Page 18 of 21

Chronicling the inability of gamedevs to make video games

Posted: March 18th, 2024, 20:37
by KnightoftheWind
I think anti-cheats are a necessary evil unfortunately, because the alternative would be getting hacked and servers being overrun with cheaters of all shapes and sizes. This has occurred with big games like Team Fortress 2, and even now with games that feature full anti-cheat support. This is a problem of online gaming being extremely popular, with tens of thousands of players and online communities dedicated to them. It's only a matter of time before a bad actor shows up with ill intent, and without some form of anti-cheat the playerbase would be at the mercy of these degenerates.

The best solution is to quit cold-turkey, and avoid all online multiplayer games, but that isn't easy for some.

Chronicling the inability of gamedevs to make video games

Posted: March 18th, 2024, 20:48
by rusty_shackleford
Boontaker wrote: March 18th, 2024, 20:31
Anti cheat is like car insurance these days, a total fucking scam
Clientside anti-cheat has always been a scam. If a game has anticheat, it means the developer probably put things in the client that enable cheating. No clientside anticheat will stop a determined cheater, the only way to do this is to analyze the data on the server for inhuman behavior.

Chronicling the inability of gamedevs to make video games

Posted: March 18th, 2024, 20:52
by Oyster Sauce
Cheating in shooters was solved decades ago with dedicated servers. I have thousands of hours in TF2 and I've never seen a single cheater.

Chronicling the inability of gamedevs to make video games

Posted: March 18th, 2024, 20:53
by Nammu Archag
Oyster Sauce wrote: March 18th, 2024, 19:18
Hackers in Apex Legends are now injecting cheats into other players' computers during tournaments

Total Streamer Death

Chronicling the inability of gamedevs to make video games

Posted: March 18th, 2024, 20:57
by Breathe
Oyster Sauce wrote: March 18th, 2024, 20:52
Cheating in shooters was solved decades ago with dedicated servers. I have thousands of hours in TF2 and I've never seen a single cheater.
Back when community servers were everywhere it was never much of a problem, because the cheaters would be booted pretty quick. Every multiplayer game needs community servers. They also help extend the life of games too.

Chronicling the inability of gamedevs to make video games

Posted: March 18th, 2024, 21:49
by Rand
BattleBit has EasyAnti-Cheat, but it was broken wide open last year and the cheating is rampant now.
Easily 20% of the players in every match are using some sort of hacks. Sometimes more.
There are whole clans known to be ALL (obviously) cheating.
Every "top" player is a dirty fucking cheat. I've literally watched most of them aimbot.
And the devs do nothing.

Chronicling the inability of gamedevs to make video games

Posted: March 18th, 2024, 22:06
by Oyster Sauce
Rand wrote: March 18th, 2024, 21:49
BattleBit has EasyAnti-Cheat, but it was broken wide open last year and the cheating is rampant now.
Easily 20% of the players in every match are using some sort of hacks. Sometimes more.
There are whole clans known to be ALL (obviously) cheating.
Every "top" player is a dirty fucking cheat. I've literally watched most of them aimbot.
And the devs do nothing.
BattleBit was so much fun. I even forged credentials to get the devs to give me veteran status. Good times.

Chronicling the inability of gamedevs to make video games

Posted: March 18th, 2024, 22:09
by gerey
Rand wrote: March 18th, 2024, 21:49
And the devs do nothing.
Didn't the same devs leave Titanfall 2 to rot for years on end, at the mercy of a bunch of cheaters that were ruining the experience for everyone, to the point the game was basically unplayable unless you used community-made fixes?

The devs only bothered to release a patch late last year.

Chronicling the inability of gamedevs to make video games

Posted: March 18th, 2024, 22:43
by Breathe
Rand wrote: March 18th, 2024, 21:49
BattleBit has EasyAnti-Cheat, but it was broken wide open last year and the cheating is rampant now.
Easily 20% of the players in every match are using some sort of hacks. Sometimes more.
There are whole clans known to be ALL (obviously) cheating.
Every "top" player is a dirty fucking cheat. I've literally watched most of them aimbot.
And the devs do nothing.
Oh that's horrible news. I played on release but was thinking about playing again.

