teams use against you defensive approaches.
In order to test how tactics performance when AI teams use a defensive approach against them you can add to the testing league teams that use 2 most popular defensive formations that AI uses when it want to use a defensive approach :
And you can tweak the AI managers that will be in charge of these teams to always use a defensive approach no matter what and in order to do that you should tweak their "Tactical Attributes" such as "Attacking", "Sitting Back", "Depth" and some other :
Also, it's possible to add some "Tendencies" such as "Will Use Counterattacking" and some other:
In order to test how tactics performance when AI teams use an attacking approach against them you can add to the testing league teams that use 2 most popular attacking formations that AI uses when it want to use a attacking approach :
Of course, you also should tweak the AI managers that will be in charge of these teams to always use an attacking approach no matter what and it can be done the same as I described above by tweaked their "Tactical Attributes" and "Tendencies".
Please notice, if you have no idea how to tweak "Tactical Attributes" and "Tendencies" to make AI manager to use a defensive or an attacking approach then you can clone Jose Mourinho few times, tweak his "Preferred Formation" to the defensive formations that I showed above and put him in charge of the teams that are supposed to use a defensive approach in your testing league because with the default settings Jose Mourinho always plays defensively no matter what(do they play vs underdog or favorite).
Also, you can clone Jurgen Klopp or Arsen Wenger few times, tweak thier "Preferred Formations" to the attacking formations that I showed above and put them in charge of the teams that are supposed to use an attacking approach in your testing league because with the default settings Jurgen Klopp and Arsen Wenger always play attackingly no matter what(do they play vs underdog or favorite).
As you can see it isn't necessary that the testing team should be the strongest or the weakest team in the testing league if you want to test how tactics perform if they used with "underdog" or "favorite" club.
It's very important to have equal number of teams that use defensive approaches and teams that use attacking approaches in your testing league because if there’s imbalance in that area then you don't get a clear picture how tactics perform against attacking approaches or defensive approaches and that greatly reduces the efficiency of your testing result for the real game.
The composition of the testing is the crucial factor that determines everything and what results you'll be getting during your tests totally depends on it.
You will be testing many different tactics/formations/approaches and if the players in the testing team even slightly favor to any specific tactics/formations/approaches then the results of such testing can be thrown into a garbage can because there's no any use of it.
You might ask how much impact it would have on the test result if some tactic gets just few better and more suited players in the starting eleven than other tactic? I'd say in this case the result might be better for this tactic up to 40% on 60 matches distance than for other tactic.
As you can see only just few better and more suite players in the starting eleven have a huge impact on the result of testing so it excludes any possibility of using the real players because there simply no real players that are equally good and all real players have different quality, it mean that you have to fill the testing team with premade players that have equal quality.
I want to tell you at once that if you are going to balance your testing team just by CA (Current Ability) then you better leave this idea immediately because CA is only amount of points that players have to spend on their attributes but what’s more important is that how these points were spent and the way how the points were spent determines the player’s efficiency for the roles.
For example, here are 2 players that have identical CA = 159 and "Natural" rating for AML position but how do you think who of them will perform better as Winger?
I'm sure that it’s obvious for everyone that Player A would perform much better as Winger than Player B even that they both have identical CA – 159 and it's because CA of Player A was spent much more efficient for Winger than CA of Players B.
As you see that you can say that 2 players are equally good for some position only when they have identical attributes and not when they have identical CA because players with identical CA might have a huge difference in their efficiency at some position.
Let's say you have Barcelona, Real Madrid and PSG in your testing league.
Here are their best starting eleven (or almost the best, there might be some optimization) :
It would greatly reduce randomness in your testing if these teams always use the same starting eleven for all their matches in the testing league but unfortunately these factors won't allow that happen :
1) Injuries ; even if you set Injury Proneness of all players in the testing league to "1" then injuries still occur and believe me they happen with Injury Proneness quite often.
2) Suspension for some matches reaching some amount of Yellow Cards or after getting a Red Card. (Btw, that's why I suggested removing any suspension after players reach some amount or Yellow Cards or Red Cards in your testing league).
3) Regular team rotation that all AI managers do.
