Announcement

Collapse
No announcement yet.

Epic Games Announces Easy Anti-Cheat For Linux - Including Wine/Proton

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Ironmask
    replied
    Originally posted by mrazster View Post

    Oh great, more Tinfoil Hats in the beautiful world of Linux, yay !
    1989 Tiananmen Square Massacre 六四事件 八九民运 茉莉花革命动态网自由门 天安門 天安门 法輪功 李洪志 Free Tibet 六四天安門事件 The Tiananmen Square protests of 1989 天安門大屠殺 The Tiananmen Square Massacre 反右派鬥爭 The Anti-Rightist Struggle 大躍進政策 The Great Leap Forward 文化大革命 The Great Proletarian Cultural Revolution 人權 Human Rights 民運 Democratization 自由 Freedom 獨立 Independence 多黨制 Multi-party system 台灣 臺灣 Taiwan Formosa 中華民國 Republic of China 西藏 土伯特 唐古特 Tibet 達賴喇嘛 Dalai Lama 法輪功 Falun Dafa 新疆維吾爾自治區 The Xinjiang Uyghur Autonomous Region 諾貝爾和平獎 Nobel Peace Prize 劉暁波 Liu Xiaobo 民主 言論 思想 反共 反革命 抗議 運動 騷亂 暴亂 騷擾 擾亂 抗暴 平反 維權 示威游行 李洪志 法輪大法 大法弟子 強制斷種 強制堕胎 民族淨化 人體實驗 肅清 胡耀邦 趙紫陽 魏京生 王丹 還政於民 和平演變 激流中國 北京之春 大紀元時報 九評論共産黨 獨裁 專制 壓制 統一 監視 鎮壓 迫害 侵略 掠奪 破壞 拷問 屠殺 活摘器官 誘拐 買賣人口 遊進 走私 毒品 賣淫 春畫 賭博 六合彩 天安門 天安门 法輪功 李洪志 Free Tibet 劉曉波动态网自由门

    Leave a comment:


  • Awesomeness
    replied
    Originally posted by jaxa View Post
    Don't like it? Don't play it. I only use GOG and Steam.
    This is about spyware that is included in plenty of GOG and Steam games. It's not about Epic Game Store. 🤦‍♂️

    Originally posted by mrazster View Post
    Oh great, more Tinfoil Hats in the beautiful world of Linux, yay !
    "EasyAntiCheat [...] monitors your PC while you’re playing" https://www.howtogeek.com/358756/wha...n-my-computer/

    How is "monitoring the PC" not spyware?

    And Epic Games is half owned by Tencent. Look it up if you don't believe it.

    Leave a comment:


  • oiaohm
    replied
    Originally posted by artivision View Post
    Please don't use as citations random articles that you read casually on the net. If you read correctly you will see something like: Then next version will be the big one with full auto aim and superman stuff, bla bla bla. Don't confuse driving simulation with fast paced FPS games, driving simulation is easy and it's developed for years, can easily be used on games. It happens to be a semi-expert on that department and i'm telling you that by the time that an AI like that exist, we will be at the age of full dive so that AI will be deprecated. Just buy and confirm, instead of auto aim you will get auto fail. Also those people never claimed that you say they did, they only claimed that you can write you own exploits training this technology (after how many years and with how much knowledge tho?).

    Its about time you stop the crap that is in the future. We have banned players who have confirmed they have been using systems exactly like I have described in different FPS. The claim about auto fail is not true I wish it was.


    Here they mention the processing side to the system to make it choices is 10ms. Remember human at best is 80ms. This is not using insanely huge GPU either can get down to 10ms this is why losing 40-60ms to a decent capture device is not a problem. That still puts automated system ahead human.

    artivision problem here I deal with people who do banning and investigate this stuff. Those write ups are behind where the high end game cheaters are. The high end competition cheaters who are cheating to earn money don't sell their cheats to other people and in fact pay their own developers from their winning to keep their cheats competitive. There is a plea bargin process at the moment I don't really agree with where if the cheaters who have won money hand over the tech the developers they get to keep their winnings and only get a 12 months ban. One of the most advances was the guy caught being in two places at once.