Chronicling the inability of gamedevs to make video games

Posted: March 18th, 2024, 23:35
by gerey
lol

lmao even


Chronicling the inability of gamedevs to make video games

Posted: March 18th, 2024, 23:49
by Rand
gerey wrote: March 18th, 2024, 22:09
Rand wrote: March 18th, 2024, 21:49
And the devs do nothing.
Didn't the same devs leave Titanfall 2 to rot for years on end, at the mercy of a bunch of cheaters that were ruining the experience for everyone, to the point the game was basically unplayable unless you used community-made fixes?

The devs only bothered to release a patch late last year.
Not the same devs, but the same, yeah.

Oh, and it was ONE GUY that was basically fucking over the (shitty) Titanfall server code in new ways all the time.

Chronicling the inability of gamedevs to make video games

Posted: March 18th, 2024, 23:51
by Rand
Breathe wrote: March 18th, 2024, 22:43
Rand wrote: March 18th, 2024, 21:49
BattleBit has EasyAnti-Cheat, but it was broken wide open last year and the cheating is rampant now.
Easily 20% of the players in every match are using some sort of hacks. Sometimes more.
There are whole clans known to be ALL (obviously) cheating.
Every "top" player is a dirty fucking cheat. I've literally watched most of them aimbot.
And the devs do nothing.
Oh that's horrible news. I played on release but was thinking about playing again.
Worse, while they never ban for cheating anymore, they ban all the time for people saying gamer words.
Cheating is fine, but calling some retard a faggot will get you banned. What a world.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 00:30
by rusty_shackleford
gerey wrote: March 18th, 2024, 23:35
lol

lmao even

Yea, devs are legally allowed to do this. They can just take your mods and sell them if they wanted.
The amount of people saying 'sue them' is amazing.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 00:35
by Oyster Sauce
rusty_shackleford wrote: March 19th, 2024, 00:30
gerey wrote: March 18th, 2024, 23:35
lol

lmao even

Yea, devs are legally allowed to do this. They can just take your mods and sell them if they wanted.
The amount of people saying 'sue them' is amazing.
RimWorld Guy always includes a few popular mods in his expansions. I'm all for it.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 00:45
by aweigh
Image

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 00:47
by aweigh
rusty_shackleford wrote: March 18th, 2024, 20:48
Boontaker wrote: March 18th, 2024, 20:31
Anti cheat is like car insurance these days, a total fucking scam
Clientside anti-cheat has always been a scam. If a game has anticheat, it means the developer probably put things in the client that enable cheating. No clientside anticheat will stop a determined cheater, the only way to do this is to analyze the data on the server for inhuman behavior.
currently the most popular forms of cheating that i know of are anti-recoil scripts and abusing aim assist. Both can be done with dedicated hardware like this:

Image

There's many such devices, XIM, Nexus, Titan2, etc. They intercept the peripheral calls and send their own.

There is nothing that client-side anti-cheat or anything devs put in an executable that can do anything about this, or if there is then i've never seen them do it. Real cheaters rarely mess around or depend on installed software.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 01:16
by Irenaeus
Oyster Sauce wrote: March 19th, 2024, 00:35
rusty_shackleford wrote: March 19th, 2024, 00:30
gerey wrote: March 18th, 2024, 23:35
lol

lmao even

Yea, devs are legally allowed to do this. They can just take your mods and sell them if they wanted.
The amount of people saying 'sue them' is amazing.
RimWorld Guy always includes a few popular mods in his expansions. I'm all for it.
There's a difference in motivation in that RimWorld guy is a great guy who wants to make a better game for gamers, while that generic soulless Disney-Star Wars vehicle is a product of hellish forces and poisonous minds.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 01:25
by Rand
aweigh wrote: March 19th, 2024, 00:47
rusty_shackleford wrote: March 18th, 2024, 20:48
Boontaker wrote: March 18th, 2024, 20:31
Anti cheat is like car insurance these days, a total fucking scam
Clientside anti-cheat has always been a scam. If a game has anticheat, it means the developer probably put things in the client that enable cheating. No clientside anticheat will stop a determined cheater, the only way to do this is to analyze the data on the server for inhuman behavior.
currently the most popular forms of cheating that i know of are anti-recoil scripts and abusing aim assist. Both can be done with dedicated hardware like this:

Image

There's many such devices, XIM, Nexus, Titan2, etc. They intercept the peripheral calls and send their own.

