What features do you want to see in an Anti-Cheat?

Discussion in 'Spigot Discussion' started by funkemunky, Mar 12, 2018.

  1. Hello,

    I’m an anticheat developer new to this website in the premium resource market, and I’m wondering what features you people would like to see in an anticheat (especially a premium one)? I have been selling my anticheat on another website and found much success. However, I see the community is very different here. If you can give me your input(s) that would be great!

    Also if you’d like to put examples of the style of anticheat and/or the link to an anticheat you like/use, you can put that in this thread.

    - funkemunky
     
    • Like Like x 1
  2. To be honest, as an admin, all I want is that the features a plugin provide help me figure out if someone's just giving me bs on their in-game behavior or if they're indeed on a non-vanilla client, using certain mods, and have done things that triggered the anti cheat. And if so, if there's any logs of this that are somewhat easy to read so you can use it as convincing evidence against the player.

    The ability to get some data on the user for the last few days after the fact, would be nice. For example .. if the AC can track the person and they have multiple warnings and violations, that are way above the average, it seems logical that it starts to log more in-depth and keeps track of this so a /ac report <player> -time:1w, would actually help us analyze the situation better.

    A feature of 'i dont want them to fly, so fix that' is in every anti cheat.
     
    • Agree Agree x 1
  3. Not a bad idea.
     
  4. better custom enchant support & better conditions if ping/tps is high don't tp. ect
     
  5. +1

    --

    I would like it to have the less false positives as possible, and a good configuration system.
     
  6. As far as player motion, it's quite possible with an NPC. I've been planning to push an update on my end doing just that for quite some time now. Each notification providing an id that staff can use to replay their action(s) as solid proof of cheating compared to just logs which can occasionally be inaccurate or due to some incompatibility that wasn't accounted for.

    If an anti-cheat's checks don't check for player and server latancy before pushing a notification, they're doing something wrong to begin with.

    To my own surprise, most people actually perfer notifications sent sooner rather than later. Obviously if an anti-cheat is consistency spamming a false positive, it would be something to report to the developer. Accuracy is a a key aspect but to what extent should a notification be delayed before being sent?

    I definitely don't mean to hijack this thread so I'll contribute as well. Most people do look for solid movement and combat cheats over minor cheats like derp so those should be higher priorities. A key aspect is limiting CPU usage as it seems to be the falling point of most anti-cheat plugins so I've heard. Updating is a key topic, not abandoning the plugin after a month of earning a quick buck before moving on. I'll of which I'm sure that you're familiar with.

    Personally, and although I may not use it constantly, I wouldn't mind seeing a GUI with player heads that display the main or current violations that a player has when hovering over it (I believe aac had something like this at some point or has it currently?). Perhaps tying into @mrfloris's idea is clicking the head to display paper or books that hold their violation history. If needed, the GUI would include arrows to the next page or page before at the bottom, in addition to the the main menu icon. And while on the topic of GUIs, which should be icing on the cake rather than the main focus, in-game config editting is always a plus. Kind of less useful for an anti-cheat but again, it should be a low priority.
     
    • Agree Agree x 1
  7. I don’t not believe in the use of NPC. Why? Because the client I have has a one line of code antibot that blocks all NPC based checks and is impossible to patch unless you use a real live player. Algorithimic and heuristic based checks are the way to go.
     
    • Agree Agree x 1
  8. I didn't even mention ncp directly, much less anything on kill aura, why bring it up? If you have an ideal kill aura patch, use it. There are several ways to pick up kill aura as depicted by this endless thread; some are better than others.
     
  9. I said NPC, not NCP. Npc as in bots.
     
  10. Oh, just saw that, my bad. I do agree with not using npcs for kill aura to some extent as they could cause lag, especially on larger servers, but they tend to be useful in picking up poorly set up kill auras that aren't set ignore different conditions like invisibility, mobs, etc. Again, if you have something that works well, take full advantage of it.

    To clarify my first post, I didn't mean an npc for detecting kill aura but an npc to recreate/replay the actions that a suspected player does. Which would include player movement, actions like eatting or swinging their sword, dropping items, etc. This could be something to implement if you want to do so after focusing on the main features of your anti-cheat. I'm not quite sure how intensive it may be on a server though, depending on the player count I would assume as it would be recording each player's every action.
     
  11. I think it's just bringing something new to the table, similar to what @mrfloris said, so many anti-cheats here all do the same thing of 'ha cheat xx detected, lets cancel them doing it' gets boring and what everyone seems to do. I think just having an anti cheat that keeps track of what your real players are performing at (because if you have a custom plugin, normal anti cheats would block it, similar to machine learning) compared to a 'hacker'.

    Would be cool to even have it flag up potential cheaters, you define a time (eg 1 week) and it'll show whether they were performing at average or of a cheating level etc.

    tl:dr an anti cheat that isn't gonna be the same as every single one, just report to staff with lots of interesting information than just block it with hardcoded values.
     
    • Agree Agree x 1
  12. That requires human interaction and is irrelevant to an anticheat. That’s not an anticheating solution that’s worth buying. That just takes out the factor of players having to record the cheaters after cheaters ruin their experience. What a stupid waste of money and resources.

    NPC’s is general aren’t a good idea. If someone is using a client from 2 years ago and doesn’t have an antibot, it’s likely a Killaura check from a leaked anticheat you can skid, if you really can’t make one, detect it. Bots arent worth the time and effort, as client devs can patch them in one line of code. Not to mention, any Killaura not properly setup is easier to detect without a bot than with. Anticheats whose detections rely on bots tend to be the worse ones.
     
    • Agree Agree x 1
  13. This is actually quite inefficient and a bad idea. Anticheats aren’t supposed to be fun, except for me at least. I love making them. But still, the point of an anticheat is to let the humans focus on other things. Now I have to agree, that would be a nice touch. However, in any real world use that isn’t acceptable.
     
  14. I still don't think you are getting their point. They want a replay system that spawns a fake player that performs all actions of another previously recorded player thus allowing you to "watch" potential cheater gameplay even when the player is not online. Like let's say there is no mod around. But in the evening a mod is joining, gets a notification that potential cheater gameplay has been recorded and he is asked to review it. That's what they want.

    And I would only partly agree to your anti bot thing. The strategy should be to use a bot only if the anti cheat already suspects something. And the bot could also be adapted to show almost human behaviour and thus breaking most of the anti bot cheats. I mean if only 1% of your player base is ever confronted with such a bot because they are being watched closely by the anti cheat it might help you to determine who is really cheating.
     
    • Like Like x 1
    • Like Like x 1
  15. konsolas

    Supporter

    I've actually got a plugin in the works which is designed to record gameplay retroactively, which could be used by anticheats to save a recording of suspected cheaters.
     
    • Informative Informative x 2
    • Creative Creative x 1
  16. That I didn’t know. Link it, please.

    No you don’t understand. There is no way to break a proper one liner antibot unless you have actual real Mojang registered players. Even then, antibots can just be made to block players with the specific uuids. I don’t believe in having staff review unless you have very experimental checks for it. If that is the case, I totally agree. But for main functionality, no.
     
  17. There are 30m Minecraft accounts out there. Of course you can use any uuid of those to make your player look real. I don't really see where you are coming from.
     
  18. still wouldn’t work. The bots would never be under the Player object not returned as null. I should have worded it differently. It isn’t a Mojang registered login session.
     
  19. I've been thinking of making mine modular. I only support legacy versions, though. If you can make Fiona able to load custom-built checks, that would be awesome for the community. If someone is not satisfied with a small check or two, they can make their own and integrate it into your anticheat.
     

Share This Page