    The question now is how many years from the high end cheaters tech from the players cheating for money to end up as your 50 dollar cheats. Yes that arstechnia is a lower end cheat developer following in the foot steps of the higher end cheat developers.

    Leave a comment:


  • artivision
    replied
    Originally posted by oiaohm View Post

    This is still going down the wrong arguements.

    The cheat maker brags here yes.



    Device maker mentions the interrogation. The point to be aware that OpenCV does not just have heavy neural net identification.

    https://csusm-dspace.calstate.edu/bi...Spring2019.pdf

    Depend on the game depends on how much GPU you need or if you need a GPU at all. Yes that is 2019 paper. Thing have been evolving. The screen capture rate required for a game to be played by computer vision processing with a movement logic AI turns out not to be that high.

    The unnatural nature of computer game engine generated worlds makes it a lot simpler to apply filters that strip away the texturing and reduce the game to basically wire frame model. This stripping the game down to a wireframe mode is why these systems can downscale the graphical input without a issue. Because they reduce the game output appearance anyhow massive and use that for making action choices. You hit boxes are designed for humans to hit them in games. So the automatic game player does not need absolute perfection just better perfection than human.

    Sorry to say AI solutions that run on second computer using screen capture and controller emulation can in fact recognise different and jump when they are not part of the game engine. They can jump faster to the appearance of the object than what a Human can because the processing loop is shorter than the natural human loop.

    Some games are more AI controllerable than others. How well the games rendered output can be converted back into a wire-frame /simplified is a big one. Lot of the top played FPS turn out to be light on processing to simplify the output.

    There is a behavour difference you do see from here as well. Current cheating AI setups will use walls and that for cover but they are not smart to use wall close to the same colour as there skin to hide in front of. Good reason for this that data has been discarded. Automated vs Automated there is no advantage to hiding that way.


    The reality here not all those working on self driving cars are about having huge power computer to-do. There are many competitions these days to make self driving cars as in toy class. These don't use big heavy CPU or GPU for the choice logic. This is why it comes how much can the games rendered world be simplified. If the game engines output can be simplified far enough then you don't need a GPU to play it. Some of your old and new FPS games can be fully auto played with a raspberry pi 4 at speeds a human cannot do.

    AI car cannot recognize all random objects. This is true but it does not need to be able drive the car safely. Something horrible in 2018 a AI was proven to be able to identify more objects with higher accuracy than humans can and faster than human can but this was totally useless for driving a car. This is why its the last 3 years in the last 3 years the way making a self driving car logic has to be designed has changed. No longer need to identify every object what it is. When playing a computer game you not identifying every single object either. That the problem the AI cars on light end CPU/GPU solutions have had to get very good at discarding non important objects and only recognising important objects. This discard of non important objects and simplify what need to be processed turns out to be the same feature auto control cheat at FPS games needs. Some new FPS games you only ID really 8 types of object.

    We basically have a perfect storm. What is being developed for self driving cars in the last 3 years highly matches up to what is required to auto cheat at games. Few groups so far have taken advantage of the overlap.
    Please don't use as citations random articles that you read casually on the net. If you read correctly you will see something like: Then next version will be the big one with full auto aim and superman stuff, bla bla bla. Don't confuse driving simulation with fast paced FPS games, driving simulation is easy and it's developed for years, can easily be used on games. It happens to be a semi-expert on that department and i'm telling you that by the time that an AI like that exist, we will be at the age of full dive so that AI will be deprecated. Just buy and confirm, instead of auto aim you will get auto fail. Also those people never claimed that you say they did, they only claimed that you can write you own exploits training this technology (after how many years and with how much knowledge tho?).
    Last edited by artivision; 27 September 2021, 07:54 PM.

