Page 20 of 21 FirstFirst ... 10161718192021 LastLast
Results 476 to 500 of 510

Thread: Tsigalko / van Nistelrooy Thread

  1. #476

    Join Date
    27-10-14
    Location
    The Netherlands
    Posts
    7,219
    vCash
    0
    What's his intrinsic score for finishing and technique, I wonder?

  2. #477

    Join Date
    07-08-19
    Posts
    36
    vCash
    500
    Quote Originally Posted by samsami View Post
    What's his intrinsic score for finishing and technique, I wonder?
    I have no idea, i don't use these tools and i don't even know what these intrinsic values are.

    But i'm buying the guy as soon as money are available

  3. #478

    Join Date
    26-09-18
    Location
    Brazil
    Posts
    757
    vCash
    1352
    I created 2 identical players (in terms of intrinsic attributes, and everything else like age, position, etc) in Greame kelly savegame editor, but one I gave CA 1 and PA 1, and the other I gave CA 197 and PA 198:

    Player 1 (CA 1 and PA 1):
    Spoiler!


    Player 2 (CA 197 and PA 198):
    Spoiler!


    Above, you can see they are identical, except for the CA and PA.

    I tested each player under exact the same conditions (trust me), like tactics, etc, and I forced their attributes to always remain constant using savegame editor.

    Player 1 results (I repeated the same test season 7 times):
    Spoiler!


    Player 2 results (I repeated the same test season 5 times):
    Spoiler!


    We can see Player 2 performed better than player 1.

    My conclusion: as opposed to what I used to believe, the value of an intrinsic attribute isn't absolute, as the performance of an intrinsic attribute is boosted by CA. For example, the two players above have 33 intrinsic for dribbling, but the higher CA player is better at dribbling than the lower CA player.

    Now, how much exactly is this CA boost? Im very curious to find out, but Im pretty sure that its less than what the highly CA biased in-game attributes suggest.
    Last edited by MadScientist; 21-11-19 at 07:27 PM.

  4. #479

    Join Date
    27-10-14
    Location
    The Netherlands
    Posts
    7,219
    vCash
    0
    The fact of the matter is that it is no rocket science.

    I have players with no high values (normal values or intrinsic values) perform remarkable well and players with high (intrinsic) values performing less well.

    There are always these kind of exceptions to keep the game interesting.

  5. #480

    Join Date
    18-03-14
    Posts
    1,232
    vCash
    7700
    Quote Originally Posted by MadScientist View Post
    I created 2 identical players (in terms of intrinsic attributes, and everything else like age, position, etc) in Greame kelly savegame editor, but one I gave CA 1 and PA 1, and the other I gave CA 197 and PA 198

    ...

    My conclusion: as opposed to what I used to believe, the intrinsic attributes aren't absolute, as their performance is boosted by CA. For example, the two players above have 33 intrinsic for dribbling, but the higher CA player is better at dribbling than the lower CA player.

    Now, how much exactly is this CA boost? Im very curious to find out, but Im pretty sure that its less than what the highly CA biased in-game attributes suggest.
    Very interesting MadS! Like you I've been thinking a bit about intrinsic vs CA, for me it was in relation to scouts and their JPA/JPP attributes. I hope to do a similar test to you in creating an identical -CA scout and +CA scout and see who performs the best.

  6. #481

    Join Date
    15-12-13
    Posts
    113
    vCash
    500
    [QUOTE=Aeelon;460456]This guy is ruining my life.

    Tactics the AI uses are my creation > realistic NWW not-centered.


    Would you mind sharing that tactic pack? It's exactly what I am looking for: better, but realistic (not-centered) AI tactics. And preferably NWW as I only use NWW myself and of course still want to have a chance of beating the AI. A harder tactic pack with WibWobbed AI-tactics is almost unbeatable if you only use NWW yourself.

