TotalFootballFan9
Member
- Joined
- Dec 10, 2013
- Messages
- 4,113
- Reaction score
- 201
- Points
- 63
Hi guys,
First of all, I want to apologize for posting this thread at this section but I see that it became kind of a tradition to post different stuff about tactic testing at this section and I see that people are fine with that.
I find that this post could be very useful people that want to learn how to make the most efficient tactics to play FM and I see there are lot of such people at this forum.
Obviously, this subject is very massive and in theory there’s no limits to the size of the info that covers it but I want to keep this post as short as possible and focus only on the areas that I find to be the most important.
Also, I want to add that I understand that there are a lot of people at this forum that aren’t interested at all in making and playing with the most efficient tactics and they prefer to play with tactics that they like due to other factors than efficiency so I want to say to such people that it might be a good idea to stop reading this post from this point because all information that follow bellow might look very boring and geeky to you.
Heh, If you are still reading this then it means that you are interested in making the most efficient tactics to play FM and I suggest that the first thing you should do is build your testing league where you’ll be testing your tactics.
Someone may ask why building a testing league and testing tactics are necessary things? Why I just can’t sit and think a tactic that would seem completely logical and flawless in my head and that’s all? To answer these questions I can say that about 95% of all my tactical ideas that looks absolutely perfect and flawless in my head and that also seem work perfectly when I try them for few matches during a normal gameplay go into garbage can after I give them a comprehensive testing and notice that it happens with me who might be called a very experienced FM player that knows all the ins and outs in FM.
Also, it important to understand that FM is just a PC game and it’s close to real life like GTA 5 is close to real life and things that you know work well in real life football might not work at all in FM and if you’d ask me then I say that FM is very far away from the real life football and if you are going to build the most efficient tactic to play FM then you’d better forget all you know about real life football… I know… it’s sad but true.
Now let’s talk about how your testing league should look like to be capable of providing you with such data that can show even a very small difference in efficiency between tactics because if we are going to build the most efficient tactics to play FM then we need such great accuracy.
You might ask how such great accuracy can be achieved? The answer is very obvious… we should remove all random factors that affect the results but not related to tactics.
Here’s a list of the most significant random factors that greatly affects the result but they aren’t related to tactics:
1) All injuries should be restricted from happening during the testing.
2) The condition and morale of all players in all teams should be set the same level before every match, it means that all players in all teams should start every match with the same level of morale and conditions.
3) All teams should use the same starting eleven for all matches, there shouldn’t be any rotation at all.
4) The AI managers and you must be protected from getting sacked before the testing ends.
5) The consistency attribute of all players in all teams should be set at the maximum possible level.
6) All teams in the testing league should be equally strong, there shouldn’t be teams that are much stronger or much weaker than other teams in the league and if you want to find out how tactics works when you use them with underdog clubs and favorite clubs then you can just change the reputation of your team and the reputation of the AI teams to force the AI teams to play defensively or attackingly against you.
7) The reputation of your team and the reputation of the AI teams must be set at such level that its changing during the test won’t cause any changes in the tactical approaches that AI teams use against your team during the testing, speaking other words, all AI teams should use the same tactics against your team all the time during the testing.
In order to further greatly increase the efficiency of our testing league without causing any significant distraction to the testing environment we can add one copy of our testing team into the testing league that will double the number of matches that we test per one test round; so there’ll be 2 teams under our control in the testing league, let’s call them “Team A” and “Team B” as I said “Team B” is just a copy of “Team A”, it has the same players, coaches and so on.
Ok, let’s assume that we have built a testing league with the properties listed above so what’s the next step? The next step is find out how random is our testing league, speaking other words, how many matches we need to test in order to get such data that will be accurate enough to catch a very small difference in efficiency between tactics. Let’s test a random tactic in our testing league and check the results:
As we can see at the Screenshot 1 “Team A” and “Team B” got almost the same result.
“Team A” got 170 Points and “Team B” got 175 Points so the difference in points is just about 3%.
“Team A” got +67 G.D. and “Team B” got +68 G.D. so difference in G.D. is less than just about 1%.
When you got such results and look at the them then you might think that you built a perfect testing system and it’s time to celebrate but I must warn you that any celebration will be a mistake and instead of celebration you should repeat the test no less than 8-10 times and if you do that then you unavoidable encounter a season that will be looking like the season that is shown at the Screenshot 2 below, the stats after 16 matches mgiht look like this:
As we can see at the Screenshot 2 above “Team A” got 38 Points after 16 matches and “Team B” got 20 Points after 16 matches so the different in Points is about 48%.