    Leave a comment:


  • oiaohm
    replied
    Originally posted by artivision View Post
    An AI car cannot recognize all the random objects, will crash if you let it alone and for that needs 1000w as well. That is why those you referring are scam, its no possible for an AI to recognize the object and jump except if its part of the game engine.
    This is still going down the wrong arguements.

    The cheat maker brags here yes.

    Computer Vision Integration (Python OpenCV)
    Seamless integration with Python OpenCV, allowing the Titan Two users download, create and share computer vision algorithms that analyzes the game video streaming in order to obtain valuable information for execution of automated actions.
    Device maker mentions the interrogation. The point to be aware that OpenCV does not just have heavy neural net identification.



    Depend on the game depends on how much GPU you need or if you need a GPU at all. Yes that is 2019 paper. Thing have been evolving. The screen capture rate required for a game to be played by computer vision processing with a movement logic AI turns out not to be that high.

    The unnatural nature of computer game engine generated worlds makes it a lot simpler to apply filters that strip away the texturing and reduce the game to basically wire frame model. This stripping the game down to a wireframe mode is why these systems can downscale the graphical input without a issue. Because they reduce the game output appearance anyhow massive and use that for making action choices. You hit boxes are designed for humans to hit them in games. So the automatic game player does not need absolute perfection just better perfection than human.

    Sorry to say AI solutions that run on second computer using screen capture and controller emulation can in fact recognise different and jump when they are not part of the game engine. They can jump faster to the appearance of the object than what a Human can because the processing loop is shorter than the natural human loop.

    Some games are more AI controllerable than others. How well the games rendered output can be converted back into a wire-frame /simplified is a big one. Lot of the top played FPS turn out to be light on processing to simplify the output.

    There is a behavour difference you do see from here as well. Current cheating AI setups will use walls and that for cover but they are not smart to use wall close to the same colour as there skin to hide in front of. Good reason for this that data has been discarded. Automated vs Automated there is no advantage to hiding that way.


    The reality here not all those working on self driving cars are about having huge power computer to-do. There are many competitions these days to make self driving cars as in toy class. These don't use big heavy CPU or GPU for the choice logic. This is why it comes how much can the games rendered world be simplified. If the game engines output can be simplified far enough then you don't need a GPU to play it. Some of your old and new FPS games can be fully auto played with a raspberry pi 4 at speeds a human cannot do.

    AI car cannot recognize all random objects. This is true but it does not need to be able drive the car safely. Something horrible in 2018 a AI was proven to be able to identify more objects with higher accuracy than humans can and faster than human can but this was totally useless for driving a car. This is why its the last 3 years in the last 3 years the way making a self driving car logic has to be designed has changed. No longer need to identify every object what it is. When playing a computer game you not identifying every single object either. That the problem the AI cars on light end CPU/GPU solutions have had to get very good at discarding non important objects and only recognising important objects. This discard of non important objects and simplify what need to be processed turns out to be the same feature auto control cheat at FPS games needs. Some new FPS games you only ID really 8 types of object.

    We basically have a perfect storm. What is being developed for self driving cars in the last 3 years highly matches up to what is required to auto cheat at games. Few groups so far have taken advantage of the overlap.

    Leave a comment:


  • artivision
    replied
    Originally posted by oiaohm View Post

    That the problem they do exist. Only have existed for the past 2-3 years. This form of ai cheating solutions are part based on work for self driving cars and other things. Sorry light based switches what a joke of a reason to win.



    These brain limits are on the presume that you keyboard/mouse/controller and monitor has zero latency so no matter how you improve your computer system in monitor or controllers or keyboards or mouse more main computer you cannot be better than these limits. So it does not matter how much you improve your hardware. It does not matter if you have light based switches.