  7. #482

    Join Date
    27-10-14
    Location
    The Netherlands
    Posts
    7,219
    vCash
    0
    [QUOTE=Ravanelli;463214]
    Quote Originally Posted by Aeelon View Post
    This guy is ruining my life.

    Tactics the AI uses are my creation > realistic NWW not-centered.


    Would you mind sharing that tactic pack? It's exactly what I am looking for: better, but realistic (not-centered) AI tactics. And preferably NWW as I only use NWW myself and of course still want to have a chance of beating the AI. A harder tactic pack with WibWobbed AI-tactics is almost unbeatable if you only use NWW yourself.
    That would be the ultimate challenge! Play a team like Spurs or Everton (NWW) with an AI-pack installed for the computer operated teams!

  8. #483

    Join Date
    26-09-18
    Location
    Brazil
    Posts
    757
    vCash
    1352
    Quote Originally Posted by MadScientist View Post
    Now, how much exactly is this CA boost? Im very curious to find out
    I found out. When the match engine wants to know how good a player is in a given attribute (like dribbling, passing, positioning, etc) during the current match being processed (either AI-only matches running in background, or human matches in foreground), it calls the function in offset 6EC010 in the cm0102.exe. So I debugged this function in real time using Olly and I discovered what it does.

    Basically, I see there is a new type of attributes, which is used only by the match engine (so it only "exists" during matches). Lets call it "in-match attributes". An "in-match attribute" determine how good a player is in a given attribute (like dribbling, passing, positioning, etc) during his current match.

    The value of an "in-match attribute" of a player is calculated with the following formula during a match:

    Code:
    in_match_value = intrinsic_value/5 + CA/20 + other_factors
    The maximum (and best) in_match_value possible is around 45 (because the maximmun intrinsic value possible is 125, the max CA possible is 200, and the other_factors max value seem to be around 10), but such high in_match_values are extremelly rare to occur. Values above 20 are already very very good, just imagine a 45!
    Note: If the in_match_value results to be negative, its clamped to 0 (and is the worst possible value).

    From those 45 possible points, 25 can come from the intrinsic value (125/5), 10 can come from the CA (200/20), and 10 can come from other_factors.

    So, we can see the boost the CA can give to the performance of an attribute is something between 0 and 10 out of 45 possible points.

    Note: What is the 'other_factors' value? I don't know, I just see there is a huge number of variables that get summed up to form this 'other_factors' value, and its a value usually between 0 and 10. Its probably derived from player morale, condition, mental atts, playing near preferred position, etc, etc, etc, but don't take my work on those, as im just guessing.

    EXAMPLE: Tsigalko has an intrinsic value of 46 for 'finishing' and a CA of 80, and lets suppose during a normal league match his other_factors value was 8. So, his in_match_value for 'finishing' during that match is calculated:
    Code:
    in_match_value_finishing = 46/5 + 80/20 + 8
    in_match_value_finishing = 9.2 + 4 + 8
    in_match_value_finishing = 13.2 + 8
    in_match_value_finishing = 21.2
    So he can really be a beast at 'finishing' during that match, as 21.2 is VERY good. But it doesn't mean his performance as a whole will be good in that match. It just means: If he gets an oportunity to finish, then there is a high chance he will make a good finish, possibly a goal.

    This discovery brings me to a next question which im also curious to find out: now that we know what values the match engine really uses for the attributes, how can we compare it with the attribute values that we see in the player profile screen in the game (the values in the 1 to 20 range)?
    Last edited by MadScientist; 26-11-19 at 04:50 PM. Reason: fixed typo, I mean 21.2, not 23.2

  9. The Following User Says Thank You to MadScientist For This Useful Post:


  10. #484

    Join Date
    18-03-14
    Posts
    1,232
    vCash
    7700
    Wow, great work MadS! Very interesting findings and very well described. That function you describe is huge, it must've taken you ages to pick it apart. Do you have any parts of it annotated anywhere?

