1. Do you need support for Assetto Corsa Competizione? Please use the proper forum below and ALWAYS zip and attach the WHOLE "Logs" folder in your c:\users\*youruser*\AppData\Local\AC2\Saved. The "AppData" folder is hidden by default, check "Hidden items" in your Windows view properties. If you report a crash, ALWAYS zip and attach the WHOLE "Crashes" folder in the same directory. Do not post "I have the same issue" in an existing thread with a game crash, always open your own thread. Do not PM developers and staff members for personal troubleshooting and support.
  2. As part of our continuous maintenance and improvements to Assetto Corsa Competizione we will be releasing small updates on a regular basis during the esports season which might not go through the usual announcement process detailing the changes until a later version update where these changes will be listed retrospectively.
  3. If ACC doesn't start with an error or the executable is missing, please add your entire Steam directory to the exceptions in your antivirus software, run a Steam integrity check or reinstall the game altogether. Make sure you add the User/Documents/Assetto Corsa Competizione folder to your antivirus/Defender exceptions and exclude it from any file sharing app (GDrive, OneDrive or Dropbox)! The Corsair iCue software is also known to conflict with Input Device initialization, if the game does not start up and you have such devices, please try disabling the iCue software and try again. [file:unknown] [line: 95] secure crt: invalid error is a sign of antivirus interference, while [Pak chunk signing mismatch on chunk] indicates a corrupted installation that requires game file verification.
  4. When reporting an issue with saved games, please always zip and attach your entire User/Documents/Assetto Corsa Competizione/Savegame folder, along with the logs and the crash folder (when reporting related to a crash).

Minorating.com (drivers thread)

Discussion in 'Chit Chat Room' started by Minolin, Aug 7, 2015.

  1. martcerv

    martcerv Alien

    I cant see the server especially a regarded popular one having any negative impact on rating. Maybe the racing isnt as close there so your not getting much in the attact range. If its mostly a follow the leader type thing with people stretching larger gaps to others there wont be much attack range.
     
  2. Berniyh

    Berniyh Alien

    Edit: Quoted comment was removed.
     
    Last edited: Aug 6, 2016
  3. Berniyh

    Berniyh Alien

    I rather thought about a positive impact on rating.

    But it's hard to tell without proper statistics.
     
  4. Andy-R

    Andy-R Alien

    as Marty said, most likely down to attack and combat range. i have higher cp than both of you but we enjoy quite a bit of close racing on aloog1

    MR.jpg
     
  5. liakjim

    liakjim Alien

    Ok, good to know.

    So another question would be : what are the limits of each grade ( in other words : at which point someone will go from A/B to C?)
     
  6. Berniyh

    Berniyh Alien

    http://www.minorating.com/Grades

    iirc, it's a bit outdated (I think @Minolin said something like that recently), but it'll give you an idea.
     
    liakjim and Andy-R like this.
  7. V8_KB

    V8_KB Alien

    Oh how awesome is that new stats thingy. And I'm not saying it because I'm B at speedwise and A at MR. :p
    Thanks Minolin.
     
    tabis likes this.
  8. liakjim

    liakjim Alien

    Ok, you mean this
    upload_2016-8-6_13-40-32.png

    It's pretty clear now, thanks.
     
    brioche and Andy-R like this.
  9. V8_KB

    V8_KB Alien

    Instead of restricted grades I have "error" next to sessions from these two servers. Is it just a bug or something is wrong with them? Other values (driven, %, CP) seem to be OK, I don't actually know if they are OK, but at least there are numbers instead of "error" so they work.

    Przechwytywanie.JPG
     
    TDS likes this.
  10. liakjim

    liakjim Alien

    Yes I have some errors too
     
  11. liakjim

    liakjim Alien

    upload_2016-8-6_14-9-34.png
     
  12. LeiF

    LeiF Alien

    I think this is amazing. Thank you so much for releasing these stats :D I seriously like statistics!

    I know there have already been a flood of requests and questions because of this so I'll keep it short.
    Possible to see the overall regarded percentages of normal, combat and attack, as on the speedwise pie chart?
    And maybe sort the sessions by clicking column titles? However seeing the regarded sessions is the main thing to help understand what is going on, nice one Mino!
     
    tabis and Andy-R like this.
  13. martcerv

    martcerv Alien

    Its possible these server could be wrongly setup, not sure if they are using MR 2.0 but you should be able to check in server list. I remember Minolin saying MR2.0 is checking to see if description and actual filtering is correct. It could also be the plugin isnt working on those servers or for some reason the MR database is rejecting the data due to its trust token thingy.
     
  14. Mathieu Labbé

    Mathieu Labbé Hardcore Simmer

    Amazing Minolin, that tool is very precious in understanding what the hell is going on :) Almost all of the servers on my list have 'error' in the restriction column like the other have said, not sure whats that saying.

    Thanks for your hard work.
     