There some way around how to make AI use the same starting eleven as often as possible. You can clone the players from the starting eleven and remove all other players that you don't want to be used but please notice if you clone Messi, Suarez and Neymar 4 times then AI manager will be choosing Messi for all positions where Messi can play so you will be seeing Barca's starting eleven where Messi and his clones play at all attacking position and AI manager don't pick Suarez and Neymar so to avoid such situation from happening you need limit the position where Messi can play to AMR position and in this case AI manager won't pick him for STC and AML position.
For example, here's how composition of Barcelona team might look in your testing league:
Yes, pinpoint tweaking of AI teams in your testing league might take some quite time but you'll be rewarded with greatly reduced randomness and with much more accurate testing results.
so I want to help people to make their testing systems more productive because there are many things that people don't pay attention to but these things ruin testing systems making their result totally useless for the real game. INTRODUCTION
First of all, I want that people understand that no matter how good your testing system is and no matter do you use FMRTE to keep the morale and fitness of your players locked or not during your testing, anyway, if you test only for 60 matches or less then the only what you can get is a sneak-peak about tactics' performance and if you want to get a really decent knowledge about tactics' performance then you should test no less than 200-300 matches.
The second, if you are going to catch any small tweaks like changing "D-line" by one notch or adding "Dribble More/Shot More Often" PIs by such testing then I strongly recommend you that you leave this idea because this game is too random that you can catch such small tweaks by this kind of testing but it can tell you do you move in the right direction or not.
The third, any testing with the real players in the testing team is just waste of your time because there are no real players that equally good and it inevitably that some tactics will get "better/more suited" players and some tactics will get "worse/less suited" players and that will have a huge impact on the testing result, recently I made a thread(you can search for it) where I showed that if you change only 2 players in the starting eleven for "more suited/better" players then it might improve the result up to 40%. You might ask that based on what I said above you can't test tactic during normal gameplay? How come? Of course, you also can test tactics during normal gameplay and the best way to do that is pick the weakest in the most competitive league such English Premier League and try to win the league with that team 2 TIMES IN A ROW and when you achieve that you'll get full knowledge about the performance of the tactic that you used.
And the last one, usually we have only about 60 matches to get a sneak-peak about tactics' performance to decided if it's decent enough so we test it for more matches and in total we have about 200-300 matches to test so we need to remove any randomness as much as possible to get the most accurate knowledge about tactics’ performance as possible and below in thread I'll tell what needs to be done to achieve that.
HOW TO KEEP AI MANAGERS IN THEIR TEAMS UNTIL THE END OF THE TESTING PREVENTING THEM FROM GETTING SACKED
It happens every time and it might happen at the begging/middle/end of tests that some of AI teams starts underachieve and its manager gets fired and it brings great randomness to the testing because this team hires a new manger and starts to play with a different formation and approach and it might with many team in the testing league.
So it has crucial importance to keep AI managers in their teams until the end of tests and also it's very important that AI teams have something to fight for until the end of test.
Here's how to do that, you need to create 2 divisions: top division and division below it.
You can add any teams to the Top Division because we don't need it for testing and it might have "None" Detail Level during tests.
We'll be testing in the lower division because it can be configured such way that AI teams have something to fight until the end of the testing and AI managers don't get sacked until the end of tests.
For example, here's how you can configure a testing league that consist of 8 teams : the first 2 place get a straight promotion then 4 places can get promotion though play-offs
In such testing league no one gets fired until the end of the season because as you can see at the screenshot Juventus and Bayern had chances to get a promotion through the play-offs until at the end of the test and that prevented their manager from sacking. Also, the team from the middle had chances to fight for the straight promotion until the end of the season so no one stopped playing until the end of the season because nothing was known until the very last match of the season.
RED CARDS, YELLOW CARDS, SUSPENSION AND DIRTINESS
"Dirtiness" is extremely important attribute that has a huge impact on the result of tests so you shouldn't leave to random.
For example, some positions in the testing team have very high "Dirtiness" and some position very low "Dirtiness" and when you test tactics that have more positions that use the players with very high "Dirtiness" then these tactics get a huge disadvantage because they'll be getting sent off more often.
So in order to eliminate this randomness I recommend setting "Dirtiness" to "11" for all defenders, defensive midfielders and central midfielders in your testing team and setting "Dirtiness" to "8" for all attacking midfielders and strikers in your testing team, it puts all tactics in equal conditions and simulates the real game environment very well.
