Bot Killers

zhult

I would like to propose a crazy idea. As in all games we know that ashes will have bots, but thanks to the pks system we will be able to kill them, surely on my server I will create a guild called bot killer with my alter, and I will dedicate myself only to searching and kill bots. Do you think that at the community level a guild would be created on each server with this name and that if they only kill bot pks they won't be persecuted? On a personal level I would not do it, how do you see this? ((sorry for my english)

Kilion

I think Intrepid will forcefully deactivate any account that shows patterns of bot activity, which will be done with more accuracy and speed than players could, as Intrepid can just look at the activity logs of an account.

And since Bots would need to be flagged (and therefore recognized by the system) as "Bots" for players to kill them them without gaining corruption I don't think that Bot hunting will become a thing.

daveywavey

Seems like an obvious way to hide your PK of actual players. "Nope, I totally got this from killing bots. Don't worry about me......"

Darkbolt

Ahah yea killing bots will be a funny game, but i hope the AoC economy system and game mecanics will make useless the use of bots.

Strevi

@Azherae said the ideal activity bots will do in AoC is to run caravans. So everybody might hunt them, without care if they are bots or not and without fear of gaining corruption.

But if bots will also gather resources, then I do not know if those resources will worth a lot. And bots will still retain some resources after being killed, making more efficient gathering the resources before them.

Azherae

Strevi wrote: »

@Azherae said the ideal activity bots will do in AoC is to run caravans. So everybody might hunt them, without care if they are bots or not and without fear of gaining corruption.

But if bots will also gather resources, then I do not know if those resources will worth a lot. And bots will still retain some resources after being killed, making more efficient gathering the resources before them.

Specifically for those unfamiliar (since I could easily be wrong about the caravans thing as a whole) this is the problem with botting, and why I tend to refer to it as botting instead of 'having bots'.

Here's a brief explanation of what a botting system normally does in a simplistic game, shooter or MMO, from a programmer:

1. Determine the functions and UI of a game, relative to what a human needs to see and use to understand, mostly by the color ranges onscreen.
2. Find an activity where the screen changes in only one small area, or a human would only need to focus on a small area of the screen (caravans can fit this, gathering less so, with VFX all the way up, PvP might)
3. Implement pixel or pixel-area based detection for that area (this can use both ranges and histogram algorithms, just assume that if a human can see it WELL, this doesn't require AI, only 'some similar small images to compare to'
4. Build a function that can 'watch the screen' in the area you want (this can be costly, since it's constantly 'taking tiny screenshots', but no moreso than anything that lets you stream content
5. Build a function that will gather information on your mouse movement and learn to emulate it with some randomness and combined linear transforms (this step is unnecessary if the game isn't likely to detect instant mouse movement)
6. Run the program while playing, even while playing the game yourself. When the condition is met, the bot 'takes over' the mouse movement for a moment and may be able to do other things such as fire a weapon or use an ability (aimbots).
7. If the game requires a 'fuller' bot, add a function similar to the mouse movement one for other movement too. Point and click games, these are the same function. Probably add a toggle key within game for quickly activating or deactivating the whole framework
8. Gather the information on the pixels/ranges for the thing to be botted, i.e. do the thing once, then customize each new loop function for whatever 'response to screen stimulus' you want the bot to take.

Things I therefore expect to be botted are:
Processing Tasks - Minigames usually require no camera movement and have clearly defined screen positions for where their visual indicators appear, it doesn't actually matter too much how many of these there are.
Fishing - Easy same for most games, if the prepwork isn't the hard part, and even then.
Lumbering specifically - Because you can get the location of tall singulars onscreen using parallax and sky lighting. Beware those 'bots' that only seem to gather at a certain time of day in game.
Healing - Most games, healing is not movement related enough, it's 'detect a change in a health bar, click it, click a skill'. This is the bare minimum.
Caravans - A combination very similar to "Lumbering + Healing". A caravan is a factor heavily under the player's understanding and control. Color, orientation toward the 'bot', expected path. A caravan defender who is a healer in a party can 'use the detection to stay close to the caravan', use the movement functions to keep up with it, and use alternate detection to watch the HP of themselves and other players and heal more-or-less correctly.

Can you think of anything else that is likely to be botted, given the explanation? I'm interested to think about it, and it might help us to focus in a way that will 'figure out what bots to kill', especially in certain weather conditions before the less skilled botting implementations can easily keep up with the on-screen context ranges.

But beware that the problem is often that a person who is running a bot or bot-adjacent script isn't even actually afk, they're just doing something else while the 'bot' handles the 'difficult' or 'tedious' parts.

And that's why I usually say that people who think that they personally will be able to easily detect bots will only catch the weakest. They're beating up on the equivalent of someone's Roomba when the better bots are 'cyborgs'.

Note: This explanation is for the type of 'bot' that does not read ingame memory at all, it is less intense than a streaming program, and can work similarly, it is not undetectable, but that method of detection is not available, and a custom implementation is not difficult enough to make them hard to churn out.

Here's a minimal-effort-search reddit thread for anyone whose curiosity is sparked now.