    After_Midnight likes this.
  15. liakjim

    liakjim Alien

    Well , the errors are in this server :

    upload_2016-8-6_15-38-29.png

    which is using 1.4.3

    So I have to inform the admin to install plugin 2.0?
     
  16. Minolin

    Minolin Staff Member KS Dev Team

    Uh. I did expect you like numbers, but this... I wanted a bit less traffic in this thread so I can keep up with the questions and comments :D


    Quick jump over the questions:

    Speedwise stats/graph: Most of the stats are speedwise stats, ie when you've last driven on a speedwise-monitored server. That's not necessarily a MR one.
    For the grade indication it uses some kind of estimated progress directly from MR (a private function I built for heNNa), but I never said it's 100% accurate. The way I calculate the grade is unfortunately not progress-friendly, at least for A and B.


    More stats (like Average total km of the players (last 24h/month/year?) ):

    No, MR 2.0 is not the right tool for that. Many limitations, especially performance and development time.


    Can you please explain how Attack % is exactly calculated?

    Exactly? No. The disadvantage of giving you data is that I also give a help for exploitations - you can try something and now directly see if it worked.
    The attack/combat exploitation protection is some kind of simple threshold mechanics, the less exploiters know about it the more effort it will mean to actually find a suitable way (if there is any).

    In general I'm very sure you will have high numbers when racing in a pack, just do your best not to be too defensive and have fun. @V8_KB is a good example.
    But no, it doesn't exaclty count passes, although you will have more Attack % when passing often. Also keep in mind the Attack % you can see is simplified (deliberate information loss) compared to the real formula.


    Does Minorating currently give increased contact points for the first absolute contact in a race session?

    Nope. I understand the logic behind and it might be a good idea, but I think the explicit mass accident count is similar and where you are thinking to.


    What about ping? Does the system account if a collisions occur with a guy that has 800-2000 ping?

    Good question. There are indicators of a huge ping or lag, and collisions beyond a specific threshold are ignored. Not 100% confident that it works well, as it is hard to test.
    But I can't tell who had the ping/lag, it's more the difference between the contact data from each client.


    the helicorsa app going totally nuts at race start

    Yes this is a result of the improved prediction code. Something changed in the velocity vectors, but this is just a mistake of mine. I fiddled around until I could ignore standing cars with a hack, now it doesn't work anymore.
    But well, it's not really a huge problem so I focus on other things.


    Comparable stats + "I'm grade B and he is grade A"

    This is just a result of the simplification. Yes, Attack% is very important here. And: The Attack % actually includes two values (Attack+Combat), where Combat is weighted/multiplied by 4 and has TONS of influence on your grade.
    There are also more parameter like the mass accidents.

    But in general you are right, it's something about Bad Points / Distance. But I don't think you (or the potential exploiter) have much chance to re-create the actual model based on this data.
    Just so much: Until up to Grade C the Distance is Total and Attack, for B it's almost only Attack. Even sharper for A.


    Do you have a magic grade A?

    Funilly - No. Nobody has, although I had to implement a AUTH wildcard for @Lord Kunos because he couldn't be bothered to do his 40 N grade km :D
    I'm usually B during testing, but didn't do much of this lately. I also dropped to C once.


    I have "error" in the Grade column

    Yep, WIP. A few days ago I published a Plugin beta (2.0), with that one MR reads the AUTH restriction so we don't need to rely on the server name.
    Many work, some do not. Need to check why, but that's usual part of beta testing. Doesn't affect you in any way (aside from reading a text value).
    And no, the beta plugin is totally optional. Please don't ask the admins to use it, we will update it several times until it is fine.




    In general, most of the "can we have X" answers will be "no". Sorry, but this was just a way to provide you with some basic stats where you can see what sessions look worse and why your grade might have dropped. Actually I just gave up some obscurity to have more time for development ^^
    The MR Lobby might show you even more, I'm just 20x faster in desktop programs compared to web stuff. But don't expect 100% insight. MR works when you are driving well and have a good grade without obeying any special rules of the system. So far it really did. For year now, right?
     
    bigbawmcgraw, baboon, Berniyh and 9 others like this.
  17. liakjim

    liakjim Alien

    Thank you!!!
     
    brioche likes this.
  18. brioche

    brioche Racer

    Thanks for sharing.

    20m attack range is an interesting concept and it fits well for A-B drivers which have a clear idea of braking points and racing lines.

    However it's not a strong argument about mixing different grades and it could partially be a reason why the rating system is actually splitting the driver classes rather than protecting good drivers in a poor drivers' environment.

    A good driver can't and shouldn't drive within 20 meters of a rookie that has no idea about braking points or even worse a wrecker.
    That's no good way to make a safe pass.

    On the contrary clever drivers need to be overly cautious and give them more room than they would do to an A-B driver (which is in full control of his car) until it is safe enough to pass.

