Is this a joke? Machine learning as anti cheat?

Okay, let’s say you cheat and change your number. Now you’re going to have to spoof your HWID’s because those are being banned. Now you have to go through the whole FTUE and all the stuff required to play Comp, you know, where you get the most benefit out of the cheat.

The phone number is just another annoying part for cheaters to get around. Pretty sure your phone carrier is going to be suspicious of you changing your phone number constantly as well.

He literally is, and anyways the fact that he is 150SR above rank 2 is highly sus. Wouldn’t you investigate this if you were Blizzard? There’s no downside to reviewing it

Lol… the game didn’t cancel with 4 cheaters in it.

Tell me, do you think Blizzard is doing a good job of handling cheaters?

There are anti cheat where you can’t even start the game if you have cheats :clown_face:
Kernel ring 0 anti cheat was one of the way but they didn’t, they trust a ML instead and only ML…

Noone was talking about human you don’t know anything about AC just quiet.

They are literally use Kernel Ring 0 for Call of Duty.

Good.

Yeah ??
We’re not playing COD here

Literally the same company.

It sounds like you really dont understand how effective a lot of ML/AI models are. A lot of people have been pushing for machine learning as the next evolution of anti-cheat because it’s capable of detecting things that not even kernel level anticheat can.

If you think this is a bad idea then you’re extremely naive.

2 Likes

I don’t get it, i’m not against kernel anti cheat …
As long as it protects the game properly even if it’s not 100%, but counting only on ML is not enough.

Same as above.

AI based anticheat should be much more effective than even kernel mode anticheat if done correctly.

4 Likes

Yeah ,we will see that in few months how efficient it is :joy: :ok_hand:

Firstly, nobody said the first iteration would be perfect (hence the “when done properly”), and it’s almost certainly still paired with their existing anticheat.

But my god, your posts read like a 12 year old who just watched his first youtube video on machine learning and has it all figured out.

2 Likes

Good post though just “naive” would probably do. :wink:

I’m only tangentially associated with ML through work. There’s a ton of different approaches they might use but they probably involve “training” a system as to what are cheats and what are not and dialing in the parameters based on human proofing during testing.

It will almost certainly identify wall hacks and head shot snapping very quickly. This is a good thing for players and bad for cheaters.

It’s not their first theyve been doing that for a while :clown_face:
Their existent anti cheat is non-existent :clown_face:
The game has been a disaster especially in high elo i know exactly what i’m talking about.

Nice to see you have nothing else to say than trash talking my comment…
See you in few months complaining buddy

Well, using the clown emoji in a reply is basically an automatic validation of my statement, so thank you.

This is the issue. Doing experimental anti-cheats is fine and anything to stop cheaters should be tested but companies these days have 0 effort put in place to attempt to reverse wrongful bans unless a mass amount of players are banned.
The only company I know that does reverse bans is Valve with CSGO.

Perhaps, but machine learning is largely just a scatter plot, with a best fit curve.

But instead of 2 variables, it’s like 15 to 100 variables.

Like what would be the housing price increase for a specific location, if they added a third bathroom.