There is nothing that client-side anti-cheat or anything devs put in an executable that can do anything about this, or if there is then i've never seen them do it. Real cheaters rarely mess around or depend on installed software.
Yep. Microcontrollers. Uncommon, but still too common.
It's the step before you install a card to direct memory access one computer using another computer. Which is hella expensive, but rich cretins are used to cheating all the time in life, so...

In BattleBit, it's mostly mouse macros and AHK stuff, with the big clan cheaters having the better hacks and microcontrollers.
And you can tell instantly because there aren't that many pro players playing the game to equal the amount of people getting pro-level scores. By which I mean that they're outpacing even good players by 3-5x, getting ridiculous kill numbers with few deaths which you only see in pro-level competitions against ordinary gamers. "But he's just a good player, bro." I wanna shoot these morons irl.
Most of the players are retards. "I've played 200 hours and never seen a cheater!" Dumbass, you've seen plenty, you're just stupid and suck at the game or are a cheater yourself.

The solution is AI anticheat.
It's supposed to get good enough that it can even detect previously banned cheaters on a new account because it can detect the playstyle.
I'm sure it won't get them all, but if it gets 80% that would be a huge boon.

Unfortunately it's server-side and most devs today are half retarded.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 01:53
by Analogue Dreams
Rand wrote: March 19th, 2024, 01:25
Yep. Microcontrollers. Uncommon, but still too common.
It's the step before you install a card to direct memory access one computer using another computer. Which is hella expensive, but rich cretins are used to cheating all the time in life, so...

In BattleBit, it's mostly mouse macros and AHK stuff, with the big clan cheaters having the better hacks and microcontrollers.
And you can tell instantly because there aren't that many pro players playing the game to equal the amount of people getting pro-level scores. By which I mean that they're outpacing even good players by 3-5x, getting ridiculous kill numbers with few deaths which you only see in pro-level competitions against ordinary gamers. "But he's just a good player, bro." I wanna shoot these morons irl.
Most of the players are retards. "I've played 200 hours and never seen a cheater!" Dumbass, you've seen plenty, you're just stupid and suck at the game or are a cheater yourself.

The solution is AI anticheat.
It's supposed to get good enough that it can even detect previously banned cheaters on a new account because it can detect the playstyle.
I'm sure it won't get them all, but if it gets 80% that would be a huge boon.

Unfortunately it's server-side and most devs today are half retarded.
How does memory access direction work, what sort of card do they use? Is it like a KVM switch to send commands from one computer to another so rootkit anticheat can't detect it? Think I've heard something about cheats like those before, but never really understood how they work. It amazes me the lengths people go to cheat at online games, especially all the bottom of the barrel ones that are popular now.

AI anticheat doesn't sound like it'd be reliable unless you mean something more specialised than what I'm guessing, which is training an LLM to do it. They can't reliably identify the time on a watch in photos or who is wearing it, although LLMs are much better at telling the time if the time just so happens to be the most prevalent time featured on a clock in advertising photos and data. Next-token prediction and the way LLMs rely on statistical prevalence of data don't sound good to me for finding cheaters.

All the attempts to solve cheating seem like the wrong approach imo, the more tech everyone has the more it seems everyone forgets how to do basic things. Way it used to work, like in old RTS at one point, was just by user reports and the dev looking into it and investigating whatever proof you sent them. You emailed them your replay of the game match and whatever other evidence you had, then some intern takes a look (or not if they were an intern at a prevalent company that decided to abandon every previous player base of theirs in 2004). If it checks out, they ban their CD key.

Games that did that were some of the biggest at the time that everyone was playing too, they weren't small, no reason devs couldn't do it now when they have way more money to employ customer servants. Better a real person finds and determines who is actually cheating than an algorithm or rootkit. Obviously the only problem is they're too incompetent now, but there's no workaround that'll make them act as if they were competent anyway.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 02:00
by Rand
Analogue Dreams wrote: March 19th, 2024, 01:53
Rand wrote: March 19th, 2024, 01:25
Yep. Microcontrollers. Uncommon, but still too common.
It's the step before you install a card to direct memory access one computer using another computer. Which is hella expensive, but rich cretins are used to cheating all the time in life, so...

In BattleBit, it's mostly mouse macros and AHK stuff, with the big clan cheaters having the better hacks and microcontrollers.
And you can tell instantly because there aren't that many pro players playing the game to equal the amount of people getting pro-level scores. By which I mean that they're outpacing even good players by 3-5x, getting ridiculous kill numbers with few deaths which you only see in pro-level competitions against ordinary gamers. "But he's just a good player, bro." I wanna shoot these morons irl.
Most of the players are retards. "I've played 200 hours and never seen a cheater!" Dumbass, you've seen plenty, you're just stupid and suck at the game or are a cheater yourself.

The solution is AI anticheat.
It's supposed to get good enough that it can even detect previously banned cheaters on a new account because it can detect the playstyle.
I'm sure it won't get them all, but if it gets 80% that would be a huge boon.

Unfortunately it's server-side and most devs today are half retarded.
How does memory access direction work, what sort of card do they use? Is it like a KVM switch to send commands from one computer to another so rootkit anticheat can't detect it? Think I've heard something about cheats like those before, but never really understood how they work. It amazes me the lengths people go to cheat at online games, especially all the bottom of the barrel ones that are popular now.

AI anticheat doesn't sound like it'd be reliable unless you mean something more specialised than what I'm guessing, which is training an LLM to do it. They can't reliably identify the time on a watch in photos or who is wearing it, although LLMs are much better at telling the time if the time just so happens to be the most prevalent time featured on a clock in advertising photos and data. Next-token prediction and the way LLMs rely on statistical prevalence of data don't sound good to me for finding cheaters.

All the attempts to solve cheating seem like the wrong approach imo, the more tech everyone has the more it seems everyone forgets how to do basic things. Way it used to work, like in old RTS at one point, was just by user reports and the dev looking into it and investigating whatever proof you sent them. You emailed them your replay of the game match and whatever other evidence you had, then some intern takes a look (or not if they were an intern at a prevalent company that decided to abandon every previous player base of theirs in 2004). If it checks out, they ban their CD key.

Games that did that were some of the biggest at the time that everyone was playing too, they weren't small, no reason devs couldn't do it now when they have way more money to employ customer servants. Better a real person finds and determines who is actually cheating than an algorithm or rootkit. Obviously the only problem is they're too incompetent now, but there's no workaround that'll make them act as if they were competent anyway.
This is a pretty good primer. It focuses on Valorant, but applies to all games. There's a timestamp for DMA cheats in the video.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 02:00
by Oyster Sauce
Analogue Dreams wrote: March 19th, 2024, 01:53
Rand wrote: March 19th, 2024, 01:25
Yep. Microcontrollers. Uncommon, but still too common.
It's the step before you install a card to direct memory access one computer using another computer. Which is hella expensive, but rich cretins are used to cheating all the time in life, so...

In BattleBit, it's mostly mouse macros and AHK stuff, with the big clan cheaters having the better hacks and microcontrollers.
And you can tell instantly because there aren't that many pro players playing the game to equal the amount of people getting pro-level scores. By which I mean that they're outpacing even good players by 3-5x, getting ridiculous kill numbers with few deaths which you only see in pro-level competitions against ordinary gamers. "But he's just a good player, bro." I wanna shoot these morons irl.
Most of the players are retards. "I've played 200 hours and never seen a cheater!" Dumbass, you've seen plenty, you're just stupid and suck at the game or are a cheater yourself.

The solution is AI anticheat.
It's supposed to get good enough that it can even detect previously banned cheaters on a new account because it can detect the playstyle.
I'm sure it won't get them all, but if it gets 80% that would be a huge boon.

Unfortunately it's server-side and most devs today are half retarded.
How does memory access direction work, what sort of card do they use? Is it like a KVM switch to send commands from one computer to another so rootkit anticheat can't detect it? Think I've heard something about cheats like those before, but never really understood how they work. It amazes me the lengths people go to cheat at online games, especially all the bottom of the barrel ones that are popular now.

