Lucho616

Member
Joined
Mar 12, 2009
Messages
120
Reaction score
0
Points
16
Most Comprehensive Tactic Test in FM-Base History

UPDATE (27-04-2016):


Hello everybody!

We made a new test!!

Now we are testing CA145 and CA120 as you can see in the new spreadsheet.

We are posting a full description of the testing environment.

We have tested like 20 tactics so you have a lot of information to see.

NEW Result table Patch 16.3

------------------------------------------------------------------------------------------------------------------------------------------------

UPDATE (23-03-2016):


Hello again everybody!

Patch 16.3 came out and we (the test team) were immersed in a lot of real life stuff (and we still are). So we have to sorry to everybody that came to the thread asking "Hey! Where are the results for the new version?", you have all the right to complain but we want to make this clear. This isn't our jobs, we can't pay our bills doing this! So, have more patience guys this isn't our priority number one in ours life.

With that cleared up, lets explain how we are going to test.

Database will remain exactly as it was in the last patch and, for now, we will test in the same two different scenarios. First is a sub-top league team (CA 145) and another that will be an underdog team (CA120). You can read more about the database and the different teams in the rest of the OP.

We are thinking of adding a new team, but we aren't sure if that will give us more/better/accurate results of a tactic for every team level (the holy grail of FM = a fully "plug&play" tactic). At the time we make our head clear about this we will communicate it to you.

We move all 16.2 results to a new G document, the URL for looking at those results is at the end of the OP.

At this point there are a lot of tactics to test, we are reviewing the possibility of adding new members to the test team, so if you want to be a part of it send us a PM.

That is all for now. Thanks you all.
------------------------------------------------------------------------------------------------------------------------------------------------

UPDATE (24-12-2015):

Hello everybody!
We have a new Database with a few changes to make this test even better:

All players of the AI teams in the league will have:
- Important Matches 20
- Loyalty 20
- Professionalism 20
- Stamina 20

We'll be testing only with the sub-top team (145CA) because we think that a next patch is round the corner. But with this modification on the DB, the Media Prediction went down to 20th.

We also moved all the results of the V16.1.1 to a new G document, and will start to upload new V16.2 results to the old document. You can find the links to them in the bottom of this OP.

Thanks everybody who had visited the thread.


Introduction:

We (Mr Langvatn, Francicampione and Lucho616) would like to introduce you to the
“Most Comprehensive Tactic Test” in the history of FM-Base. The amount of work this project requires is not possible to do as a solo tester, so we hope that we will “stay a team”. Moreover, we hope that you appreciate all the work we do.

Mr Langvatn had created a custom database in order to test his and others tactics and get the most accurate test results. This database has been fully customized to remove all randomness that can be removed in the game.

Custom Database description (What did we modify of the game):

In Editor (General):
- Removed all injuries
- Removed all loans
- Removed all bans
- Occurrence ratio for all injuries set to 0.000001
- Removed All media stuff
- Removed every other competition in the whole world to increase testing speed (Going instantly from match to match now)
- Test league has 32 teams, meaning 62 matches tested per tactic!
- Test Team has 60 REAL players with 145/120 CA and no registration needed for the league, so ALL are available for every match. This covers every single role equally good.
- Added Transfer Embargo for all clubs

All players in the whole league got these attributes fixed:

This is because we don’t want that only our test team performs stable. If the opposition teams has “luck” or “bad luck”, this would also make our test a lot less accurate.

- Adaptability 20
- Loyalty 20
- Professionalism 20
- Consistency 20
- Dirtiness 10
- Important Matches 10
- Versatility 20
- Injury Proneness 1
- Added the German language on all players (so that they can understand each other)
- Made all players retire after season, to avoid unhappiness
- Set all players nationality to the same as the league to avoid homesickness/unhappiness

The test team:
- Set all “tactical attributes” to 10
- Assistant Manager CA 200 (Testing suggests that this had close to zero significance though)
- 6200 Reputation
- Players: Modified “pressure” to set value: 10
- Players: Removed all PPM’s

In FMRTE:
- Freeze players fitness and morale to 100%, so that “bad luck” in one game doesn’t affect the next. Equal chance every match for every tactic.
- Set the tactic tested to 100% fluid

In Game:
Other testing instructions/information:
- Match preparation is set to defensive positioning
- Training is set to balanced
- Instant result with option "simulate match" (no match plan)
- Using the same save file for every tactic tested (so they get the same schedule)
- There are lots of different player types available, and several of each type
- There are no team talks, as this is a random element that we want to omit.
- Assistant manager does not do OI’s, as this is also a random element that we want to omit.
- Only testing one tactic at a time, to avoid tactics facing each other, as that would result in less accurate results for players who don't play online (like myself).


How do we test tactics?

We run two tests: one with a 145 CA team and another with 120 CA. The 145 CA team emulates a mid-team/sub-top team in the league, and the 120 CA team emulates an underdog team.

With the results of each test we make an “overall” table with the sum of both results. The overall table is much more accurate than both the sub-top and the underdog table, so please keep that in mind when reading the tests. 62 matches isn’t enough. 124 is much better...