    In case of contact, an A-B driver shouldn't get more contact points for being cautious and staying farther away before attempting a safe pass on a rookie or a wrecker.

    I think the number of safe contact-less passes at the end of a race is a better indicator of the skill and fairness of a driver when mixing grades.

    If we want to "save" the concept of attack range when mixing grades, the only thing I can think of, is making the attack range 10m longer per each positive difference step in grades between attacking and defending.

    If A-B attacks A-B the attack range could be 20m.
    If A-B attacks C the attack range could be 30m.
    If A-B attacks D the attack range could be 40m.
    If A-B attacks W the attack range could be 50m which is probably very smart as wreckers should be handled with caution.

    Similar could be done for C on C, C on D and so on.

    In any case I think the attack range is a risky concept and that it can be abused.
    One driver could drive the full length of a race within 20m of another driver and then push out the opponent at the last turn.
    And wreckers could find a way to exploit this to improve their grade (while they wouldn't be able to do the same with the count of overtakes at the end of the race which requires finishing the race and cleanly passing others).

    I think counting contactless passes over "polished-normalized" end race results (i.e. cleaned of people that didn't drive to the end, disconnected or respawned) as suggested in my previous post is better because in the end it doesn't really matter how close you drive to others but how cleanly you can pass them.

    In my opinion a good deal of clean overtakes is the only meaningful proof that can reduce the seriousness of making an accidental contact to other cars and who knows, it might also encourage faster C fallen drivers to join ABCN servers and cleanly storm through the field to restore their grade (maybe teaching something to the others).

    And for the fast drivers that like to start in front and run solo, it won't be that hard to get a clean solitary race to restore their rank (cause that's also a sign of intelligence and skill and shouldn't get penalized just because they are not driving within 20m of somebody else).
     
    Last edited: Aug 6, 2016
  19. V8_KB

    V8_KB Alien

    Oh snap indeed, time flies. I remember at the beginning I was just watching from the outside how it will proceed and then jumped in and never left.

    I like stats, but I also agree with your approach regarding "know it all". It would be like cheating if we knew exactly how MR works in all situations and how to "boost" your grade in the most efficient way, and would heavily affect how (some/most) people drive.

    Exactly, but you seem to be missing the point. Which is that a good driver shouldn't drive in "poor drivers environment" as you called it, he should drive with like-minded people. But because most of the huge 12000 AB drivers group also miss that point, it's hard to be in the "good drivers environment". I'll stop here as I'd have to rant about the same thing for 95836th time.

    Lets imagine that:
    1. I'm racing with you and we have an epic battle from start to finish, we never closed up on anyone else and nobody closed up on us either. You are in front all the time, we have lots of 2 wide action through multiple corners but I never actually pass you. We finish in the same positions as we started, so because of the fact that there was no pass it means it doesn't matter? All that great door to door driving with absolutely zero contacts for nothing?
    2. Same story but I pass you near the end of the race. I think it's unfair if I'm being rewarded for clean pass, and the driving itself "doesn't matter". So we were both doing our best throughout the whole race distance but only I'm getting something out of it.
     
    Ace Pumpkin, Andy-R and Minolin like this.
  20. Minolin

    Minolin Staff Member KS Dev Team

    Don't think too complicated here. The attack range doesn't tell anything about what we expect from a driver. It's just an indication if you are more or less driving alone or not. 20m is really enough, even when driving around rookies.
    Go a step backwards: We have one big safety rating system out there, that simply counts contacts/distance (or corners).

    I didn't like this at all, because:

    - A good driver that is involved in fights often should be allowed to have more contacts
    - A wrecker should not

    In an extreme case I could just count contacts/attack range and change the grade requirements accordingly. Actually this is what you really want to have if there are no newbies or rookies around, just W and A drivers:
    A wrecker that offensively waits for other drivers may have a few meters of attack distance until the crash, while two fighting A drivers might have kilometers even if they tough several times a lap. You have a ratio of 100, maybe even 1000 between both situations - that's the point you want to use to distinct both.

    But in the end it's all about statistics. MR has proven that if you look at the data this way the good drivers will have significant attack and combat ranges, while the not-so-good drivers don't have.
    An A or B grade will be given to anybody with good attack/combat range results and reasonable contact points.

    What you are suggesting is too complicated imho - you are missing many special cases for a more complicated model that won't really improve the outcome.

    I need to repeat: The whole MR model does respect you have accidents. It's build around the fact that even an A driver will cause an accident, and for sure be victim. You don't need to find ways to protect good drivers. There is an A-Driver with 38 contacts and one mass accident. B-drivers can have 80 CPs. 80 accidents in 1000km.
     
    Berniyh and Andy-R like this.
Similar Threads
Forum Title Date
ACC Multiplayer Entrylist-Bug (Driverswap) after a driver got crash/disconnect Aug 8, 2020
Troubleshooting - Workarounds Not able to play with 3-4 cars or more (drivers fault) Jan 25, 2014

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice