Don't make this mistake and don't test tactics this way...

[TotalFootballFan9]

*** First of all, I want to say that everything that I'll show here can be easily reproduced by anyone and anyone can make sure that it isn't a fake or a lie otherwise I'd never do this thread if I wasn't able to backup my words. ***​

Guys, I want to show you that Will1981's tactic testing method is total waste of your time that only can be and I hope that this knowledge will help to save someone's time.

Knowing that Will1981 tends to remove his threads I copied some necessary data from his testing thread here so it'll be always available.

Here you can download the original save that Will1981 uses to test tactics ---> TTL

Here you can see how Will1981's testing table looks at the current moment :

[TABLE="width: 674, align: center"]
[TR]
[TD="width: 176"]TACTIC[/TD]
[TD="width: 42"]PLD[/TD]
[TD="width: 50, align: center"]WON[/TD]
[TD="width: 47"]DRN[/TD]
[TD="width: 40"]LST[/TD]
[TD="width: 44"]FOR[/TD]
[TD="width: 35"]AG[/TD]
[TD="width: 72"]GD[/TD]
[TD="width: 39"]PTS[/TD]
[TD="width: 41"]YEL[/TD]
[TD="width: 45"]RED[/TD]
[TD="width: 43"]GFC[/TD]
[/TR]
[TR]
[TD="width: 176"]CALLY MAC[/TD]
[TD="width: 42"]70[/TD]
[TD="width: 50"]42[/TD]
[TD="width: 47"]12[/TD]
[TD="width: 40"]16[/TD]
[TD="width: 44"]161[/TD]
[TD="width: 35"]80[/TD]
[TD="width: 72"]81[/TD]
[TD="width: 39"]138[/TD]
[TD="width: 41"]117[/TD]
[TD="width: 45"]4[/TD]
[TD="width: 43"]24[/TD]
[/TR]
[TR]
[TD="width: 176"]Bantons2[/TD]
[TD="width: 42"]70[/TD]
[TD="width: 50"]37[/TD]
[TD="width: 47"]19[/TD]
[TD="width: 40"]14[/TD]
[TD="width: 44"]123[/TD]
[TD="width: 35"]59[/TD]
[TD="width: 72"]64[/TD]
[TD="width: 39"]130[/TD]
[TD="width: 41"]78[/TD]
[TD="width: 45"]3[/TD]
[TD="width: 43"]9[/TD]
[/TR]
[TR]
[TD="width: 176"]V1 Twister - CE[/TD]
[TD="width: 42"]70[/TD]
[TD="width: 50"]37[/TD]
[TD="width: 47"]13[/TD]
[TD="width: 40"]20[/TD]
[TD="width: 44"]108[/TD]
[TD="width: 35"]69[/TD]
[TD="width: 72"]39[/TD]
[TD="width: 39"]124[/TD]
[TD="width: 41"]119[/TD]
[TD="width: 45"]3[/TD]
[TD="width: 43"]8[/TD]
[/TR]
[TR]
[TD="width: 176"]SuperMike - New Era[/TD]
[TD="width: 42"]70[/TD]
[TD="width: 50"]32[/TD]
[TD="width: 47"]19[/TD]
[TD="width: 40"]19[/TD]
[TD="width: 44"]111[/TD]
[TD="width: 35"]65[/TD]
[TD="width: 72"]46[/TD]
[TD="width: 39"]115[/TD]
[TD="width: 41"]134[/TD]
[TD="width: 45"]3[/TD]
[TD="width: 43"]16[/TD]
[/TR]
[TR]
[TD="width: 176"]TFF CONQUEROR (FOR INSTANT RESULT) V5[/TD]
[TD="width: 42"]70[/TD]
[TD="width: 50"]30[/TD]
[TD="width: 47"]19[/TD]
[TD="width: 40"]21[/TD]
[TD="width: 44"]116[/TD]
[TD="width: 35"]92[/TD]
[TD="width: 72"]24[/TD]
[TD="width: 39"]109[/TD]
[TD="width: 41"]146[/TD]
[TD="width: 45"]4[/TD]
[TD="width: 43"]9[/TD]
[/TR]
[TR]
[TD="width: 176"]Lillies[/TD]
[TD="width: 42"]70[/TD]
[TD="width: 50"]31[/TD]
[TD="width: 47"]14[/TD]
[TD="width: 40"]25[/TD]
[TD="width: 44"]127[/TD]
[TD="width: 35"]101[/TD]
[TD="width: 72"]26[/TD]
[TD="width: 39"]107[/TD]
[TD="width: 41"]176[/TD]
[TD="width: 45"]6[/TD]
[TD="width: 43"]9[/TD]
[/TR]
[/TABLE]

As you can see TFF CONQUEROR (FOR INSTANT RESULT) V5 got only 109 points and it's almost the worst result so according to Will1981's testing TFF CONQUEROR (FOR INSTANT RESULT) is a bad approach for FM16.

But what if take a closer look at the way how Will1981 makes his testing, he uses "Quick Pick" before every match and then he uses "Instant Result" to play the match. His testing team consists of the 30 best English players and Will1981 is fully confident that it gives a guaranty that all tactics are provided with equally good players and when he uses "Quick Pick" then the assistant manager pick the best suited players for the tactic.

I have nothing against testing with "Instant Result" even it has some issues because the advantages that "Instant Result" provides compensate these issues BUT using "Quick Pick" without pre-configuration of the testing team "kills" the testing and makes it totally useless and here's why...

Because AI is totally dumb when it comes to pick the team and what make it even worse that he also can't pick the substitution, for example, your tactic has only 1 striker and 3 midfielders but when you use "Quick Pick" AI might pick 3 strikers for the substitutions and the same time he might pick no midfielders at all for the substitutions, yup, AI is that dumb but when you use "Instant Result" the AI might make 3 substitutions just after 45 min of the match and you will be player with the worse players the half of the match. As I already said "Quick Pick" function wasn't designed for any complex task and it just doesn't possess any significant intelligence, it just picks fit players for positions in your tactics that's all and it doesn't have a clue what modification were made to positions in your tactic and what players it needs to performs good so trusting to "Quick Pick" when you picking the team is just ridiculous, especially, when you test tactics because in this case even a small detail might have crucial impact on the final result and you'll see it.

Ok, back to the testing.... I saw that during the Will1981's testing of TFF CONQUEROR (FOR INSTANT RESULT) tactic AI picked totally unsuitable players for some key positions, for example, James Milner as MR or central defender John Stones as DR and the most important AI picked just awful substitution so there weren't any substitutions for some key positions at all and if something happened during the matches then it would be tragic.

Also I don't know that did Will1981's follow or not the instruction change the PIs of ML/MR position depends on their main foot from "Cuts Inside" to "Run Wide".

The logic behind the AI's decision what players to pick for the starting eleven and the substitutions when you use "Quick Pick" is very simple and it isn't hard at all to make AI do the "right" choices and in order to achieve that I simple "blacklisted" some players so AI can't pick them when I use "Quick Pick”, these players were James Milner, Daniel Sturridge, Danny Welbeck, John Stones, Jamie Vardy. After that AI started to pick the players that I wanted for the starting eleven and for the substitutions.

How did I "blacklisted" these players? It's very simple I gave to all players in the team "First Team" squad status and these players I gave "Rotation" status and then I configured the "Quick Pick" such way that AI only can pick players with "First Team" squad status and higher, it can be made this way:

Also I made that all right footed players that might be picked for MR position had "Run Wide" PI instead "Cuts Inside" PIs:

Here you can download a save where I made the modification that I mentioned above so the only what is left to you just push "Quick Pick" and "Instant Result" to test TFF CONQUEROR (FOR INSTANT RESULT) V5 tactic in Will1981's testing league and see what result you'll get ----> TTL ( TFF mod)

If you don't trust me then you can download the original Will1981's testing save and make the modification that I described above by yourself step by step as I explained and then test TFF CONQUEROR (FOR INSTANT RESULT) V5 in Will1981's testing league and see what result you'll get.

Please notice that you'll be still testing by just pushing 2 buttons "Quick Pick" and "Instant Results"... that's all.



