• This is a reminder of 3 IMPORTANT RULES:

    1- External self-promotion websites or apps are NOT allowed here, like Discord/Twitter/Patreon/etc.

    2- Do NOT post in other languages. English-only.

    3- Crack/Warez/Piracy talk is NOT allowed.

    Breaking any of the above rules will result in your messages being deleted and you will be banned upon repetition.

    Please, stop by this thread SoccerGaming Forum Rules And Guidelines and make sure you read and understand our policies.

    Thank you!

NeuralNet Hacking

E

Exocyst

Youth Team
I thought I would open a discussion on hacking the files in data/neuralnet/

I'll start off with nnettackle.dat because I want to get the game to be a bit rougher.

1st, Neural nets are basically a way to extrapolate complex decisions given a limited amount of information. A data file for a neural net should contain numeric data points which the program will use to calculate output behaviour based on input information. There are three parts of a neural net. INPUT LAYER, CALCULATION MATRIX, and OUTPUT LAYER. If we interpret the nnettackle.dat as a little endian file containing 32-bit float numbers, we observe some patterns:

Section 1: ¿¿ Input layer parameters ??
The first five numbers are exact integers: 4.0, 7.0, 10.0, 10.0, 1.0

Section 2: ¿¿ Computational layer matrix ??
The next 201 numbers are non-integers valued between -7.5 and +7.5


Section 3: Low Range scale of calculation matrix output updated

Section3[0]= -11.204
Section3[1]= -102.513
Section3[0] Scales the probability of a tackle using an aggressive or poor tackle animation. At the default value of -11.204, these animations are rare. When increased to a range from -100.00000 to -200.00000, these animation sequences are more frequent. I have not determined the purpose of Section3[1] yet.

Section 4: High Range scale of calculation matrix output updated
Section4[0]= 94.148003
Section4[1]= 118.445
Section3[0] and Section4[0] work together to define the range of output from the calculation matrix, so default range is -11.204 <--> 94.148003
Decreasing Section4[0] will decrease the probability that a clean tackle animation sequence will be used. I have not determined the purpose of Section4[1] yet.

Section 5: ¿¿ Output layer parameters ??
The last five numbers are exact integers: 1.0, 180.0, 180.0, 1.0, 98.0
Changing these values has no effect in my test so far.

CHRISTMAS UPDATE
Link: Improved Tackle behavior version 0.1
Changelog(v0.1): Decreased Section3[0] 2-fold to -205.026001
Changelog(v0.1): Decreased Section4[0] by 1/2 to 49.074001
 
regularcat

regularcat

Manager
Moderator
i thought about doing something like this also but there was just too much unknown data to be understood and too little time to try to figure it all out.
 
Fidel Gameplay

Fidel Gameplay

Senior Squad
Hello bro,
good investigation as always!

Anyway, sincerely bro, don't lose your time with neuralnet database files, I've checked these files latest year on 12, when there was a problem of poor number of cpu sliding tackes.

If you compare into an hex editor the 13 neuralnet files with 12 and 11 neuralnet files, the EA DEV Team have never edited/updated these files.
Now I think who you remember the high foul's number in fifa 11 gp/ai.

On my latest gp mods for fifa 12, I've incresed to the max value cpu/ai tackles by putting 0 or 1 to the following strings:

MAX_REASON_FOR_SLIDE_TACKLE = 0
MAX_REASON_FOR_STAND_TACKLE = 1

decresing the max reason to commit a tackle increase the tackles.
I put 0 to the slide tackle string for solve fifa12 sliding tackes issue (I think community remember).

Now I remember that just fifa 13 came out, I tried to use the same method, but I have seen that with 0 on slide tackle string I did not have the same 12's effects. This is very strange and requires study, and perhaps you have discovered the value 180 (high) and I read that this works.

Anyway I've partially incresed cpu tackles with value 1 for both strings for increase all general fouls.
Another tips is to disable new defending face ball (DEFENDING_USE_NEW_FACE_BALL = 0), anyway this depends on the tastes, who likes who's not.
With new face ball disabled the defense is still strong, but commit more fouls.

I hope you have understood the meaning of this post: don't waste your time with these files.
There are many other things that you could work, and then use your enormous skills!

Unfortunately, since new NG gp/ai is difficult to see a cpu sliding tackles from behind, desperate saving.

Another news: on NBA 2k13 there are cpu fouls percent sliders, but not in fifa13, that's really incredible (big ea dev team's failure!).

