Valorant Dev Tracker

Yes! I'd like to have an Overwatch system one day, I think it'll need to wait until after we build a general purpose replay system.

When building our aimbot detection AI research project we definitely were paying attention to how VACnet works. John McDonald from Valve talks about it here: https://www.youtube.com/watch?v=ObhK8lUfIlc (it's a good watch if you're interested in this kind of thing)

Definitely!

The current plan is to reduce packet rate by not sending packets that have little to no new information in them. For example right now in VALORANT if you don't move the mouse on a frame we still send a packet. These packets don't actually help responsiveness so we can skip them to improve the experience at no cost.

If we do end up reducing packet rate in a way that could impact responsiveness I think we would want to make it an opt-in setting (like a fps cap setting for packets). Currently you can kind of get that effect from the fps cap but maybe in the future we can make fps and packet rate independently configurable.

• SFW only please!
• Don't forget to credit yourself and link your art pages, if you have any!
• I'd love to retweet some of my favorites; please let me know if you're not comfortable with that!

Yeah I chose a deliberately outlandish example because I don't want to completely rule out the possibility countermeasures for any techniques just yet :)

Currently our aimbot detection AI is a research project that is used to inform analyst investigations. Once we're more confident in its results we hope to let it ban players unsupervised but due to the error prone / probabilistic nature of AI we think that it'll primarily be an analyst investigation tool for a while yet.

We try to make our anti-cheat systems hard to analyze for cheat developers but there are some limits because at the end of the day the client is running on the cheat deverloper's computer.

It's not too unusual for a game to send one packet per frame though often there's a cap. VALORANT ends up sending a lot because frame rates are high and we're trying to make the game as responsive as possible.

Round end in VALORANT is a time where a lot of cleanup is happening, a lot of objects are being created and destroyed and that makes for longer frame times and more packets. It's not too surprising that you can get a ping spike at the start of the round. Hopefully we can improve this in the future!

I've written an article about the Fog of War system, hopefully it'll come out next week. It includes a gameplay clip that shows a wallhack both before and after Fog of War. Hopefully that'll be helpful!

I agree!

It's going to take a lot of time and effort to keep cheaters under control but I think it's necessary if we want to create a competitive shooter that lasts.

Thanks for the kind words!

One of the reasons I like talking about Fog of War is that it's one of the rare anti-cheat systems where knowing how it works doesn't make it less useful. We often don't share details of our security measures in fear of helping out cheaters but this is an area where I think being transparent is helpful.

I experimented with this a little bit and it was harder than I thought.

The problem I ran into is that how the sound is "rendered" depends a lot on the exact location and not just the attenuation due to distance. There are factors like occlusion and special effects like reverb that get applied based on position. Even the sound asset itself can vary based on exact location (imagine footsteps on grass versus footsteps on wood).

I concluded that it'd be possible to get it mostly working but the effort wasn't worth it. You'd end up with a mostly correct sound (probably a noticeable difference to what was originally intended) but a hacker could still guess a location from it with reasonable accuracy by taking the direction and volume of the sound.

Maybe someone cleverer than me will figure out a solution, if so I'd be happy to hear it!

It's similar to web security bounties, reports are eligible for a reward if they lead to a change in the software to address the reported flaw.

So for example if a researcher reports a way to defeat the mechanism that ensures that the game client executable hasn't been modified and in response we make a change to fix the attack (by improving the mechanism) they'd be eligible for a reward.

On the other hand if a researcher created a general purpose AI that could play the game impossibly well by using a keyboard and mouse that report is unlikely to be eligible for a reward since there's no improvements we can make.

Though this would be the least of their (and our) worries as the tac shooter robot uprising would soon begin likely destroying human civilization.

Yeah, the need to play sounds for enemies even if you can't see them does weaken the Fog of War system, especially if a sound visualizing hack is used.

Sorry for the inconvenience! If we find a way to support cloud gaming services without weakening our anti-cheat protections we're definitely open to supporting them. We don't have any plans for this right now though.

Because pushing you out of a safe position in to something like a Vandal trained on you is a deadly threat. It takes threat of death to force repositioning - you can soak a Diva arrow, but Raze is about small area, high threat position clearing

We're working on it! It's not anti-cheat related though, it's because of how our netcode works.

What's happening is that the game is sending a packet every frame (sometimes more) to try and maximize game responsiveness. However, if your frame rate gets high enough the amount of packets can overwhelm your router (or your upstream bandwidth).

The workaround for now is just what you suggest (turning the FPS limiter on) but in an upcoming patch we'll have a fix that reduces the number of packets sent so that we don't accidentally crush routers.

Well said!

I wouldn't say that it's fine exactly but I don't think the measure of success is whether or not there can be cheats, instead it's what impact cheats are having on the game as a whole.

I'm hoping to get the bans as close to the detection as possible, even terminating a match if a player in it is detected. Sometimes there will be a delay because the cheater will be detected by some offline processing but I believe that where possible cheaters should be removed quickly.

It's a balancing act though because each ban gives cheat developers more information about what is detected and what is not, that's one reason that banwaves are a thing. Sometimes a banwave makes sense to try for maximum disruption of a cheat developer's operation.

We'll probably use both strategies but I'm going to advocate for immediate bans as much as possible because I want to remove cheaters from the environment before they damage any more matches.

Yeah, having anti-tamper technology is important but it's definitely not all we're doing!

Support tickets are great for bugs because the support agents know how to route the reports to the right developers. Other than that I know that devs are reading the bug megathread so that's another decent spot!

Edit: I'm pretty sure we've received other reports of this map bug, I heard someone mention it today so if I'm not confused we're working on a fix at the moment.