Now let's test and see what result we'll get if we test TFF CONQUEROR (FOR INSTANT RESULT) V5 in "proper way":



[TABLE="width: 674, align: center"]
[TR]
[TD="width: 176, align: center"][/TD]
[TD="width: 176, align: center"]TACTIC[/TD]
[TD="width: 42, align: center"]PLD[/TD]
[TD="width: 50, align: center"]WON[/TD]
[TD="width: 47, align: center"]DRN[/TD]
[TD="width: 40, align: center"]LST[/TD]
[TD="width: 44, align: center"]FOR[/TD]
[TD="width: 35, align: center"]AG[/TD]
[TD="width: 72, align: center"]GD[/TD]
[TD="width: 39, align: center"]PTS[/TD]
[TD="width: 41, align: center"]YEL[/TD]
[TD="width: 45, align: center"]RED[/TD]
[TD="width: 43, align: center"]GFC[/TD]
[/TR]
[TR]
[TD="width: 176, align: center"]Proper testing result[/TD]
[TD="width: 176, align: center"]TFF CONQUEROR
(FOR INSTANT RESULT) V5[/TD]
[TD="width: 42, align: center"]70[/TD]
[TD="width: 50, align: center"]43[/TD]
[TD="width: 47, align: center"]12[/TD]
[TD="width: 40, align: center"]15[/TD]
[TD="width: 44, align: center"]148[/TD]
[TD="width: 35, align: center"]61[/TD]
[TD="width: 72, align: center"]87[/TD]
[TD="width: 39, align: center"]141[/TD]
[TD="width: 41, align: center"]137[/TD]
[TD="width: 45, align: center"]2[/TD]
[TD="width: 43, align: center"]16[/TD]
[/TR]
[TR]
[TD="width: 176, align: center"]Will1981's results[/TD]
[TD="width: 176, align: center"]TFF CONQUEROR (FOR INSTANT RESULT) V5[/TD]
[TD="width: 42, align: center"]70[/TD]
[TD="width: 50, align: center"]30[/TD]
[TD="width: 47, align: center"]19[/TD]
[TD="width: 40, align: center"]21[/TD]
[TD="width: 44, align: center"]116[/TD]
[TD="width: 35, align: center"]92[/TD]
[TD="width: 72, align: center"]24[/TD]
[TD="width: 39, align: center"]109[/TD]
[TD="width: 41, align: center"]146[/TD]
[TD="width: 45, align: center"]4[/TD]
[TD="width: 43, align: center"]9[/TD]
[/TR]
[/TABLE]

As you can see the difference is huge 109 point vs 141 points and +24 GD vs +87 GD and the only what was changed just few more suited players in the starting eleven and few more suited player for the substitutions, of course, also I made that all right footed players that might be picked for MR position have "Run Wide" PI in this case but as I already said I don't know did Will1981 do the same or not.

This example once more proves how few suited players for the tactic can turn everything apart and how it's important to provide all tactics in the testing with equally good and suited players otherwise the testing would be just waste of time and you should have total control about what players AI picks when you use "Quick Pick" and not only what players he pick for the starting eleven but also what players he picks for the substitutions.

What make the situation with “Quick Pick” even worse is that AI doesn’t pick equally “bad players” for all tactics in the testing and AI might pick for some tactic “good players” and other tactic “bad players”, for example, I notice that AI makes good choices when he picks players for Inside Forward/Winger/Wing Back and some others roles due to the way how he was coded so any tactic in the testing that has these roles has a huge advantages above other tactics if you use "Quick Pick" to pick the team because as you just saw only few more suited players for the tactic might greatly boost the tactic’s performance.

NOW THE OTHER IMPORTANT THING - HOW RANDOM THE RESULT MIGHT BE IN THE TESTING LEAGUE: ​

I re-tested TFF CONQUEROR (FOR INSTANT RESULT) V5 for more 3 seasons and my worst season was looking like this:

As you can see the final result might vary greatly at 70 matches distance(118 points vs 141, GD 38 vs 87) so 70 matches distance is like a drop in ocean and you need to test no less than 5 seasons and then take average points.

It's possible greatly reduce the randomness in the testing league by preventing the AI chairman from sacking AI mangers until the end of the season because it causes tactic change for AI teams and adds great randomness to the testing but I see that Will1981 still hasn't figured out how to do that and I saw that some AI managers were fired at the mid of the season.

Btw, the testing league that I shared also has this issue but I've "defeated" this issue at a new testing league that I use right now, also, I made many other improvements that help greatly reduce the randomness so I might share it soon.




As Leonardo da Vince said : "He who wishes to be rich in a day will be hanged in a year."