    The zen hardware and equal is critical as well. Remember you have the computer processing the screen making a choice on action. Then the zen device is macro based so for the next processing loop the zen base hardware can do a stack inputs while the computer processing the video and making the tactical choices goes back to that.

    Code:
    30 frame a second realtime video processing is 34ms of overhead+90ms for video capture device+ 40ms of device faking input
    The 40ms there for the faked input is using hacked up junk. Using something like a zen the input latency from the point the video processing has made it choice of action to the inputs starting to be entered is a 2ms. That setup with zen bit of hardware is under 150. With a better grade of capture device this is under 80ms.

    The problem here the computer capturing the screen processing it making it choice on what todo and having that choice enter the computer its monitoring is faster than you ever can be. Yes the fact input have been offloaded to the zen and equal bits of hardware is another reason why this is so lethal these days.

    The fake input device receiving directions is only slightly slower than your physical keyboard/mice and controllers with absolutely perfect button press. There is a reason why you don't see people typing at 300 words a min all the time. 1ms to physically press key is about the best you do. Yes the world record typing speed is 216 word a min. Problem here is a item like zen can smash that into the ground. You can have the zen device to typing at 3000 words a min. Yes this is have a computer OCR a page with camera and feed the zen macro device directions.

    This is why I call this stuff inhuman. Zen device does not have to fight with the keyboard spring at all so is always able todo perfect keypresses. It can enter data faster than human by a large magin. 2ms delay in changing the zen devices directions from computer controlling it instructions. But the zen device could be doing single set of directions for up 30 mins basically this gives the automated system like a human reflex system just one problem its even faster than a human reflex system.

    Yes I have messed around with a zen like device not for cheating. I was messing around with it for quality control. 40ms instead of the zen 2ms would be using a raspberry pi zero W in device mode(to emulate the input) hooked up to the pc by either bluetooth or wireless. Yes it was 40ms latency when I worked out a automated system could make input into a computer be at 3000 words per min typing. We are talking at least 10 times faster than a human at pressing buttons.
    Code:
    30 frame a second realtime video processing is 34ms of overhead+90ms for video capture device+ 40ms of device faking input
    This line is if you have used all cheap parts. The cheapest video capture you can buy and the cheapest solution you can use to fake input yes this is a 50USD stack of peripherals . The only thing somewhat decent here is the PC and even then its not a top of the line.

    Yes the first people caught cheating and winning against pros were using close to these cheap hardware. The quality of hardware to cheat this way has improved. If a person puts 200USD into the peripherals instead of 50USD that a good quality capture and a good quality fake input device beating human reflex is straight forwards its just about having processing time optimised enough. Yes games allow the screen processing to be optimised to avoid heavy AI processing.
    An AI car cannot recognize all the random objects, will crash if you let it alone and for that needs 1000w as well. That is why those you referring are scam, its no possible for an AI to recognize the object and jump except if its part of the game engine.

    Leave a comment:


  • oiaohm
    replied
    Originally posted by artivision View Post
    That is another thing that doesn't exist. There is no external AI that can control your character via mouse and keyboard input from image analysis. If there was, humanity would be in danger. Bots are modified game AI, that is why they caught. Also any aimbot except the fastest ram one is doomed to fail, that is because there is so much that have to be done in order to hit your opponent, not only targeting correctly, a wrong move and aim doesn't matter. Auto-evasion how? With good players like me, only one trigger is needed to end it. With light based switches we also win there.
    That the problem they do exist. Only have existed for the past 2-3 years. This form of ai cheating solutions are part based on work for self driving cars and other things. Sorry light based switches what a joke of a reason to win.

    Something else here important max human reflex time with brain in loop is 0.15 of a second but the average is 0.2 of a second(yes 150ms/200ms). So in frame per second that ~7 frames per second. Max reflex with brain not in loop is 0.08 of second or roughy 13 frames a second(80ms). This is not this fast.
    These brain limits are on the presume that you keyboard/mouse/controller and monitor has zero latency so no matter how you improve your computer system in monitor or controllers or keyboards or mouse more main computer you cannot be better than these limits. So it does not matter how much you improve your hardware. It does not matter if you have light based switches.