“Team A” got +28 G.D. and “Team B” got +2 G.D. so the different in G.D. is about 97%
I'd say that the difference is huge and that after we removed many random factors that that persists during normal gameplay
Let’s look at the stats of Scoring Chances in those 16 matches for both teams: “Team A” and “Team B”:
As we can see at the screenshots above in those 16 matches both teams created almost the same amount of Scoring Chances “For” and had almost the same amount Scoring Chances “Against” but despite that they have very different results after 16 matches in the league… Why? The answer is obvious it’s because “Team A” was extremely lucky, it scored all its chances and the opposition teams missed all their chances and everything was opposite for “Team A”, it was extremely unlucky, it missed all its chances and the opposition scored all their chances and as the result we see such big different in Points and in G.D. after 16 matches played in the league.
Another important factor that greatly contributed to the difference in the result of “Team A” and “Team B” is that all teams in the testing league are equally strong which excludes any possibility of domination one team over other team in a match due to significant difference in the players’ quality and the only what makes one team win over other team is the difference in the efficiency of their tactics and as my practice shows even the most efficient tactics in FM18 don’t give any extraordinary advantage over AI teams and when you use just an average tactic or just a good tactic vs AI team and you both have similar quality players then the result will be very random, I’d say you won’t have more than 60% on success over the AI team in this case.
Ok, let’s finish the season and look at the results:
As we can see at the screenshot above, after 95 matches the result of “Team A” and “Team B” became much closer than they were when we checked them after 16 after matches but there still difference in results between "Team A" and "Team B" that is significantly bigger than we got during our first test that is shown at the Screenshot 1.
What conclusion can we make from what we have seen? I’d say the conclusion is that even if we make a “perfect” testing league where many random factors are eliminated then we still might get a great randomness after 16 matches and we still might get a quite high randomness after 95 matches, of course, all that is true if all teams in the testing are equally strong.
As my practice shows, I get an accurate result that is capable of revealing small tweaks’ efficiency only if I give a tactic 3 full runs in the testing league and take the average of the 3 runs results so it means that a tactic must be tested for about 570 matches in a “perfect” testing league in order to reveal small tweaks’ efficiency… can you imagine how hard it to find out small tweaks’ impact if you try to catch it during normal game when there are 100 times more random factors than in the testing league.
Also, I see some people think that if you want to give a tactic a real stress test that will reveal its efficiency then you must try it the lowest league possible but I can say that such thinking is a total mistake because in the lower leagues AI managers are very poor and the players are very poor so they aren’t capable of exploiting any weakness in your tactic and as my practices show all tactics work few times better in the worst leagues than they works in the best leagues.
If you still prefer to test tactics during regular gameplay then It might be a good idea to test tactics with small clubs or mid clubs in the best leagues but such testing also might be tricky because when you start to play with a small club or a mid club then it means that your club’s reputation is small and almost all AI teams in the league have higher reputation than you so they will be using their attacking approaches against you but if you use a good tactic then it’s inevitably that you will be doing good and overachieving so your reputation will be increasing and more and more AI teams will be using defensive approaches against you which might greatly reduce the efficiency of your tactic, so you need to find how your tactic work when you use it with a low reputable team and a high reputable team.
Please, notice that I don’t say that it’s impossible to test tactics during normal gameplay and that what the most people do, I just want to show how random FM is and how many efforts it requires to put into the testing if you want to see the effect from small tweaks.
At the end I want to give another important tip that greatly helps to find a small difference in tactics' efficiency when you analyze testing data: "Pay more attention to G.D. than Points".
For example, “Team A” use “Tactic A” and won 3 matches during some very small testing that consist only of 3 matches and the results of those matches are below:
1 – 0
1 – 0
1 – 0
So the stats of “Team A” after the testing look like this: Points = 9 and G.D. = +3
“Team B” use “Tactic B” also won 3 matches during the same testing and the result of those are below:
10 – 0
10 – 0
10 – 0
So the stats of “Team B” after the testing look like this: Points = 9 and G.D. = +30
If we look at the points then we can conclude that both “Tactic A” and “Tactic B” are equally efficient because the teams using them got the same amount of points 9 but if we look at G.D. then we can conclude that the “Tactic B” is 10 times more efficient than “Tactic A” because the G.D. of “Tactic B” is 10 times better that the G.D. of “Tactic A”
So as you can see G.D. is much more flexible than Points and it reveals much better the difference in tactics’ efficiency.