  11. #485

    Join Date
    26-09-18
    Location
    Brazil
    Posts
    757
    vCash
    1352
    Quote Originally Posted by saturn View Post
    Wow, great work MadS! Very interesting findings and very well described. That function you describe is huge, it must've taken you ages to pick it apart. Do you have any parts of it annotated anywhere?
    The function is huge but what really matters is just one line of code:

    Spoiler!


    Some other anotations I made:

    Spoiler!


    By the way, there are some places in the match engine code (but out of function 6EC010) where some intrinsic values of attributes are read directly without being summed with CA, but apparently those places don't use the intrinsic value to determine how good the player is in the attribute, those places seem to use it just to calculate something unrelated to that. So I assume function 6EC010 is the only place where it determines how good a player is in a given attribute.
    Last edited by MadScientist; 27-11-19 at 08:21 PM.

  12. The Following User Says Thank You to MadScientist For This Useful Post:


  13. #486

    Join Date
    16-10-14
    Posts
    23
    vCash
    500
    Incredible work ...

  14. #487

    Join Date
    26-09-18
    Location
    Brazil
    Posts
    757
    vCash
    1352
    Below are just some tests I made when I was debuging the formula above:

    As I said, the max in_match_value to occur naturally is around 45. But with some hack in the game .exe I gave a 120 in_match_value for dribbling during a match for the striker player below (much more than 45):



    We can see above he made 33 sucessful dribbles.



    Now I did the same, but for off the ball too (120 in_match_value dribbling AND 120 in_match_value off the ball) and replayed the same match:



    Above, he made 11 headers and 12 shoots, because he is a master in off the ball (so he is always in a good position to receive the ball and has many opportunities to head and shoot to goal).



    Now, for reference, this is how the player plays the same match without any hacks (with his normal values for in_match_values):


  15. #488

    Join Date
    15-12-11
    Location
    Your Mother
    Posts
    32,801
    vCash
    606362
    Can you do this for GK's and find they exact thing that makes a good/great GK? PLEASE??

  16. #489

    Join Date
    26-09-18
    Location
    Brazil
    Posts
    757
    vCash
    1352
    Ha! Guys you need to see this:

    Well, first, I noticed the following attributes receive an additional boost from CA besides the boost from formula above: heading, finishing, marking, positioning, crossing, passing and creativity. I won't get into detail, but its worth mention this additional CA boost can set the in_match_values of those attributes to a max of around 75.

    So far so good, nothing special there. But here comes the interesting part: While I was checking the code of that additional CA boost, I found a really funny bug in the match engine code:

    When the CA of a player is higher than 124 AND the intrinsic value for creativity of that player is higher than 10, then the creativity formula ends up using the intrinsic value of REFLEXES instead of the intrinsic value of CREATIVITY! This is exactly the same bug reported by @Trip on 12/12/12 (very nice date, by the way):

    Quote Originally Posted by Trip View Post
    if anyone cares, I found the exact CA value above which high intrinsic creativity stops yielding high number of passes-per-game. It's 124. CA = 125 and above means it won't work.

    Doing an experiment right now with Gascoigne's regen, who has insane intrinsic creativity (106). Have set his PA to 142, so his CA stays in the low 120's. As long as it's there, he bangs in around 80 (edit: make that 90, just did a calculation) passes a game on average. If his CA rises above 124, as it does occasionally through bouts of training zeal, his number of passes drops dramatically and his creativity basically means fuck-all, he might as well not have the att at all. When his CA returns below 125, up go the passes, up goes the avg.
    This screenshot shows CR7 in Oct2019 data in the original .68 .exe (with the bug):



    In the first season his intrinsic for creativity was 8, so the bug didn't occur and he made lots of assists. In the second season, his intrinsic for creativity grow to 13. So now the bug occur and it started to use the reflex value rather than creativity, so his performance and assists dropped.