Also, if you don’t use FMRTE and you use Quick Pick to pick the team then you don't rotate your team because Quick Pick doesn't rotate the team it just picks the fit players for the starting eleven so all other players expect players from the starting won't have Match Sharpness and if something happens to the players from starting eleven then you'll be using players with very low Match Sharpness and that strongly affects their performance and adds randomness to the testing.
So in order to avoid the issue when you have to use players with low Match Sharpness due to suspension of the players from the starting eleven you should remove any suspension for the next match/matches in the league because of high amount of yellow cards or red card.
It can be done by switching to the advanced rules in the editor and changing the disciplinary rules of the league
Please notice that players still will be getting sent off in matches due to red card or second yellow card but they won't miss the next matches because of it and for those people who don't use FMRTE it's a great opportunity to reduce randomness because they don't have to use players with very low Match Sharpness for the next matches.
MATCH PREPARATION
NEVER LEAVE IT TO YOUR ASSISTANT MANAGER because the match preparation is very important, it gives a buff to your players for the next match but the assistant manger very often makes terrible choices the same as he makes terrible choices when you use Quick Pick to pick the team.
I recommend setting it to "Match Tactic" because it doesn't give any buff at all, it just speed up the rate at which tactics get "Familiarity" so all tactics will be in equally conditions all the time during testing, you won't have any buff at all but anyway it's much better than if some tactic get more useful buff and other tactic get less useful buff, it'll greatly reduce randomness in tests.
CONFIGURATION OF THE TESTING LEAGUE
This is one of the most important things because how valuable your test results will be for the real game depends on how you configured the testing league.
First of all, It's important to understand that you don’t have all time in the world and you have only 40-60 matches to test in order to get a sneak-peek about tactics' performance in order to decide if it's enough decent and does it deserve that you test it for more matches to get better understand about its performance.
So you have only 40-60 matches to test in order to get a sneak-peek about tactics' performance and it means that you want to give a real "Stress Test" during that distance in order to unveil any weaknesses in the tactic as soon as possible and it means that you can't allow situations when the testing team wins matches because it's is stronger than its opponents so the testing team should be weaker or equal strong with the teams in the testing league.
Someone might ask if the testing team is the weakest team in the testing league then you only simulate situation when you manage "underdog" club and don't simulate situation when you manage "favorite" club but tactics might perform different when you use it with "underdog" club and when you use it with "favorite" club because opponents play different against you in both cases.
That would be a good question but let's see what different when you play with "underdog" club and when you play with "favorite" club:
1) When you play with "underdog" club then AI teams use against you attacking approaches.
2) When you play with "favorite" club then AI teams use against you defensive approaches.
In order to test how tactics performance when AI teams use a defensive approach against them you can add to the testing league teams that use 2 most popular defensive formations that AI uses when it want to use a defensive approach :
And you can tweak the AI managers that will be in charge of these teams to always use a defensive approach no matter what and in order to do that you should tweak their "Tactical Attributes" such as "Attacking", "Sitting Back", "Depth" and some other :
Also, it's possible to add some "Tendencies" such as "Will Use Counterattacking" and some other:
In order to test how tactics performance when AI teams use an attacking approach against them you can add to the testing league teams that use 2 most popular attacking formations that AI uses when it want to use a attacking approach :
Of course, you also should tweak the AI managers that will be in charge of these teams to always use an attacking approach no matter what and it can be done the same as I described above by tweaked their "Tactical Attributes" and "Tendencies".
Please notice, if you have no idea how to tweak "Tactical Attributes" and "Tendencies" to make AI manager to use a defensive or an attacking approach then you can clone Jose Mourinho few times, tweak his "Preferred Formation" to the defensive formations that I showed above and put him in charge of the teams that are supposed to use a defensive approach in your testing league because with the default settings Jose Mourinho always plays defensively no matter what(do they play vs underdog or favorite).
Also, you can clone Jurgen Klopp or Arsen Wenger few times, tweak thier "Preferred Formations" to the attacking formations that I showed above and put them in charge of the teams that are supposed to use an attacking approach in your testing league because with the default settings Jurgen Klopp and Arsen Wenger always play attackingly no matter what(do they play vs underdog or favorite).