First of all, I want to apologize for posting this thread at this section but I see that it became kind of a tradition to post different stuff about tactic testing at this section and I see that people are fine with that.
I find that this post could be very useful people that want to learn how to make the most efficient tactics to play FM and I see there are lot of such people at this forum.
Obviously, this subject is very massive and in theory there’s no limits to the size of the info that covers it but I want to keep this post as short as possible and focus only on the areas that I find to be the most important.
Also, I want to add that I understand that there are a lot of people at this forum that aren’t interested at all in making and playing with the most efficient tactics and they prefer to play with tactics that they like due to other factors than efficiency so I want to say to such people that it might be a good idea to stop reading this post from this point because all information that follow bellow might look very boring and geeky to you.
*****************************
Heh, If you are still reading this then it means that you are interested in making the most efficient tactics to play FM and I suggest that the first thing you should do is build your testing league where you’ll be testing your tactics.
Someone may ask why building a testing league and testing tactics are necessary things? Why I just can’t sit and think a tactic that would seem completely logical and flawless in my head and that’s all? To answer these questions I can say that about 95% of all my tactical ideas that looks absolutely perfect and flawless in my head and that also seem work perfectly when I try them for few matches during a normal gameplay go into garbage can after I give them a comprehensive testing and notice that it happens with me who might be called a very experienced FM player that knows all the ins and outs in FM.
Also, it important to understand that FM is just a PC game and it’s close to real life like GTA 5 is close to real life and things that you know work well in real life football might not work at all in FM and if you’d ask me then I say that FM is very far away from the real life football and if you are going to build the most efficient tactic to play FM then you’d better forget all you know about real life football… I know… it’s sad but true.
*****************************
Now let’s talk about how your testing league should look like to be capable of providing you with such data that can show even a very small difference in efficiency between tactics because if we are going to build the most efficient tactics to play FM then we need such great accuracy.
You might ask how such great accuracy can be achieved? The answer is very obvious… we should remove all random factors that affect the results but not related to tactics.
Here’s a list of the most significant random factors that greatly affects the result but they aren’t related to tactics:
1) All injuries should be restricted from happening during the testing.
2) The condition and morale of all players in all teams should be set the same level before every match, it means that all players in all teams should start every match with the same level of morale and conditions.
3) All teams should use the same starting eleven for all matches, there shouldn’t be any rotation at all.
4) The AI managers and you must be protected from getting sacked before the testing ends.
5) The consistency attribute of all players in all teams should be set at the maximum possible level.
6) All teams in the testing league should be equally strong, there shouldn’t be teams that are much stronger or much weaker than other teams in the league and if you want to find out how tactics works when you use them with underdog clubs and favorite clubs then you can just change the reputation of your team and the reputation of the AI teams to force the AI teams to play defensively or attackingly against you.
7) The reputation of your team and the reputation of the AI teams must be set at such level that its changing during the test won’t cause any changes in the tactical approaches that AI teams use against your team during the testing, speaking other words, all AI teams should use the same tactics against your team all the time during the testing.
In order to further greatly increase the efficiency of our testing league without causing any significant distraction to the testing environment we can add one copy of our testing team into the testing league that will double the number of matches that we test per one test round; so there’ll be 2 teams under our control in the testing league, let’s call them “Team A” and “Team B” as I said “Team B” is just a copy of “Team A”, it has the same players, coaches and so on.
*****************************
Ok, let’s assume that we have built a testing league with the properties listed above so what’s the next step? The next step is find out how random is our testing league, speaking other words, how many matches we need to test in order to get such data that will be accurate enough to catch a very small difference in efficiency between tactics. Let’s test a random tactic in our testing league and check the results:
Screenshot 1
As we can see at the Screenshot 1 “Team A” and “Team B” got almost the same result.
“Team A” got 170 Points and “Team B” got 175 Points so the difference in points is just about 3%.
“Team A” got +67 G.D. and “Team B” got +68 G.D. so difference in G.D. is less than just about 1%.
When you got such results and look at the them then you might think that you built a perfect testing system and it’s time to celebrate but I must warn you that any celebration will be a mistake and instead of celebration you should repeat the test no less than 8-10 times and if you do that then you unavoidable encounter a season that will be looking like the season that is shown at the Screenshot 2 below, the stats after 16 matches mgiht look like this:
Screenshot 2
As we can see at the Screenshot 2 above “Team A” got 38 Points after 16 matches and “Team B” got 20 Points after 16 matches so the different in Points is about 48%.