    This is really a bug and not something intentional, because its very clear in the code that the programmer forgot to update the intrinsic variable in this specific case, so it ends up using the value of the previous attribute in the loop, which by pure coincidence is reflexes.

    The good news is its a pretty easy fix:

    Spoiler!




    Now, some more examples of the bug, and of the fix:

    Messi and De Bruyne are affected by this bug in Oct2019 data because they have CA higher than 124 and intrinsic for creativity higher than 10. So this is how they perform in vanilla .68 .exe, for reference:
    Spoiler!


    Above, they perform worse than expected with few assists because the creativity value is totally ignored and the reflexes value is being used instead!



    Then to confirm the bug, I changed Messi's intrinsic value of reflexes using Greame kelly savegame editor to the max possible value (125):
    Spoiler!


    Above, we can see Messi instantly becomes assist master and performance improves as a whole (even dribbling increases) because it uses reflex value (125) for creativity due to the bug. Thats still in vanilla .68 .exe (without my fix, but with Messi's reflex increased to the max 125 intrinsic with savegame editor).




    Now, back to the original Messi (with low reflexes attribute). And also see De Bruyne below. This is the .68 .exe with my fix to this bug, so now it really uses the creativity value rather than the reflexes:
    Spoiler!


    We can see above that the fix worked fine, because now their performance improved and assists too, due to value of creativity finally being used instead of reflexes. Keep in mind Messi's intrinsic for creativity is 17 (not the 125 from previous screenshot) so thats why he doesn't perform as good as that, but now performs as expected.




    Quote Originally Posted by Dermotron View Post
    Can you do this for GK's and find they exact thing that makes a good/great GK? PLEASE??
    I bet its a similar bug, GKs look so random in the game, i bet its using the wrong attributes somewhere in the code.
    Last edited by MadScientist; 30-11-19 at 07:41 PM. Reason: added link to .patch file with the fix for the bug

  17. The Following User Says Thank You to MadScientist For This Useful Post:


  18. #490

    Join Date
    06-04-12
    Posts
    91
    vCash
    500
    This is an amazing find - so this could go some way to explain why some low CA/PA players outperform higher value ones.

    Trip even mentions in the "Good players original" thread that players like Dariusz Wocsz (sp.?) outperform Ballack, Effenberg, Habler - despite having apparently worse stats.

    Opening up CM Scout on a new game and look at the players with under 140PA and around 100-124 CA you get the likes of Mike Duff, the 2 x Olympiakos keepers, Luis Andrade, Wael Reyad, Moroni, Gasperoni, Richie Wellens, Matt Hill and a load of other bargain over performers

    Does this Creativity/Reflexes bug also apply to other stat combinations? I wonder if it works in combination with other favourable stats (ie. tackling/marking or finishing/long shots)
    Last edited by 1369phil; 30-11-19 at 12:29 PM.

  19. #491

    Join Date
    26-09-18
    Location
    Brazil
    Posts
    757
    vCash
    1352
    Quote Originally Posted by 1369phil View Post
    so this could go some way to explain why some low CA/PA players outperform higher value ones.
    Yes, thats definitelly one of the reasons why. Creativity is a powerful attribute, it can greatly improve the performance and the number of assists of a player. So with reflexes value being used for creativity of high CA players (above 124 CA), they have a huge disavantage compared to low CA players.

    This bug is also one of the reasons for the drain in goals that happen a few seasons later in the game. After all, with reflexes value being used for creativity, there is no way to fight against the hord of high positioning defenders.

    Quote Originally Posted by 1369phil View Post
    Does this Creativity/Reflexes bug also apply to other stat combinations? I wonder if it works in combination with other favourable stats (ie. tackling/marking or finishing/long shots)
    Not sure I get what you mean by "combination", but I didn't notice any other attribute with a similar bug so far.