Due to the randomness of the game, if you test tactics in a normal game, without any altering, then it would probably take around 500 matches to get accurate test results. With a custom database and the help of FMRTE, you can reduce this number multiplicatively many times. This also means that if you decide to test a tactic yourself, in a normal game, even a season-long result may be misleading as to whether it's a good tactic or not. A good way to see this is to make 2 save files in the same game. Play a season on both save files. Then you will most probably see quite different results, but if you COMBINE these two season-long results, and divide everything by 2, this average will be the most accurate result of the three options. To achieve test results that are as accurate as our tests, you would have to play through that same season on many more save files, then find the average.In addition, there are no teams in the game which have more or less equally good players for all positions. Thus, any real team will favour some tactics more than others, which then results in very inaccurate test results.

What do we get from every test?

Those are the parameters we measure from every tactic:
  • Results (Overall, Home and Away):
    • Wins, Draw, Losses.
    • Goals scored and goals conceded.
  • Team Stats:
    • Avg Possession
    • Passes Completed
    • Pass %
    • Shots on Target
    • Shot Ratio
    • Dribbles Per Game
    • Tackle Ratio
    • Yellow Cards
    • Red Cards
    • Set Pieces (scored and conceded)
      • Corners
      • Direct Free Kicks
      • Indirect Free Kicks

Important Thoughts:

We would like to mention that even though we removed a lot of randomness of the game, there is still a lot of stuff we cannot change, like the match engine for instance (stupid stuff happening that we cannot control).


Another thing we want to clarify is that we cannot say that the top tactic is the BEST tactic… This test, beside being the most comprehensive in the forum, isn’t 100% accurate
(because that's impossible). Even after 124 matches, we need to be aware of that if a tactic was re-tested, it may end up 10-15 points differently than the previous test (this can go both ways).


In addition, keep in mind that this test is done 100% plug and play, without any OI’s or team talks. So, some of the tactics may perform better (or worse) with OI’s. Another thing to remember is that this test of ours does not tell us much about which tactic is the best for a “superstar” team. It can maybe give an indication, but with a superstar team with 9500 reputation, the opposition will play very differently than they do in our test, because in our test, we have set reputation to 5000 for underdog and 6000 for the mid/sub-top team. In the final patch of the game, we might add a test for “superteams”, but it’s just too much work when we know that a patch will come and make all our test results useless.


So... Where are the results?

We made a Google Spreadsheet so you can read the results at the same time as we update with new test results, easily and dynamically.


Result table Patch 16.3

Result table Patch 16.2

Result table Patch 16.1.1




Stay tuned for more, we are testing more tactics.

Big thanks to all of you in the forum.
Best wishes,
 
Last edited:
Great! Thanks for this! By the way, did you guys set 100% match prep/fitness/morale for all teams facing the tactic testing team?
 
A very complete and informative test, good job. I am not surprised with the fact Giving it Laldy and MAHLUK are at the top end of the table.

Great work!
 
Great! Thanks for this! By the way, did you guys set 100% match prep/fitness/morale for all teams facing the tactic testing team?

Sadly, that's not possible to do (FMRTE can only freeze one team at a time). We wish it was :(
 
can u upload the database for us to download ?

Sorry, we want to control that every test is done according to our own instructions/parameters, so we would like to keep this to ourselves. However, I now possess the knowledge to create whichever database I'd like, so maybe later when we are done testing I can create one for "the public".
 
The USV 433 is by far the best tactic made for FM16 but nobody seems to give a ****.
[FM16] USV 4-3-3 Tactic
Trust me, it's way better than the tactics that are tested here...
 
The USV 433 is by far the best tactic made for FM16 but nobody seems to give a ****.
[FM16] USV 4-3-3 Tactic
Trust me, it's way better than the tactics that are tested here...

Can you point us to a full first season result of the tactic? Then we can test it for sure.

We have only focused on FM-base tactics now in the beginning. But we will move on to other forum's tactics soon.

(I must say though, I highly doubt that it's "way better" than all tactics on FM-base, because we are many tacticians here at this forum who have tested pretty much every possible combination of formations, TI's and PI's)

Edit: Lol, I just read the tactic test you linked to. There was one guy who tested each tactic for 17 games in a normal game (this may be the reason why nobody "gives a ****" - in your words :)). Have you even seen the extent of our test? :)
 
Last edited:
The USV 433 is by far the best tactic made for FM16 but nobody seems to give a ****.
[FM16] USV 4-3-3 Tactic
Trust me, it's way better than the tactics that are tested here...

I actually used that tactic when it came out on Steam Workshop, it's very average. I don't know where you got the idea it is the best tactic in the game but considering some of the tactics here on FM BASE helped me win the PL title with ****** Watford then I think that's why no body cares about it.

Also it got me 6th with Liverpool that tactic you linked to.
 
Very good job guys, thnx for the effort. Im a bit dissapointed that we dont have a clear winner tho. Do you guys think Mahlukv1 can be superior to the new one? In previous tests(done by will) v1 was performing slightly better compared those tactics.
 
Very good job guys, thnx for the effort. Im a bit dissapointed that we dont have a clear winner tho. Do you guys think Mahlukv1 can be superior to the new one? In previous tests(done by will) v1 was performing slightly better compared those tactics.

There will never be a "clear winner" as that would require a 20+ point lead on the others.... About Mahluk v1, perhaps we can test that later, but first we will test tactics from more creators.
 
Last edited:
Good effort guys - I have been having success with gatling project (updated)
 
Top