Mate, in this testing league “Randomness” is about 100 times less compare with the real game-play, take only one improvement in this testing league that forces all AI teams in the testing league to use almost the same starting eleven for all matches, also, the testing teams use the same starting eleven for all matches, only that improvement already immensely reduces “Randomness” in the testing league compare with the real game-play.
But you see that even in such “perfect” testing league there might be like about 30 points difference between Team A and Team B at the end of the season when you test the same tactic with both teams, all that clearly shows that it’s just impossible to see the effect from small changes if you test only for 85 matches even in such “perfect” testing league where “Randomness” is about 100 times less compare with the real game-play.
If you compare two very similar tactics or looking to see the effect from very small change then 85 matches or 170 matches is like a drop in ocean even in such “perfect” testing league and in order to see it you need to test for about 10-20 seasons; 1700 – 3400 matches.
As you can see FM is very well protected from “Brutal” force approaches, you just can’t make some small change then test for few matches and clearly see what effect it has a positive or negative; as you can see even 85 matches in a “perfect” testing league don’t give you answer.
P.S. Don’t treat the results of Team A and Team B as two separate results, treat them as one result and sum up the results of Team A and Team B, these teams play between each other during the season and during these matches they share 15 possible points and it's random which team gets points so you should always sum up the results of Team A and Team B and the end of the season.