AI anticheat doesn't sound like it'd be reliable unless you mean something more specialised than what I'm guessing, which is training an LLM to do it. They can't reliably identify the time on a watch in photos or who is wearing it, although LLMs are much better at telling the time if the time just so happens to be the most prevalent time featured on a clock in advertising photos and data. Next-token prediction and the way LLMs rely on statistical prevalence of data don't sound good to me for finding cheaters.

All the attempts to solve cheating seem like the wrong approach imo, the more tech everyone has the more it seems everyone forgets how to do basic things. Way it used to work, like in old RTS at one point, was just by user reports and the dev looking into it and investigating whatever proof you sent them. You emailed them your replay of the game match and whatever other evidence you had, then some intern takes a look (or not if they were an intern at a prevalent company that decided to abandon every previous player base of theirs in 2004). If it checks out, they ban their CD key.

Games that did that were some of the biggest at the time that everyone was playing too, they weren't small, no reason devs couldn't do it now when they have way more money to employ customer servants. Better a real person finds and determines who is actually cheating than an algorithm or rootkit. Obviously the only problem is they're too incompetent now, but there's no workaround that'll make them act as if they were competent anyway.
Cheats regularly advance and become more difficult to detect, guys watching replays to find evidence of cheating have peaked and will never get better.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 02:02
by Rand
Analogue Dreams wrote: March 19th, 2024, 01:53
All the attempts to solve cheating seem like the wrong approach imo, the more tech everyone has the more it seems everyone forgets how to do basic things. Way it used to work, like in old RTS at one point, was just by user reports and the dev looking into it and investigating whatever proof you sent them. You emailed them your replay of the game match and whatever other evidence you had, then some intern takes a look (or not if they were an intern at a prevalent company that decided to abandon every previous player base of theirs in 2004). If it checks out, they ban their CD key.

Games that did that were some of the biggest at the time that everyone was playing too, they weren't small, no reason devs couldn't do it now when they have way more money to employ customer servants. Better a real person finds and determines who is actually cheating than an algorithm or rootkit. Obviously the only problem is they're too incompetent now, but there's no workaround that'll make them act as if they were competent anyway.
That's why BattleBit gave up.
Three cheap-ass devs.
Can't be arsed to ban cheaters anymore.
It's a literal full-time job with their garbage anti-cheat implementation.
And they opened up community servers, who themselves can't be bothered.
A couple of community servers are run by clans who cheat, too.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 02:02
by aweigh
Analogue Dreams wrote: March 19th, 2024, 01:53
Way it used to work, like in old RTS at one point, was just by user reports and the dev looking into it and investigating whatever proof you sent them. You emailed them your replay of the game match and whatever other evidence you had, then some intern takes a look (or not if they were an intern at a prevalent company that decided to abandon every previous player base of theirs in 2004). If it checks out, they ban their CD key.
This is still exactly how they do it. Just not for cheating.

This is what they do when you send a screenshot of someone saying nigger.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 02:03
by rusty_shackleford
Oyster Sauce wrote: March 19th, 2024, 02:00
guys watching replays to find evidence of cheating have peaked and will never get better.
AI models are trained on cheaters and used to detect them. Valve already does this, don't know if others do.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 02:04
by Oyster Sauce
rusty_shackleford wrote: March 19th, 2024, 02:03
Oyster Sauce wrote: March 19th, 2024, 02:00
guys watching replays to find evidence of cheating have peaked and will never get better.
AI models are trained on cheaters and used to detect them. Valve already does this, don't know if others do.
Ok, but AI models aren't people watching replays. I agree that technology is the way forward rather than human intuition.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 02:05
by Rand
Oyster Sauce wrote: March 19th, 2024, 02:00
Cheats regularly advance and become more difficult to detect, guys watching replays to find evidence of cheating have peaked and will never get better.
Yeah. For example, the new triggerbots are hard for a human to detect since it just fires the weapon only when you manually line up the shot.
It just has effective perfect human accuracy, firing only when there's a possible hit and stops the instant the target is dead.
Add in normal weapon spread and some shots still miss. If you're smart, you program in just a little randomness to the bot.
I think I see it pretty often in BattleBit, but you can't be sure unless you watch someone play a match and be just too damned good all the time.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 02:06
by Shillitron
rusty_shackleford wrote: March 19th, 2024, 02:03
Oyster Sauce wrote: March 19th, 2024, 02:00
guys watching replays to find evidence of cheating have peaked and will never get better.
AI models are trained on cheaters and used to detect them. Valve already does this, don't know if others do.
After 20 epochs of crunching data, using entire data centers... the AI has deduced.. any female streaming playing well - is wall hacking.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 02:06
by aweigh
They don't bother with cheating because there's no BlackRock incentive to police cheaters, however there is huge monetary incentive in the industry to police what people say in chat, specifically what they refer to as "hate speech".