    The zen hardware and equal is critical as well. Remember you have the computer processing the screen making a choice on action. Then the zen device is macro based so for the next processing loop the zen base hardware can do a stack inputs while the computer processing the video and making the tactical choices goes back to that.

    Code:
    30 frame a second realtime video processing is 34ms of overhead+90ms for video capture device+ 40ms of device faking input
    The 40ms there for the faked input is using hacked up junk. Using something like a zen the input latency from the point the video processing has made it choice of action to the inputs starting to be entered is a 2ms. That setup with zen bit of hardware is under 150. With a better grade of capture device this is under 80ms.

    The problem here the computer capturing the screen processing it making it choice on what todo and having that choice enter the computer its monitoring is faster than you ever can be. Yes the fact input have been offloaded to the zen and equal bits of hardware is another reason why this is so lethal these days.

    The fake input device receiving directions is only slightly slower than your physical keyboard/mice and controllers with absolutely perfect button press. There is a reason why you don't see people typing at 300 words a min all the time. 1ms to physically press key is about the best you do. Yes the world record typing speed is 216 word a min. Problem here is a item like zen can smash that into the ground. You can have the zen device to typing at 3000 words a min. Yes this is have a computer OCR a page with camera and feed the zen macro device directions.

    This is why I call this stuff inhuman. Zen device does not have to fight with the keyboard spring at all so is always able todo perfect keypresses. It can enter data faster than human by a large magin. 2ms delay in changing the zen devices directions from computer controlling it instructions. But the zen device could be doing single set of directions for up 30 mins basically this gives the automated system like a human reflex system just one problem its even faster than a human reflex system.

    Yes I have messed around with a zen like device not for cheating. I was messing around with it for quality control. 40ms instead of the zen 2ms would be using a raspberry pi zero W in device mode(to emulate the input) hooked up to the pc by either bluetooth or wireless. Yes it was 40ms latency when I worked out a automated system could make input into a computer be at 3000 words per min typing. We are talking at least 10 times faster than a human at pressing buttons.
    Code:
    30 frame a second realtime video processing is 34ms of overhead+90ms for video capture device+ 40ms of device faking input
    This line is if you have used all cheap parts. The cheapest video capture you can buy and the cheapest solution you can use to fake input yes this is a 50USD stack of peripherals . The only thing somewhat decent here is the PC and even then its not a top of the line.

    Yes the first people caught cheating and winning against pros were using close to these cheap hardware. The quality of hardware to cheat this way has improved. If a person puts 200USD into the peripherals instead of 50USD that a good quality capture and a good quality fake input device beating human reflex is straight forwards its just about having processing time optimised enough. Yes games allow the screen processing to be optimised to avoid heavy AI processing.

    Leave a comment:


  • artivision
    replied
    Originally posted by oiaohm View Post

    The sad reality here people using the cheat system I described have have fact won online tournaments. So beating pro players is not a question its been done. Yes this includes having deal with real world lag and accuracy loss.

    There is a major differences between the ram based aimlock and these new second computer solutions.
    1) more processing power.
    2) hardware downscaling(this is a big thing)

    Yes that hardware down-scaling from 1080/4k to 480p and 360p means the processing load to find targets is lighter. Turns out AI target identification need less pixels than human for a game. So you could say this is DLSS in reverse that helps cheaters out. Where you historically would only loss 40% time now comes against these newer cheat methods only able to win 1% of the time if at all if you don't have server side anti-cheat looking for inhuman actions so forcing the cheat to obfuscate their actions so leveling the playing field.