As you can see it isn't necessary that the testing team should be the strongest or the weakest team in the testing league if you want to test how tactics perform if they used with "underdog" or "favorite" club.
It's very important to have equal number of teams that use defensive approaches and teams that use attacking approaches in your testing league because if there’s imbalance in that area then you don't get a clear picture how tactics perform against attacking approaches or defensive approaches and that greatly reduces the efficiency of your testing result for the real game.
COMPOSITION OF THE TESTING TEAM // CA (CURRENT ABILITY)
The composition of the testing is the crucial factor that determines everything and what results you'll be getting during your tests totally depends on it.
You will be testing many different tactics/formations/approaches and if the players in the testing team even slightly favor to any specific tactics/formations/approaches then the results of such testing can be thrown into a garbage can because there's no any use of it.
You might ask how much impact it would have on the test result if some tactic gets just few better and more suited players in the starting eleven than other tactic? I'd say in this case the result might be better for this tactic up to 40% on 60 matches distance than for other tactic.
As you can see only just few better and more suite players in the starting eleven have a huge impact on the result of testing so it excludes any possibility of using the real players because there simply no real players that are equally good and all real players have different quality, it mean that you have to fill the testing team with premade players that have equal quality.
I want to tell you at once that if you are going to balance your testing team just by CA (Current Ability) then you better leave this idea immediately because CA is only amount of points that players have to spend on their attributes but what’s more important is that how these points were spent and the way how the points were spent determines the player’s efficiency for the roles.
For example, here are 2 players that have identical CA = 159 and "Natural" rating for AML position but how do you think who of them will perform better as Winger?
PLAYER A
PLAYER B
PLAYER B
I'm sure that it’s obvious for everyone that Player A would perform much better as Winger than Player B even that they both have identical CA – 159 and it's because CA of Player A was spent much more efficient for Winger than CA of Players B.
As you see that you can say that 2 players are equally good for some position only when they have identical attributes and not when they have identical CA because players with identical CA might have a huge difference in their efficiency at some position.
COMPOSITION OF AI TEAMS IN TESTING LEAGUE
Let's say you have Barcelona, Real Madrid and PSG in your testing league.
Here are their best starting eleven (or almost the best, there might be some optimization) :
It would greatly reduce randomness in your testing if these teams always use the same starting eleven for all their matches in the testing league but unfortunately these factors won't allow that happen :
1) Injuries ; even if you set Injury Proneness of all players in the testing league to "1" then injuries still occur and believe me they happen with Injury Proneness quite often.
2) Suspension for some matches reaching some amount of Yellow Cards or after getting a Red Card. (Btw, that's why I suggested removing any suspension after players reach some amount or Yellow Cards or Red Cards in your testing league).
3) Regular team rotation that all AI managers do.
There some way around how to make AI use the same starting eleven as often as possible. You can clone the players from the starting eleven and remove all other players that you don't want to be used but please notice if you clone Messi, Suarez and Neymar 4 times then AI manager will be choosing Messi for all positions where Messi can play so you will be seeing Barca's starting eleven where Messi and his clones play at all attacking position and AI manager don't pick Suarez and Neymar so to avoid such situation from happening you need limit the position where Messi can play to AMR position and in this case AI manager won't pick him for STC and AML position.
For example, here's how composition of Barcelona team might look in your testing league:
3 x Suarez (STC)
3 x Messi (AMR/MR)
3 x Neymar (AML/ML)
4 x Iniesta (CM)
3 x Busquets (DM)
3 x Alba (DL)
4 x Pique (CD)
3 x Dani Alves (DR)
3 x Bravo (GK)
3 x Messi (AMR/MR)
3 x Neymar (AML/ML)
4 x Iniesta (CM)
3 x Busquets (DM)
3 x Alba (DL)
4 x Pique (CD)
3 x Dani Alves (DR)
3 x Bravo (GK)
Yes, pinpoint tweaking of AI teams in your testing league might take some quite time but you'll be rewarded with greatly reduced randomness and with much more accurate testing results.
TO BE CONTINUED...
I'LL ADD MORE TIPS LATER
I'LL ADD MORE TIPS LATER
Last edited:
