I've been noticing there're some discussions going on the topic so I'd like to briefly explain the reason why the tactic testing slows down toward the end of each cycle.
First of all, here are some numbers:
There're about 1,144,066 possible formation variations in the game and if we look at the positions then every position has up to 3 duties and up to 8 roles. That gives us about 1,000,000,000,000,000 possible different tactic variations.
Also, every tactic might have up to 100 different Team Instructions and that further increases the number of possible tactic variations to some crazy number like 1,000,000,000,000,000,000 or even more if take into consideration Set Pieces and Individual Player Instructions.
So to test all possible tactic variations it would required running a extreme high-end PC non-stop 24 hours per day for 6,000,000,0000,000,000 days.
As you can see any attempt to brute-force cracking the Match Engine by testing all possible tactic variations would be an absurd idea and that's why when we pick tactics for the testing we use some algorithm that helps us to filter tactics in order to exclude those tactics that are very similar to what's been tested already. Also, at this point our algorithm has learned very well the impact of different Team Instructions/Roles/Duties and it helps to avoid testing something that in theory would make a very small difference like 1%-5% or even smaller.
At this point we're gonna test a tactic only if we see that it offers "really" something new and it's also backed up with outstanding results and feedbacks by many different people.
Zippo said: As you can see any attempt to brute-force cracking the Match Engine by testing all possible tactic variations would be an absurd idea Expand I more than agree with that
Zippo said: There're about 1,144,066 possible formation variations in the game Expand even if you remove some crazy formations like only 1 defender and so on and leave only about 40 formations that can be considered as the most reliable formation then you still have like 60,466,176 different tactics to test and to get this number I took into consideration only the duties and roles but if I would considered the Team Instruction and Set Piece then I can only image how many additional zeros that number would get
Nikko said: even if you remove some crazy formations like only 1 defender and so on and leave only about 40 formations that can be considered as the most reliable formation then you still have like 60,466,176 different tactics to test and to get this number I took into consideration only the duties and roles but if I would considered the Team Instruction and Set Piece then I can only image how many additional zeros that number would get Expand Yeah... obviously, as I said if you want your tactic testing were useful then the only way to achieve that is analyzing its result and fiddling patters because a blind testing of random tactics is a complete waste of time.
For example, if you look at the Passing Directness slider then you'll see it has 5 levels
We've tested many different tactics with the Passing Direction slider set at the lowest level and the highest level and the difference was about 5% at the best.
Can you imagen how little difference it makes when you adjust the slider only by one notch? But from my experience it isn't a rare thing to see people adjust the Passing Direction slider only by one notch and ask to test their tactics again saying that such change turns their tactics into world-beaters.
Hey,
I've been noticing there're some discussions going on the topic so I'd like to briefly explain the reason why the tactic testing slows down toward the end of each cycle.
First of all, here are some numbers:
There're about 1,144,066 possible formation variations in the game and if we look at the positions then every position has up to 3 duties and up to 8 roles. That gives us about 1,000,000,000,000,000 possible different tactic variations.
Also, every tactic might have up to 100 different Team Instructions and that further increases the number of possible tactic variations to some crazy number like 1,000,000,000,000,000,000 or even more if take into consideration Set Pieces and Individual Player Instructions.
So to test all possible tactic variations it would required running a extreme high-end PC non-stop 24 hours per day for 6,000,000,0000,000,000 days.
As you can see any attempt to brute-force cracking the Match Engine by testing all possible tactic variations would be an absurd idea and that's why when we pick tactics for the testing we use some algorithm that helps us to filter tactics in order to exclude those tactics that are very similar to what's been tested already. Also, at this point our algorithm has learned very well the impact of different Team Instructions/Roles/Duties and it helps to avoid testing something that in theory would make a very small difference like 1%-5% or even smaller.
At this point we're gonna test a tactic only if we see that it offers "really" something new and it's also backed up with outstanding results and feedbacks by many different people.
I hope this clears the situation a bit.
Cheers.
Zippo said: As you can see any attempt to brute-force cracking the Match Engine by testing all possible tactic variations would be an absurd idea
I more than agree with that
Zippo said: There're about 1,144,066 possible formation variations in the game
even if you remove some crazy formations like only 1 defender and so on and leave only about 40 formations that can be considered as the most reliable formation then you still have like 60,466,176 different tactics to test and to get this number I took into consideration only the duties and roles but if I would considered the Team Instruction and Set Piece then I can only image how many additional zeros that number would get
Nikko said: even if you remove some crazy formations like only 1 defender and so on and leave only about 40 formations that can be considered as the most reliable formation then you still have like 60,466,176 different tactics to test and to get this number I took into consideration only the duties and roles but if I would considered the Team Instruction and Set Piece then I can only image how many additional zeros that number would get
Yeah... obviously, as I said if you want your tactic testing were useful then the only way to achieve that is analyzing its result and fiddling patters because a blind testing of random tactics is a complete waste of time.
For example, if you look at the Passing Directness slider then you'll see it has 5 levels
We've tested many different tactics with the Passing Direction slider set at the lowest level and the highest level and the difference was about 5% at the best.
Can you imagen how little difference it makes when you adjust the slider only by one notch? But from my experience it isn't a rare thing to see people adjust the Passing Direction slider only by one notch and ask to test their tactics again saying that such change turns their tactics into world-beaters.