Everything that everyone here wants done about cheating is what they're already doing to combat "hate speech". No money to police cheaters, however.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 02:07
by rusty_shackleford
Rand wrote: March 19th, 2024, 02:05
Oyster Sauce wrote: March 19th, 2024, 02:00
Analogue Dreams wrote: March 19th, 2024, 01:53

How does memory access direction work, what sort of card do they use? Is it like a KVM switch to send commands from one computer to another so rootkit anticheat can't detect it? Think I've heard something about cheats like those before, but never really understood how they work. It amazes me the lengths people go to cheat at online games, especially all the bottom of the barrel ones that are popular now.

AI anticheat doesn't sound like it'd be reliable unless you mean something more specialised than what I'm guessing, which is training an LLM to do it. They can't reliably identify the time on a watch in photos or who is wearing it, although LLMs are much better at telling the time if the time just so happens to be the most prevalent time featured on a clock in advertising photos and data. Next-token prediction and the way LLMs rely on statistical prevalence of data don't sound good to me for finding cheaters.

All the attempts to solve cheating seem like the wrong approach imo, the more tech everyone has the more it seems everyone forgets how to do basic things. Way it used to work, like in old RTS at one point, was just by user reports and the dev looking into it and investigating whatever proof you sent them. You emailed them your replay of the game match and whatever other evidence you had, then some intern takes a look (or not if they were an intern at a prevalent company that decided to abandon every previous player base of theirs in 2004). If it checks out, they ban their CD key.

Games that did that were some of the biggest at the time that everyone was playing too, they weren't small, no reason devs couldn't do it now when they have way more money to employ customer servants. Better a real person finds and determines who is actually cheating than an algorithm or rootkit. Obviously the only problem is they're too incompetent now, but there's no workaround that'll make them act as if they were competent anyway.
Cheats regularly advance and become more difficult to detect, guys watching replays to find evidence of cheating have peaked and will never get better.
Yeah. For example, the new triggerbots are hard for a human to detect since it just fires the weapon only when you manually line up the shot.
It just has effective perfect human accuracy, firing only when there's a possible hit and stops the instant the target is dead.
Add in normal weapon spread and some shots still miss. If you're smart, you program in just a little randomness to the bot.
I think I see it pretty often in BattleBit, but you can't be sure unless you watch someone play a match and be just too damned good all the time.
That's most likely something an AI model could detect easily. Anything that's abnormal gets picked up and flagged for review.
The guys who worked on it for Valve did a talk at GDC about it.

Chronicling the inability of gamedevs to make video games

Posted: March 19th, 2024, 02:11
by Oyster Sauce
rusty_shackleford wrote: March 19th, 2024, 02:07
Rand wrote: March 19th, 2024, 02:05
Oyster Sauce wrote: March 19th, 2024, 02:00


Cheats regularly advance and become more difficult to detect, guys watching replays to find evidence of cheating have peaked and will never get better.
Yeah. For example, the new triggerbots are hard for a human to detect since it just fires the weapon only when you manually line up the shot.
It just has effective perfect human accuracy, firing only when there's a possible hit and stops the instant the target is dead.
Add in normal weapon spread and some shots still miss. If you're smart, you program in just a little randomness to the bot.
I think I see it pretty often in BattleBit, but you can't be sure unless you watch someone play a match and be just too damned good all the time.
That's most likely something an AI model could detect easily. Anything that's abnormal gets picked up and flagged for review.
The guys who worked on it for Valve did a talk at GDC about it.
Hope so, but CS2 is still rife with cheaters as far as I know.