    Remember the old ram based aimlock was not moving the player to avoid being targeted as well. So these newer systems are not just locking on to you to shot you but are also attempting to move the player char they are controlling to avoid you getting getting a target hit on them as well. These are on a totally different level to your old school ram based aimlock. Remember I did mention that one of these was caught using a system like what I described was because a person was attending a real world tournament while winning a online one at the exact same time. So this is serous everyone in the online tournament was facing a total AI controlled char and losing.

    At this stage the those setting up a decanted computer to cheat with are still quite rare. But there numbers are increasing. So anti-cheat solutions are having to evolve to deal with this problem. The fact that tactics are unable to over come the amount of advantage these systems have is the big problem. Yes the old ram based aimlock was not far enough in skill above you that tactics could not beat it and good tactics could beat it quite solidly. Now this generation with more processing power where the item doing the fake inputs can also be automatically doing counter tactics as well this is what results in human players totally screwed if anti-cheat system does not catch them.

    Reality here there is a level past autoaim. Yes auto-aim and auto-evasion and auto-tactical movement though the maps. This has to be the most boring way to play the game because you are no longer playing the game but letting the computer do it for you.
    That is another thing that doesn't exist. There is no external AI that can control your character via mouse and keyboard input from image analysis. If there was, humanity would be in danger. Bots are modified game AI, that is why they caught. Also any aimbot except the fastest ram one is doomed to fail, that is because there is so much that have to be done in order to hit your opponent, not only targeting correctly, a wrong move and aim doesn't matter. Auto-evasion how? With good players like me, only one trigger is needed to end it. With light based switches we also win there.

    Leave a comment:


  • oiaohm
    replied
    Originally posted by artivision View Post
    The real world lag and the accuracy loss make this system unable to stand against pro players like me or even middle range ones. Sorry but we have tested our strength against any kind of cheaters and we only lose a 40% from those who use original ram based aimlock at instant lock speed. Even those who try to obfuscate it using a slower turn / lock speed, they loose badly.
    The sad reality here people using the cheat system I described have have fact won online tournaments. So beating pro players is not a question its been done. Yes this includes having deal with real world lag and accuracy loss.

    There is a major differences between the ram based aimlock and these new second computer solutions.
    1) more processing power.
    2) hardware downscaling(this is a big thing)

    Yes that hardware down-scaling from 1080/4k to 480p and 360p means the processing load to find targets is lighter. Turns out AI target identification need less pixels than human for a game. So you could say this is DLSS in reverse that helps cheaters out. Where you historically would only loss 40% time now comes against these newer cheat methods only able to win 1% of the time if at all if you don't have server side anti-cheat looking for inhuman actions so forcing the cheat to obfuscate their actions so leveling the playing field.

    Remember the old ram based aimlock was not moving the player to avoid being targeted as well. So these newer systems are not just locking on to you to shot you but are also attempting to move the player char they are controlling to avoid you getting getting a target hit on them as well. These are on a totally different level to your old school ram based aimlock. Remember I did mention that one of these was caught using a system like what I described was because a person was attending a real world tournament while winning a online one at the exact same time. So this is serous everyone in the online tournament was facing a total AI controlled char and losing.

    At this stage the those setting up a decanted computer to cheat with are still quite rare. But there numbers are increasing. So anti-cheat solutions are having to evolve to deal with this problem. The fact that tactics are unable to over come the amount of advantage these systems have is the big problem. Yes the old ram based aimlock was not far enough in skill above you that tactics could not beat it and good tactics could beat it quite solidly. Now this generation with more processing power where the item doing the fake inputs can also be automatically doing counter tactics as well this is what results in human players totally screwed if anti-cheat system does not catch them.

    Reality here there is a level past autoaim. Yes auto-aim and auto-evasion and auto-tactical movement though the maps. This has to be the most boring way to play the game because you are no longer playing the game but letting the computer do it for you.