“Team A” got +28 G.D. and “Team B” got +2 G.D. so the different in G.D. is about 97%
I'd say that the difference is huge and that after we removed many random factors that that persists during normal gameplay
Let’s look at the stats of Scoring Chances in those 16 matches for both teams: “Team A” and “Team B”:
As we can see at the screenshots above in those 16 matches both teams created almost the same amount of Scoring Chances “For” and had almost the same amount Scoring Chances “Against” but despite that they have very different results after 16 matches in the league… Why? The answer is obvious it’s because “Team A” was extremely lucky, it scored all its chances and the opposition teams missed all their chances and everything was opposite for “Team A”, it was extremely unlucky, it missed all its chances and the opposition scored all their chances and as the result we see such big different in Points and in G.D. after 16 matches played in the league.
Another important factor that greatly contributed to the difference in the result of “Team A” and “Team B” is that all teams in the testing league are equally strong which excludes any possibility of domination one team over other team in a match due to significant difference in the players’ quality and the only what makes one team win over other team is the difference in the efficiency of their tactics and as my practice shows even the most efficient tactics in FM18 don’t give any extraordinary advantage over AI teams and when you use just an average tactic or just a good tactic vs AI team and you both have similar quality players then the result will be very random, I’d say you won’t have more than 60% on success over the AI team in this case.
Ok, let’s finish the season and look at the results:
As we can see at the screenshot above, after 95 matches the result of “Team A” and “Team B” became much closer than they were when we checked them after 16 after matches but there still difference in results between "Team A" and "Team B" that is significantly bigger than we got during our first test that is shown at the Screenshot 1.
What conclusion can we make from what we have seen? I’d say the conclusion is that even if we make a “perfect” testing league where many random factors are eliminated then we still might get a great randomness after 16 matches and we still might get a quite high randomness after 95 matches, of course, all that is true if all teams in the testing are equally strong.
As my practice shows, I get an accurate result that is capable of revealing small tweaks’ efficiency only if I give a tactic 3 full runs in the testing league and take the average of the 3 runs results so it means that a tactic must be tested for about 570 matches in a “perfect” testing league in order to reveal small tweaks’ efficiency… can you imagine how hard it to find out small tweaks’ impact if you try to catch it during normal game when there are 100 times more random factors than in the testing league.
Also, I see some people think that if you want to give a tactic a real stress test that will reveal its efficiency then you must try it the lowest league possible but I can say that such thinking is a total mistake because in the lower leagues AI managers are very poor and the players are very poor so they aren’t capable of exploiting any weakness in your tactic and as my practices show all tactics work few times better in the worst leagues than they works in the best leagues.
If you still prefer to test tactics during regular gameplay then It might be a good idea to test tactics with small clubs or mid clubs in the best leagues but such testing also might be tricky because when you start to play with a small club or a mid club then it means that your club’s reputation is small and almost all AI teams in the league have higher reputation than you so they will be using their attacking approaches against you but if you use a good tactic then it’s inevitably that you will be doing good and overachieving so your reputation will be increasing and more and more AI teams will be using defensive approaches against you which might greatly reduce the efficiency of your tactic, so you need to find how your tactic work when you use it with a low reputable team and a high reputable team.
Please, notice that I don’t say that it’s impossible to test tactics during normal gameplay and that what the most people do, I just want to show how random FM is and how many efforts it requires to put into the testing if you want to see the effect from small tweaks.
At the end I want to give another important tip that greatly helps to find a small difference in tactics' efficiency when you analyze testing data: "Pay more attention to G.D. than Points".
For example, “Team A” use “Tactic A” and won 3 matches during some very small testing that consist only of 3 matches and the results of those matches are below:
1 – 0
1 – 0
1 – 0
So the stats of “Team A” after the testing look like this: Points = 9 and G.D. = +3
“Team B” use “Tactic B” also won 3 matches during the same testing and the result of those are below:
10 – 0
10 – 0
10 – 0
So the stats of “Team B” after the testing look like this: Points = 9 and G.D. = +30
If we look at the points then we can conclude that both “Tactic A” and “Tactic B” are equally efficient because the teams using them got the same amount of points 9 but if we look at G.D. then we can conclude that the “Tactic B” is 10 times more efficient than “Tactic A” because the G.D. of “Tactic B” is 10 times better that the G.D. of “Tactic A”
So as you can see G.D. is much more flexible than Points and it reveals much better the difference in tactics’ efficiency.
Last edited:
