Joined
Dec 10, 2013
Messages
4,113
Reaction score
201
Points
63

I strongly recommend that before you start any testing you take a look at this thread; it should give you some hint what to expect from such testing and how much effort you should put into it to get results that would worth anything ---> http://www.fm-base.co.uk/forum/share-download-fm-16-tactics/357725-how-random-game-can.html

Also, you can take a look at this thread; I explained there some key principles that are used in this testing system ---> http://www.fm-base.co.uk/forum/shar...ate-best-tactics-testing-system-possible.html






INTRODUCTION​


First of all, guys if you think that you may not know anything about tactics and how ME works and just test any tactic for 40-60 matches with "Instant Result" spending 25-35 minutes on it and get answer to question what works "the best" in this game then you are greatly mistaken because if such were possible then we already had "the best" tactic only after few weeks since the game release.

Even if you have such "perfect" testing system like this and even if you use FMRTE to freeze the morale and the conditions of all players in the testing league during testing then the minimal amount of matches you need to test is about 200(4-5 seasons) and if you don't use FMRTE to freeze the morale the condition than I'd say you need to test for about 500 matches(about 10 seasons).

Also, if you are going to catch some small changes in your tactic like adding "Shot Less" PI or "Shot More" PI to your striker by doing such testing then I suggest you leave this hopeless endeavor at once because it would be just waste of your time and such testing is supposed to give you general understanding about your tactics performance and confirm that you are on the right path and all "micro" tweaks that you make in your tactic should be based on what you see when you watch matches in full.

Please notice that many of my tactics that I abandoned might show better performance than Ultimate V3 if you test it only for 1-2 seasons but if you test for 5-6 seasons then Ultimate V3 take the lead.

Also, it's important to understand that all these "testing leagues" isn't a necessary thing to play FM or to creating a good tactic and you can test any tactic by playing the real game, just pick the weakest team in the strongest leagues in the game such English Premier League or LIGA BBVA and try to win the league with this team 2 times in row in order to exclude any luck, of course, you can make transfers and when you achieve that you'll have a full knowledge about the tactic’s performance that you used then try to do the same with other tactic and compare your experience.







Here's a list of things that have strong impact on testing results and you should be aware of them:


Match Preparation​


Match Preparation gives a strong temporary buff to testing team for the next match and sufficiently buff overall performance during testing so if you tested "Tactic A" when you have it and forgot to set it when you test other "Tactic B" then in this case you can't compare results between these tactic because "Tactic A" had strong advantage due to Match Preparation. Also, it's very important if you give or not one day rest after match so pay close attention to the Match Preparation.

By default Match Preparation is set to this:​

MP1.png


I recommend setting it to this just 2 day before the 1st match in the season:​

MP2.png





Quick Pick or Manual Pick of the starting eleven​


If you don't use FMRTE to "freeze" the morale and the conditions of the players in the testing team then you need to decide how you are going test and your decision would greatly impact testing results. There are 2 ways to test:

1) Using "Quick Pick" to pick the starting eleven before every match

2) Manually pick the stating eleven before every match and manually rotate the team during testing

In general you are choosing here between "speed" and "quality" and If you choose to manually pick the starting and manually rotate the team then your result would be much better and you can't compare them with any testing that were made with "Quick Pick" because "Quick Pick" has these issues:

A) "Quick Pick" doesn't rotate the team at all it just pick fit players so soon or later all players(except those who were used in the starting eleven) will lose "Match Sharpness" and when something happens with the players from the starting eleven then you have to use players without "Match Sharpness" and such players perform very poorly

B) When you use "Quick Pick" the assistant manager poorly picks substitutions and he might not pick substitutions for some positions at all even that he can pick up to 12 substitutions in the testing league

C) "Quick Pick" sometimes might pick injured players that have less that 75% condition even there are the same quality healthy players available.

As I said if choose to manually pick the starting eleven and rotate the team then your testing results would be much better with any tactic because you'll be keeping the "Match Sharpness" of all players in the testing team almost at 100% and you can use "fresh" players with 100% conditions for all matches which won't happen if you used "Quick Pick".

I'm not insisting on any approach to test because it's up to personal preferences "Speed" vs "Quality", I just want to give understanding that you can't compare testing result that were made in different ways.





SOME RECOMMENDATIONS HOW TO TEST:​


The testing team consists of 57 players:

3 x GK
6 x Central Defenders (3 left footed and 3 right footed)
6 x Full Back (3 left footed and 3 right footed)
6 x Inverted Full Backs (3 left footed and 3 right footed)
6 x Central Midfielders (3 left footed and 3 right footed)
6 x Inside Forwards (3 for the left flank and 3 for the right flank)
6 x Wingers (3 for the left flank and 3 for the right flank)
6 x Attacking Midfielders (3 left footed and 3 right footed)
6 x Fast Strikers (3 left footed and 3 right footed)
6 x Strong Strikers (3 left footed and 3 right footed)


Let's say we need to test Ultimate V3 so for it we don't need these roles in the testing team Wingers, Inverted Full Backs, Fast Strikers and some excessive amount of Attacking Midfielders and Strong Strikers so we put them all into U17 team in order that assistant manger doesn't pick them if you use "Quick Pick" or if you don't use "Quick Pick" then you might leave them if you want.

Just hold "Shift" and click to pick a group of players because sending all these players one by one to U17 team would be boring as **** ;)

u17.png
 
Last edited:
inb4 your own tactics come out on top again
 
Is this supposed to happen?


If you are talking about the starting eleven of Juventus where Marchisio is used at MCL/MC/MCR positions and Morata is used at STCR/STCL positions than yes, it's supposed to happen because it how AI script works when it pick the team.

For example, let's take Barcelona , I tried cloning Iniesta 3 times and Rakitic 3 times in hope that AI manager always would pick one clone of Iniesta and one clone on Rakitic for the starting eleven but that didn't happened and AI manager always picked Iniesta for both MCL/MCR positions over Rakitic, or other example let's take Juventus, I tired cloning Morata 3 times and Dybala 3 times in hope that AI manager always would pick one clone of Morata and one clone of Dybala for the starting eleven but it didn't happen and AI manager always picked Morata for both STCL/STCR positions over Dybala.

For AI manager Iniesta is always better than Rakitic for any midfielder role and Morata is always better than Dybala for any striker role due the way how AI was coded so for some positions where I can't control AI choice by editing players' position rating I cloned only the same players because anyway AI would pick only this player over any other.

It means that Iniesta and his clones are always used at MCL/MCR positions in the starting eleven of Barcelona and Pique and his clones are always used at DCL/DCR positions in the starting eleven of Barcelona, of course, I made some clones to be left footed and some clone to be right footed.

For example, it's easy to prevent AI picking Messi over Suarez and Neymar at all attacking positions by just edition his playing position to only AMR position but there are some position where it isn't possible to do that like MCL/MCR positions or STCL/STCR or DCL/DCR positions so if opposition formation has this position then you'll see the same player and his clones at these positions, of course, some clone will be left footed and some clones will be right footed.

But it's the only way to archive one crucial thing that 100 times reduces randomness in testing and to make AI teams to use the same starting eleven for all matches during testing.
 
Last edited:
Just finish one round of testing. So far so good just a couple annoying problems like injuries still happens more frequent than I expect and quite a few players ask to renew contract during the season.
 
So far so good just a couple annoying problems like injuries still happens more frequent than I expect

"Injury Proneness" was set to the minimal possible level "1" for all players in the testing league and as you can see even despite that injuries happen quite often and that's why it's important to have backup of identical players for every position in the testing team and in all AI teams too because it greatly reduces randomness in testing and allows you to test for less matches to get clear picture about tactics' performance.

quite a few players ask to renew contract during the season.

They don't ask for "renewal" of their contracts... they ask for "improvement" of their contracts and it randomly happen when some player performed quite well in the last match and got very high rating and it can't be avoided...
 
"Injury Proneness" was set to the minimal possible level "1" for all players in the testing league and as you can see even despite that injuries happen quite often and that's why it's important to have backup of identical players for every position in the testing team and in all AI teams too because it greatly reduces randomness in testing and allows you to test for less matches to get clear picture about tactics' performance.



They don't ask for "renewal" of their contracts... they ask for "improvement" of their contracts and it randomly happen when some player performed quite well in the last match and got very high rating and it can't be avoided...
did you set the chances of any injuries occuring to 0.00001
 
did you set the chances of any injuries occuring to 0.00001

As I recall I did it but might forgot it so I'll double check it when I get home and have access to DB.

Even if I forgot to set "Injury Occurrence" ratio to "0.000001" I see that all players have Injury Proneness = 1 and all AI teams and the testing team have deep backup of identical players for every position so it shouldn't have any significant impact.

P.S. Anyway... Will, thank you for point it out because I might forgot to set it!
 
noticed a major issue with your database on the opposition teams were you have cloned the existing players the original players still have all their PPM's but the cloned players do not have any PPM's I feel this could have a major impact on any results of any tests.
 
As I recall I did it but might forgot it so I'll double check it when I get home and have access to DB.

Even if I forgot to set "Injury Occurrence" ratio to "0.000001" I see that all players have Injury Proneness = 1 and all AI teams and the testing team have deep backup of identical players for every position so it should have any significant impact.

P.S. Anyway... Will, thank you for point it out because I might forgot to set it!
I think you did set it mate I have played over half a season with no injuries at all.
 
noticed a major issue with your database on the opposition teams were you have cloned the existing players the original players still have all their PPM's but the cloned players do not have any PPM's I feel this could have a major impact on any results of any tests.

Yup, it's a very good observation, Will!

I also thought about it and I didn't make any final decision about what to do with it... so I'm still thinking to remove it from all players or add to all players... haven't decided yet.

Btw, I have other small improvements that I'm planning to add for the next version of the testing league. :)
 
Yup, it's a very good observation, Will!

I also thought about it and I didn't make any final decision about what to do with it... so I'm still thinking to remove it from all players or add to all players... haven't decided yet.

Btw, I have other small improvements that I'm planning to add for the next version of the testing league. :)
Just did a quick season with your database and I like the balance of the league and the test team and the fact that no managers got sacked during the season.

here are a few suggestions to improve it:

1. Set the injury chances to 0.00001

2. Fix the PPM issue I would suggest adding the PPM's to the cloned players as I feel they will make the opposition teams more realistic.

3. I would increase the number of times you play each team to 6 that would increase the amount of games played from 44 to 66 and would give a better test sample.

4. Add a few more key staff to the test team especially a director of football.

But all in all a very good test environment.
 
J
3. I would increase the number of times you play each team to 6 that would increase the amount of games played from 44 to 66 and would give a better test sample.

Will, here are some reasons why I have "a short" season in the testing and why I think that it’s better to re-play the season few times than have "a long" season:

1) When you have a very long season(about 70-90) matches and you don't use FMRTE to "freeze" the condition of the players and you also use "Quick Pick" to pick the team then already after 40 matches stacking "Jadedness" start to "kick in" and have impact on the testing result because when you use "Quick Pick" you don’t rotate the team but AI teams rotate their team and they keep "Match Sharpness" for many players in their teams and not only for the starting eleven players and AI teams start to get advantage after 40-50 matches and it might have negative impact on the testing results.

2) When you have "a long" season then it might happen that many AI teams already don't have any chances to achieve something already in the middle of the season and they stop "fighting" but if you have "a short" season then almost all AI teams still have some chances to achieve something better until the very last matches of the season and they keep fighting until the last matches.


"Jadedness" is a hidden attributes that can only be seen by 3rd party programs and it increases for every match played

Jad.png
 
That's true TFF, and is a huge advantage of having FMRTE also, playing 90 matches in a single row with a macro makes our time worth.
 
*****

THE TESTING LEAGUE HAS BEEN UPDATED TO V2


CHANGES:

- Injury Occurrence Ratio has been set to 0.000001

- Now all clone players in the AI teams have the same PPMs as their prototype
 
What's the best Jadedness value for a player to have?
 
Top