    Leave a comment:


  • artivision
    replied
    Originally posted by oiaohm View Post


    Do read the link artvision there have already been people playing for money using 1080 and greater screen setting caught using this tech. Yes this is part of the ban plea bargain process to admit and provide the cheat technology you were caught with by automated systems to move from a permanent ban to a year only ban..

    So this is not if it worked this is this technology does work and it detected and documented in use on particular e-sports titles. It is a evolution on some of the hyper visor based cheating as well.

    There are different ways char detection can be implemented on screen that have different levels of GPU requirement.

    Is it possible to implement object detection models with real-time performance without GPU? faced is a proof of concept that it is possible…


    There is work to remove the GPU from being required.

    Something else here important max human reflex time with brain in loop is 0.15 of a second but the average is 0.2 of a second(yes 150ms/200ms). So in frame per second that ~7 frames per second. Max reflex with brain not in loop is 0.08 of second or roughy 13 frames a second(80ms). This is not this fast.

    30 frames per second feeding into a AI to controller a computer is enough to beat 99.9% of all computer game players 100 percent of the time. 0.1 is players with great tactics. Scare part here is your 720p is over quality most games to have AI ahead of human you only need 480p and this can be a forced downscale in the capture device most of the time.

    This is the inhuman problem. 30fps at 720p analysis real-time is better than human by quite a margin. Lot of people would not be thinking that analysis real-time running at 30fps at 480p or even 360p is enough to beat the crud out of a human player in computer games when in scary reality that is exactly what is required for 99% of e-sports titles. At 720p you have covered all e-sport titles even the titles that have 4k graphics and the like. 30 frames per second is enough to allow the ai computer to move the controller twice as fast as a real human could at max 1.1x faster than a real human is enough to beat the crap out of a real human in game play so 2x gives quite a lot of processing time slack. This human limitation is not fixed by using 260hz or 300hz... screen. The AI does not need the same quality vision as a human to be competitive and superior in games due to faster reflex speed.

    Remember the AI is racing against a human that can only really move at 7-13 frames a second in a reflex event as in when both players see each other. 30fps vision is enough that AI can always get the killing shot off first

    Some players in e-sports using this tech early have been caught due to total stupidity. Take part in a in real world competition while winning a online competitor is how one fo the first was caught. Human cannot be in 2 places at the same time but when you have a computer playing for you that can absolutely be the case.

    Scary part here a 16 core AMD x86 processor without GPU can process a video output from a game with the right software to beat 99% of all players. Yes the software had to tuned on a GPU at first but once tuned no GPU required. So 2 computers 1 GPU is enough to cheat. Over time this bar will get even lower. This is all due to the human reflex time limits and that fact games are generally not that complex. Driving a car and other interactions with the real world you have to be looking for more types of objects at the same time than you do in games of course that is what the human brain is optimised for. This is why a AI can be way more optimised at playing games than a human and that level of optimisation that results in the AI need way worse video quality and refresh rates to beat human.

    Remember you don't even need quality capture device for a lot of games either. You bottom teir 90ms latency capture devices.still enough to get to 1.1x faster than human for brain in loop games most of the time, 30 frame a second realtime video processing is 34ms of overhead+90ms for video capture device+ 40ms of device faking input yes this is 180ms this is better than the 200ms the human averages and this is enough lead to win competitions . High grade capture card puts makes system able to beat the Max reflex brain not in loop as well because this gets under a 40ms of overhead but that also need 60fps of realtime video processing again human still loses.

    The reality here the lag of the capture device and the fake input device is a bigger problem than the video processing to if a cheating system setup can win against humans or not.

    Yes you need to lose the idea that cheat system has to process the screen at screen output res to beat human.
    The real world lag and the accuracy loss make this system unable to stand against pro players like me or even middle range ones. Sorry but we have tested our strength against any kind of cheaters and we only lose a 40% from those who use original ram based aimlock at instant lock speed. Even those who try to obfuscate it using a slower turn / lock speed, they loose badly.

    Leave a comment:

Working...
X