Speaking other words, if someone thinks that it's possible to spend just 30-40 mins for testing a tactic during 50-70 matches in some testing league just pushing 2 buttons: "Quick Pick" and "Instant Result" without making any brain efforts and get all answers that he is looking for then he is greatly wrong and just wasting his time.

 

[DoubleR]

Would you have made this thread if you were on top? Looking at the results I got with the tactics will has posted, his table is pretty much how I would make it. Will has made some great effort to the community with his tactic testing, while you have misused it as your own personal testing squad. Talking about wasting time with your 100 tactics... I'm with will on this

Besides it's not about being presize, it's about giving an indication of which tactics stand out. Because yours is at the bottom doesn't mean it's bad. It's still part of the better ones.

 

[Jadax]

I still can't believe people actually take these testing leagues on face-value. There are so many variables (even with the 'perfectly' made dbs etc) that it just doesn't make sense to base your decision to try/not try a tactic, on the results of some games played by someone else.

Mind you, I'm not trashing anyone who puts time/effort into making/simming these leagues. I'm just saying that what you play can/will be very different from the way someone else does. These leagues have their own value, but should not be the deciding factor imo.

 

[MrGrieve]

I've deleted about 9 different responses to this as im trying to be diplomatic. Best to say i think this thread should be closed. Have a nice day everyone.

 

[Tacticseeker]

I've deleted about 9 different responses to this as im trying to be diplomatic. Best to say i think this thread should be closed. Have a nice day everyone.

TFF puts a lot of time and effort in to making arguably the best tactics for this and the last FM on these forums

when someone goes and posts a testing league that discredits his work by showing incorrect results, you expect him to just sit there and effectively allow his work to be misrepresented?

credit to will for what he's doing, but the results in tff's threads from users show that he definitely has the best most consistent tactics this patch on these forums

maybe if you put that amount of effort in to something only to see it misrepresented you'd understand, have a nice day

 

[dabootzi]

TTF mate no one can argue that your tactics are the best .. Even if they fail in some special testing your tactics still the best cheers

 

[alanbessa]

http://www.fm-base.co.uk/forum/shar...er-picks-team-vs-human-player-picks-team.html

TFF, i'm really sorry that you need to point out this again.

 

[dginola]

I am with tff on the quick pick part. I test with Knap's 442 and the result was way better if I pick the team myself instead of using quick pick. I'll say the less involvement of cpu the better the test is unless we have alphago AI in FM

 

[misterpuzza]

pathetic thread

 

[shirajzl]

One one hand, it can be annoying with TFF and 100 versions of his tactics, some of them not very good.
On the other hand, no one can denied that his tactics are by far the best and most consistent in the long run above all others on the forums.

 

[penza]

Of course its not 100% but what if 'every' tactic was done the 'right' way as you say well then every single tactic would finish above yours anyway^^ the fact that I won the treble with Roma first season and Septuple second season with Cally MAC proves that it does work!

 

[TotalFootballFan9]

Would you have made this thread if you were on top?

First of all, I've already made the similar posts/threads that were pointing at very obvious issues of such testing that Will1981 makes and in the past at the moment when I made the similar thread my tactic was at the top of current Will1981 testing and I intentionally brought attention to that fact in order not to be accused that I was doing It only because my tactic isn't at the top and if you search the forum you can find it.

The second, I clearly understand if some person "wants" to see that I do what I do only because of childish intention like: "/cry my tactic isn't at the top and you all are bad people that making a bad testing! Make it right and put my tactic at the top immediately" then this person will be seeing it like that and I can't do nothing about BUT I want say that any person that is "impersonal" should see that I do this because of as people say : "give a man a fish and you feed him for a day; teach a man to fish and you feed him for a lifetime".

I try to help people understand "how the game works" and "how to make the tactic that would work very efficient" and if I see that someone give wrong advises about how the game works and what needs to be done to make a good tactic then I won't just sit and I'll act to warn people about it.

The third, I only can wish that long time ago when I only started to play FM and create/test tactics, I'd encountered a thread like this because it would saved a ton of my time and I didn't have to figure out everything by myself spending a lot of time on it.

Of course its not 100% but what if 'every' tactic was done the 'right' way as you say well then every single tactic would finish above yours anyway^^ the fact that I won the treble with Roma first season and Sextuple second season with Cally MAC proves that it does work!

Mate, the thing is that the AI doesn't pick the team equally "good" or "bad" for all tactics in the testing and the issue is that for some tactics he might pick suited players and for other he might pick totally unsuited players.

For example, I notice that AI makes good choices when he picks players for Inside Forward/Winger/Wing Back and some others roles due to the way how he was coded so any tactic in the testing that has these roles has a huge advantages above other tactics if you use "Quick Pick" to pick the team because as you can see from the example that I showed at the OP only the presence few more suited players for the tactic in the team might turn everything apart.






Now back to what this thread about... I'll keep point to issues that such testing has and why these issues make such testing useless.

For example, when I made another re-testing somewhere at the middle of the season I saw that AI teams used these formations:

Rest of the World // Argentina​

View attachment 168908 View attachment 168913​

Spain // Brazil​

View attachment 168907 View attachment 168912​

Italy // France​

View attachment 168909 View attachment 168911


Germany
View attachment 168910​

As you can see they all team in the testing league were using 4-2-3-1 or 4-4-2 formations, how do you think does it simulates the real game environment or not? Of course, it does not because 2 important formations that many clubs use in the game are absent, including, the two best clubs Barcelona and Bayern, I'm talking about these formations:

View attachment 168906 View attachment 168905​

In Will1981's testing league Spain and Brazil are supposed to use these formations but somewhere at the middle of the season these teams started "underachieve" and their manager were sacked so the half of test were made without these important formations which greatly reduces any use of such testing and make it useless.

So when you make a testing league you need to be sure that AI teams will keep their managers until the end of the season or the testing wouldn't have any use, btw, it isn't hard to make that AI teams keep their managers until the end of the season, of course, if you bother with it...

 

[sigtekorn]

First of all, I've already made the similar posts/threads that were pointing at very obvious issues of such testing that Will1981 makes and in the past at the moment when I made the similar thread my tactic was at the top of current Will1981 testing and I intentionally brought attention to that fact in order not to be accused that I was doing It only because my tactic isn't at the top and if you search the forum you can find it.

The second, I clearly understand if some person "wants" to see that I do what I do only because of childish intention like: "/cry my tactic isn't at the top and you all are bad people that making a bad testing! Make it right and put my tactic at the top immediately" then this person will be seeing it like that and I can't do nothing about BUT I want say that any person that is "impersonal" should see that I do this because of as people say : "give a man a fish and you feed him for a day; teach a man to fish and you feed him for a lifetime".

I try to help people understand "how the game works" and "how to make the tactic that would work very efficient" and if I see that someone give wrong advises about how the game works and what needs to be done to make a good tactic then I won't just sit and I'll act to warn people about it.

The third, I only can wish that long time ago when I only started to play FM and create/test tactics, I'd encountered a thread like this because it would saved a ton of my time and I didn't have to figure out everything by myself spending a lot of time on it.






Mate, the thing is that the AI doesn't pick the team equally "good" or "bad" for all tactics in the testing and the issue is that for some tactics he might pick suited players and for other he might pick totally unsuited players.

For example, I notice that AI makes good choices when he picks players for Inside Forward/Winger/Wing Back and some others roles due to the way how he was coded so any tactic in the testing that has these roles has a huge advantages above other tactics if you use "Quick Pick" to pick the team because as you can see from the example that I showed at the OP only the presence few more suited players for the tactic in the team might turn everything apart.






Now back to what this thread about... I'll keep point to issues that such testing has and why these issues make such testing useless.

For example, when I made another re-testing somewhere at the middle of the season I saw that AI team used these formations:

Rest of the World // Argentina​

View attachment 1032651 View attachment 1032646​

Spain // Brazil​

View attachment 1032652 View attachment 1032647​

Italy // France​

View attachment 1032650 View attachment 1032648


Germany
View attachment 1032649​

As you can see they all team in the testing league were using 4-2-3-1 or 4-4-2 formations, how do you think does it simulates the real game environment or not? Of course, it does not because 2 important formations that many clubs use in the game are absent, including, the two best clubs Barcelona and Bayern, I'm talking about these formations:

View attachment 1032653 View attachment 1032654​

In Will1981's testing league Spain and Brazil are supposed to use these formations but somewhere at the middle of the season these teams started "underachieve" and their manager were sacked so the half of test were made without these important formations which greatly reduces any use of such testing and make it useless.

So when you make a testing league you need to be sure that AI teams will keep their managers until the end of the season or the testing wouldn't have any use, btw, it isn't hard to make that AI teams keep their managers until the end of the season, of course, if you bother with it...

Also notice some of the players used in the different positions. Quite amusing some of them.
Regarding the testing it is what it is. They bring attention to different tactics which is good - after using them you quickly find out if they work or not. At least for me since I play everything on 'extended'.

Still using TFFs Sunfire V2 with succes

 

[zyndar]

Hi TFF!

Are you planning to create tactics with different Team Instructions from the 8 you are using and which is not using Control/Fluid mentality and team shape?

 

[TotalFootballFan9]

Hi TFF!

Are you planning to create tactics with different Team Instructions from the 8 you are using and which is not using Control/Fluid mentality and team shape?

Hey mate,

The fact that I haven't uploaded tactics with other TIs/PIs/Formations doesn't mean that I haven't tried them.

Heh, I can say that I've tried all possible Formations/TIs/PIs that only can be imagined in FM16 but I saw that everything else performed weaker than my current tactics so uploading a tactic that's weaker than my current tactics makes no sense to me.

Of course, many times before I thought that there wasn't any better tactic than my currents tactic but the amount of tactics and tweaks that I've uploaded so far proves that I was wrong many times before so right now I can easily be wrong about my current tactics and there might be some better tactics and approaches.

Btw, I'm such person that doesn't care about what formation my tactic has or what TIs/PIs it has, the only thing I care about is it must the most efficient approach to play in the game and the fact that my tactics have wide midfielders it's just because I find they are the most efficient way to play in FM16 and for example, in FM14 I used the strikerless tactic without wide midfielders at all because it worked efficient in FM14.

 

[wkdsoul]

One one hand, it can be annoying with TFF and 100 versions of his tactics, some of them not very good.
On the other hand, no one can denied that his tactics are by far the best and most consistent in the long run above all others on the forums.

I've had better luck with Bophonets and Frankyz tbf, you can have success if you know what you're doing

 

[zyndar]

Hey mate,

The fact that I haven't uploaded tactics with other TIs/PIs/Formations doesn't mean that I haven't tried them.

Heh, I can say that I've tried all possible Formations/TIs/PIs that only can be imagined in FM16 but I saw that everything else performed weaker than my current tactics so uploading a tactic that's weaker than my current tactics makes no sense to me.

Of course, many times before I thought that there wasn't any better tactic than my currents tactic but the amount of tactics and tweaks that I've uploaded so far proves that I was wrong many times before so right now I can easily be wrong about my current tactics and there might be some better tactics and approaches.

Btw, I'm such person that doesn't care about what formation my tactic has or what TIs/PIs it has, the only thing I care about is it must the most efficient approach to play in the game and the fact that my tactics have wide midfielders it's just because I find they are the most efficient way to play in FM16 and for example, in FM14 I used the strikerless tactic without wide midfielders at all because it worked efficient in FM14.

Yes mate, I was thinking the same, that this might be the reason. I was only curious about how you would set up a Counter tactic or a Highly Structured one for example. But I can totally agree with you that every one of us have their own preferences.

 

[penza]

TFF - I respect your hard work keep it up and I appreciate the fact that you want to help everyone. I guess that is the most important thing! thanks for your comment.

 

[TotalFootballFan9]

I've had better luck with Bophonets and Frankyz tbf, you can have success if you know what you're doing

Guys, I urge you(even think that it hardly can't be achieved but I'll still try ) don't turn this thread into a flame war about which tactic is the best and which tactic is the worst...

This thread is about how to test tactics in order to find the most efficient approach to play in FM16 and this approach exists because FM16 is a PC game.

 

[wkdsoul]

Guys, I urge you(even think that it hardly can't be achieved but I'll still try ) don't turn this thread into a flame war about which tactic is the best and which tactic is the worst...

This thread is about how to test tactics in order to find the most efficient approach to play in FM16 and this approach exists because FM16 is a PC game.

That really wasn't my intention TTF. Sorry if it read that way.

Np dude.

I've been with you the this for years, I've put the testing leagues in with custom databases, fun to play and see how they pan out but doesn't really tell you much in a applying a system to a group of players.