Example of work for you:
open fifa13.exe into OllyDbg and try to discover some parameters and functions with relative values.
Or try to attach the debug to fifa13 with cheat engine (fifa13 debug attach work only on windows xp).
For attach the debug to fifa 13 with cheat engine use this method: http://forum.cheatengine.org/viewtopic.php?t=2338
(you must change the DebugPort on the process, this avoids the inevitable game crash).

Another work for you:
For example with cheat engine is possible to get physics coordinates and play with them (scan for float), good tutotial here: http://forum.cheatengine.org/viewtopic.php?t=519072 (debug attach isn't required).

bye bye bro
 
E

Exocyst

Youth Team
I have updated the original post with a new file (v 0.1) with increased aggressive tackle animation sequence probability.
 
X

x119

Youth Team
Exocyst;3369517 said:
I have updated the original post with a new file (v 0.1) with increased aggressive tackle animation sequence probability.
Click to expand...


Wow man, thanks a lot!!! I'll try it!
 
regularcat

regularcat

Manager
Moderator
exocyst my gameplay is ready so when you read this pm me your email address and i will send it to you.

i may make a few more tweaks here and there and then maybe test it with your edited neuralnet file but i am more than happy with the tackling and foul counts.
if i can achieve even more fouls with it with the way my files attack fouls already i will use it an credit you of course.
 
O

Oxifan

Club Supporter
What I must do with files ".DS_Store" and "nnettackle.dat"??
Put in "...game/data/neuralnet/"? I have not neuralnet folder. Create it, put the files and regen, right?
Thx.
 
razvanultra55

razvanultra55

Reserve Team
Oxifan;3370584 said:
What I must do with files ".DS_Store" and "nnettackle.dat"??
Put in "...game/data/neuralnet/"? I have not neuralnet folder. Create it, put the files and regen, right?
Thx.
Click to expand...
if u don't have the folder, u create it, and after by using i68regenerator, regenerate files...and play.
 
goldfinger

goldfinger

Youth Team
Thanks for this mod Exocyst.

I think that file .DS_Store is not needed; that is Mac OS file for storing custom attributes of a folder such as the position of icons or the choice of a background image.

I'll test nnettackle.dat and report back.
 
A

AudiS3

Senior Squad
Exocyst;3363974 said:
I thought I would open a discussion on hacking the files in data/neuralnet/

I'll start off with nnettackle.dat because I want to get the game to be a bit rougher.

1st, Neural nets are basically a way to extrapolate complex decisions given a limited amount of information. A data file for a neural net should contain numeric data points which the program will use to calculate output behaviour based on input information. There are three parts of a neural net. INPUT LAYER, CALCULATION MATRIX, and OUTPUT LAYER. If we interpret the nnettackle.dat as a little endian file containing 32-bit float numbers, we observe some patterns:

Section 1: ¿¿ Input layer parameters ??
The first five numbers are exact integers: 4.0, 7.0, 10.0, 10.0, 1.0

Section 2: ¿¿ Computational layer matrix ??
The next 201 numbers are non-integers valued between -7.5 and +7.5


Section 3: Low Range scale of calculation matrix output updated

Section3[0]= -11.204
Section3[1]= -102.513
Section3[0] Scales the probability of a tackle using an aggressive or poor tackle animation. At the default value of -11.204, these animations are rare. When increased to a range from -100.00000 to -200.00000, these animation sequences are more frequent. I have not determined the purpose of Section3[1] yet.

Section 4: High Range scale of calculation matrix output updated
Section4[0]= 94.148003
Section4[1]= 118.445
Section3[0] and Section4[0] work together to define the range of output from the calculation matrix, so default range is -11.204 <--> 94.148003
Decreasing Section4[0] will decrease the probability that a clean tackle animation sequence will be used. I have not determined the purpose of Section4[1] yet.

Section 5: ¿¿ Output layer parameters ??
The last five numbers are exact integers: 1.0, 180.0, 180.0, 1.0, 98.0
Changing these values has no effect in my test so far.

CHRISTMAS UPDATE
Link: Improved Tackle behavior version 0.1
Changelog(v0.1): Decreased Section3[0] 2-fold to -205.026001
Changelog(v0.1): Decreased Section4[0] by 1/2 to 49.074001
Click to expand...


where do I put them? and which of the two should I use?
 
goldfinger

goldfinger

Youth Team
Here are my test result with 24 career matches (12 default and 12 with this mod) on legendary difficulty, 10 minute halves :
....................... default/this mod
my tackles per game: 8.5/7.3
CPU tackles per game: 18.1/14
total tackles per game: 26.6/21.3
my fouls per game: 1.7/1.9
cpu fouls per game: 2.6/1.6
total fouls per game: 4.3/3.5
 
You must log in or register to reply here.


Top