  20. #492

    Join Date
    07-08-19
    Posts
    36
    vCash
    500
    So, we buy midfielders with reflexes now? Just in case

  21. #493

    Join Date
    06-04-12
    Posts
    91
    vCash
    500
    Quote Originally Posted by MadScientist View Post



    Not sure I get what you mean by "combination", but I didn't notice any other attribute with a similar bug so far.
    Firstly this is a huge discovery, Trip mentioned Creativity being one of the more important attributes and listed low PA players who outplayed their stats - this is the reason why.

    I meant other intrinsic attributes being adversely effected by incorrectly pairing with non related attributes - ie. Creativity is adversely paired with Reflexes. It sounds like this doesn't work with other attributes - as you've noted, defenders just get better and better as their positioning increases.

    How do I get the code you've written into my game? (Explain it like I'm 5 please!)
    Last edited by 1369phil; 30-11-19 at 07:24 PM.

  22. #494

    Join Date
    26-09-18
    Location
    Brazil
    Posts
    757
    vCash
    1352
    Quote Originally Posted by Aeelon View Post
    So, we buy midfielders with reflexes now? Just in case
    Yes , or apply the patch fix with the link below:

    Quote Originally Posted by 1369phil View Post
    How do I get the code you've written into my game? (Explain it like I'm 5 please!)
    I created a patch for this fix:

    UnlockYourCreativity.patch

    Right click the link above, select "Save link as", then save it somewhere in your computer.

    Then open Nick's tool, click Browse and choose your cm0102.exe (.68, tapani or saturn), then click Tools, Apply patchfile, select the .patch file you saved above. You will see something like "patch applied successfully". Then just close the tool. Dont click Apply. Thats all.

  23. #495

    Join Date
    21-06-14
    Posts
    187
    vCash
    500
    Excellent findings. Since years I thought low CA players with nice number/intrinsics in a particular attribute performed so good because they were just able to do this "attribute" and nothing else they did it way much than the others and better CA players. Your works will open something like a new era of creating and judging players.

  24. #496

    Join Date
    27-10-14
    Location
    The Netherlands
    Posts
    7,219
    vCash
    0
    Quote Originally Posted by MadScientist View Post
    I created a patch for this fix:

    UnlockYourCreativity.patch

    Right click the link above, select "Save link as", then save it somewhere in your computer.

    Then open Nick's tool, click Browse and choose your cm0102.exe (.68, tapani or saturn), then click Tools, Apply patchfile, select the .patch file you saved above. You will see something like "patch applied successfully". Then just close the tool. Dont click Apply. Thats all.
    Does this work for your save-game if you don't have Nick's tool installed but only Saturn's patch?

  25. #497

    Join Date
    03-03-12
    Posts
    3,428
    vCash
    1295
    This is phenomenal. Thank you MC for putting so much research into this. I was always baffled as to why low CA players with high creativity were assist machines. Now it is clear why!

    Also many thanks for posting the fix. I am very excited to add this to my exe, and wait in anticipation for any more findings!

  26. #498

    Join Date
    26-09-18
    Location
    Brazil
    Posts
    757
    vCash
    1352
    Quote Originally Posted by samsami View Post
    Does this work for your save-game if you don't have Nick's tool installed but only Saturn's patch?
    Yes, the fix is compatible with saturn patch and works on existing savegames too. To apply the fix to your existing saturn .exe, you can download nicks tool and follow the steps:

    Quote Originally Posted by MadScientist View Post
    UnlockYourCreativity.patch

    Right click the link above, select "Save link as", then save it somewhere in your computer.

    Then open Nick's tool, click Browse and choose your cm0102.exe (.68, tapani or saturn), then click Tools, Apply patchfile, select the .patch file you saved above. You will see something like "patch applied successfully". Then just close the tool. Dont click Apply. Thats all.

  27. #499

    Join Date
    07-08-13
    Posts
    188
    vCash
    250
    Bravo. Brillant find, thank you for the fix.

  28. #500

    Join Date
    27-10-14
    Location
    The Netherlands
    Posts
    7,219
    vCash
    0

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •