Aim-bot-recognizer
13 posts
• Page 1 of 1
Aim-bot-recognizer
Hey all,
I was wondering about something. Would it be possible to build an aim-bot-recognizer? The idea is to determine the chance of that a player uses an aim-bot.
Suppose that a relative high second time derivative of the position of the cross hair (i.e. acceleration of the cross hair) in combination with statistics of the accuracy and the movements of both players says something about the chance that a player is using an aim-bot. For example, five or four head shots on a very fast moving opponent does not have an high probability, so that could be suspicious.
If my theory works out, it may could be used for building a tool which alerts admins, if a (new) player appears to have a pretty high chance for using an aim-bot. Of course it should not be used to determine if a person uses an aim-bot. The suspicion is just based on statistics which means that there is a chance that the tool can be wrong if a player has some lucky shots. It becomes even harder if a player is very good which means that the chance on "lucky" shots increases.
I would like to do some experiments, but I have a lack of data. Is it possible to retrieve data from the server(s) on the position of players and several other variables of human players and of bot players?
I am used to use MATLAB which is pretty handy for analyzing and experimenting. (MATLAB is a program for doing everything with data: modifying, analyzing and much more. It is pretty much used by engineering (studies))
I am very curious about your reactions.
Yours sincerely,
Egglaf
I was wondering about something. Would it be possible to build an aim-bot-recognizer? The idea is to determine the chance of that a player uses an aim-bot.
Suppose that a relative high second time derivative of the position of the cross hair (i.e. acceleration of the cross hair) in combination with statistics of the accuracy and the movements of both players says something about the chance that a player is using an aim-bot. For example, five or four head shots on a very fast moving opponent does not have an high probability, so that could be suspicious.
If my theory works out, it may could be used for building a tool which alerts admins, if a (new) player appears to have a pretty high chance for using an aim-bot. Of course it should not be used to determine if a person uses an aim-bot. The suspicion is just based on statistics which means that there is a chance that the tool can be wrong if a player has some lucky shots. It becomes even harder if a player is very good which means that the chance on "lucky" shots increases.
I would like to do some experiments, but I have a lack of data. Is it possible to retrieve data from the server(s) on the position of players and several other variables of human players and of bot players?
I am used to use MATLAB which is pretty handy for analyzing and experimenting. (MATLAB is a program for doing everything with data: modifying, analyzing and much more. It is pretty much used by engineering (studies))
I am very curious about your reactions.
Yours sincerely,
Egglaf
- PolygonWindow
-
- Posts: 373
- Joined: Sun Jun 29, 2008 0:00
Should'nt be that easy since u can set up different hit-regions in "modern" hacks + theres a humanized aim(fails included, aimkey..) in most hacks today which make it hard to detect. Nice idea but imo build to fail, since sometime i have a acc like 40 or more and sometimes 25 oO. So its for me at least depending on a good day. So a soft only looking for stats wont give any real proofs mostly...
I may be wrong there..
I may be wrong there..
I guess you haven't realise to receive such an answer but such a tool is not reliable even it might be possible to develop. Nowadays even punkbuster doesn't recognize any kind of humanized bots as I was experiencing similar at my home server.
![Image](http://img706.imageshack.us/img706/4048/shad.gif)
Once upon a time there was a Drum and a Bass, on first date they kissed, second date they planned to marry, then they married and became Drum'n'Bass. On they honey moon they had sex and after 9 months DUBSTEP WAS BORN !
YO,
Thanks for your reactions guys. I was indeed expecting that it would be really though to spot a modern bot, but there are some signatures that which are unique for a bot, humanized or not. For example, the shooting of the humanized bot in the first 20 seconds, in the movie that Diamond posted, through the grass and into a bush. While you can hardly spot your enemies there.
It wouldn't be a soft spot for statistics PolygonWindow but a comprehensive look on it, of you could set up something to get a coefficient for the ability to see a enemy, and combine that with the aim, and some properties for the movements of the cross hair I think that it is still possible to produce a number which represents the chance that someone uses an aim-bot.
Thanks for your reactions guys. I was indeed expecting that it would be really though to spot a modern bot, but there are some signatures that which are unique for a bot, humanized or not. For example, the shooting of the humanized bot in the first 20 seconds, in the movie that Diamond posted, through the grass and into a bush. While you can hardly spot your enemies there.
It wouldn't be a soft spot for statistics PolygonWindow but a comprehensive look on it, of you could set up something to get a coefficient for the ability to see a enemy, and combine that with the aim, and some properties for the movements of the cross hair I think that it is still possible to produce a number which represents the chance that someone uses an aim-bot.
- Justkidding
-
- Posts: 151
- Joined: Wed Jun 27, 2007 0:00
Well, the most common aimbotters are easy to spot. After-all, 95% are noobs that use downloaded hacks too show their fake skills. However I don't think you can spot pro-hackers easily. Most pro-hackers are skilled in programming skills that can be used to make "intelligent" aimbots, or any other hack. For example an aimbot that's especially designed for a particular map. It is programmed in a way, that when a human can't see the target. The bot can neither, well theoretically it can ofcource, but the bot is not allowed to shoot it. Lets say, a player is in the grass, it's hard to spot that player, than the bot will not respond on the player, unless the player starts shooting himself. Or, when the pro-hacker behind the screen decides that the bot is allowed to shoot whatever it can see.
And these bots are not only aiming on the players their head. Ofcourse not, that would be stupid as hell. It's calculating the HP the player got left, and react on that. So lets say, after 2 body-shots the enemy is still alive and the pro-hacker got 25HP left from the, lets say 125. The bot will automatically aim on the enemies head and still shoot the player to dead first.
Lets say two enemies are shooting at the pro-hacker, the aimbot has to make a choice. "If I kill these players, do I look like a bot?" - "Yes" - "alright, the bot rejects to shoot". Well, the bot wont communicate like that, but hell, you know what I mean. If the bot make a decision that it's clear to shoot, and not look-a-like aimbot, it fires.
But aimbot is extremely fast, you can see a mouse movement difference between a player, or a bot?
Well these pro-hackers aren't stupid, and thought about this too, the answer is simple: They make the bot aim slower and sometimes little bit faster, therefor it looks like a normal player.
But if people keep making kills without failure hits?
The bot, as far I remember. Keeps that to in "mind" and it will also make shots that don't hit the player.
How the [...] did you know all this?
I saw it at a lan party where "hacking" was the topic of the day, and discussions about making bots better, harder to spot.
Why can't I find this shit on google?
Because hackers don't like to share things, when a hacker shares a program, the anti-hackers "community" can abuse this, and make a prevention against it.
The same counts for "Why can't I find programs to hack passwords?". This is shared in a close community of people, they are careful sharing this kind of programs.
Is it possible to make an aim-bot-recognizer
Yes, but some programs are just to "intelligent" to reconize with another program. Only when the programmer of the aim-bot-recognizer can recognize the pattern of the "intelligent" bot. It can be build, but it's almost impossible without having the program in your property.
Is an aim-bot-recognizer good?
Well, the idea may seems great. Really, catching cheaters while sitting in your chair smoking a joint, doing nothing... However there are to many problems with this idea. What if a player plays amazing, hes skilled as hell. And the bot spot him and use the programmed statistics to calculate if hes a cheater or not. The outcome: "Yes, he is" but this player isn't, his just playing great... Than you're banning people when they are doing nothing. Also what if, you create a program with a bug in it? it bans people that are innocent. Their reputation is messed up, they couldn't play for days/weeks/never because of it. Seems harsh to me. The great idea may become a great failure. I don't know, the theory is good. But I think it stays there.
Are you allowed to talk about it?
Sure I am, I'm just not "allowed" to give information how it's created, and the deep inside protection, and more options on the bot...
I can say whatever I want, but it's not always appreciated.
Do you have this bot?
Nope,
their policy says I am only allowed to have any of their programs until I master the language.
Why do you tell all this?
I'm bored.![Laughing :lol:](./images/smilies/icon_lol.gif)
And these bots are not only aiming on the players their head. Ofcourse not, that would be stupid as hell. It's calculating the HP the player got left, and react on that. So lets say, after 2 body-shots the enemy is still alive and the pro-hacker got 25HP left from the, lets say 125. The bot will automatically aim on the enemies head and still shoot the player to dead first.
Lets say two enemies are shooting at the pro-hacker, the aimbot has to make a choice. "If I kill these players, do I look like a bot?" - "Yes" - "alright, the bot rejects to shoot". Well, the bot wont communicate like that, but hell, you know what I mean. If the bot make a decision that it's clear to shoot, and not look-a-like aimbot, it fires.
But aimbot is extremely fast, you can see a mouse movement difference between a player, or a bot?
Well these pro-hackers aren't stupid, and thought about this too, the answer is simple: They make the bot aim slower and sometimes little bit faster, therefor it looks like a normal player.
But if people keep making kills without failure hits?
The bot, as far I remember. Keeps that to in "mind" and it will also make shots that don't hit the player.
How the [...] did you know all this?
I saw it at a lan party where "hacking" was the topic of the day, and discussions about making bots better, harder to spot.
Why can't I find this shit on google?
Because hackers don't like to share things, when a hacker shares a program, the anti-hackers "community" can abuse this, and make a prevention against it.
The same counts for "Why can't I find programs to hack passwords?". This is shared in a close community of people, they are careful sharing this kind of programs.
Is it possible to make an aim-bot-recognizer
Yes, but some programs are just to "intelligent" to reconize with another program. Only when the programmer of the aim-bot-recognizer can recognize the pattern of the "intelligent" bot. It can be build, but it's almost impossible without having the program in your property.
Is an aim-bot-recognizer good?
Well, the idea may seems great. Really, catching cheaters while sitting in your chair smoking a joint, doing nothing... However there are to many problems with this idea. What if a player plays amazing, hes skilled as hell. And the bot spot him and use the programmed statistics to calculate if hes a cheater or not. The outcome: "Yes, he is" but this player isn't, his just playing great... Than you're banning people when they are doing nothing. Also what if, you create a program with a bug in it? it bans people that are innocent. Their reputation is messed up, they couldn't play for days/weeks/never because of it. Seems harsh to me. The great idea may become a great failure. I don't know, the theory is good. But I think it stays there.
Are you allowed to talk about it?
Sure I am, I'm just not "allowed" to give information how it's created, and the deep inside protection, and more options on the bot...
I can say whatever I want, but it's not always appreciated.
Do you have this bot?
Nope,
![Sad :(](./images/smilies/icon_sad.gif)
Why do you tell all this?
I'm bored.
![Laughing :lol:](./images/smilies/icon_lol.gif)
Dude ive never played against players with such "smart" bots but theyr useless those bots would only work "good" on servers filled with noob i mean if someone starts shooting me body shots and i shoot head ill win no matter how fast he moves his mouse from body to head...
U can shoot someone 3 body shots+2 head but u dont move good ull just die from fast triple to the head..
my 2 cents
U can shoot someone 3 body shots+2 head but u dont move good ull just die from fast triple to the head..
![Razz :P](./images/smilies/icon_razznew.gif)
my 2 cents
Fefe wrote:Dude ive never played against players with such "smart" bots but theyr useless those bots would only work "good" on servers filled with noob i mean if someone starts shooting me body shots and i shoot head ill win no matter how fast he moves his mouse from body to head...
U can shoot someone 3 body shots+2 head but u dont move good ull just die from fast triple to the head..
my 2 cents
You haven't played against them, because you don't play at that league. The guys who would use these bots are known to play at high ranks in the competition scene. They are buying their bots (like $10 i think) and are pretty serious about it, and some of these bot users aren't that noob without it. In ET even private bots have been made with unique characteristics, making detection impossible.
I don't think many would want to waste money on a bot just for public servers with no cash prize matches.
Human Aimbots are not that complex, still they are undetectable if they are set right.
Tho, the skillboost is mostly minimal.. imo.
- but yeah, there are guys in the proscene which set it to a really high value and are suspicious every game they are playing and most of them getting busted soon or later.
- the thing which really gives a great skillboost in the pro scence is the damage feedback removal.
(removes the kickback when being shot)
First: It's undetectable, as you can't see if someone uses it with human 'eyes'
Second: It gives a great skillboost.
Such detection methods would be used if they wouldn't be difficult to add / possibility.. whatever.
Chaplja has great experiance because being a pretty well coder, he knows the ways howto successfully block hacks imo.
You can see it already as slac provides a better service than pb in my eyes.
Tho, the skillboost is mostly minimal.. imo.
- but yeah, there are guys in the proscene which set it to a really high value and are suspicious every game they are playing and most of them getting busted soon or later.
- the thing which really gives a great skillboost in the pro scence is the damage feedback removal.
(removes the kickback when being shot)
First: It's undetectable, as you can't see if someone uses it with human 'eyes'
Second: It gives a great skillboost.
Such detection methods would be used if they wouldn't be difficult to add / possibility.. whatever.
Chaplja has great experiance because being a pretty well coder, he knows the ways howto successfully block hacks imo.
You can see it already as slac provides a better service than pb in my eyes.
eLemenT wrote:Chaplja has great experiance because being a pretty well coder, he knows the ways howto successfully block hacks imo.
slac sux tbh.
My clanmates uncle hacked slac with 2hours and then he contacted chaplja to check his slac account if there is any cheat or not and chaplja said he is clean.
He is programmer and working somewhere on a high level security (i hope i remember right) in his country and it took only 2 hours to put wallhack on slac and he isnt still busted. That wallhack worked like r_shownormals. My clanmate (i wont say name) was next to him and saw he played with wallhack & slac
13 posts
• Page 1 of 1
Who is online
Users browsing this forum: No registered users and 15 guests