Go to article list
04/15/22Game Updates

VALORANT Gameplay Consistency Update

Share:

Hey folks, VALORANT Gameplay team here with a quick update on the investigation we’re running on game-to-game inconsistencies that some of our community has reported. TL;DR, We’re still actively working on this, and it remains one of our top priorities for the team. In the meantime, here’s an update on the approach we’re taking and findings so far.

WHAT WE’RE HEARING

When it comes to reports of inconsistent combat, these are the main themes that we’ve heard from some of you:

  1. “I'm feeling a different amount of delay from one game to the next, even when my network conditions are stable.”
  2. “It feels like other players can respond to me faster / slower from one game to the next”
  3. “I feel very inconsistent when I’m playing VALORANT. Some games I'm the one one-tapping headshots, but in the next game it’s happening to me.”
  4. ”I feel high ping players have an unfair advantage when peeking me compared to low ping players”

INVESTIGATION OVERVIEW

Many factors influence how combat feels from one match to the next, but you can lump them in two major categories: “individual player performance” and “competitive integrity (fairness).”

Individual player performance acknowledges that flow state is a thing. When players are in the zone, they achieve a higher level of focus where their game sense/awareness is increased, their reaction times are decreased, and their flicks are landing with higher accuracy. Consistency in performance is something that high skill players strive to improve with training and practice, but most players perform inconsistently within and across matches.

To further complicate things, your opponents skill levels and flow states directly impact your own perceived performance. If that enemy Jett is feeling it, her low reaction times and high accuracy may be indistinguishable from her having more time to line up a shot due to a network advantage.

So that’s the challenge:

  • Setting aside individual player performance, how can we quantify and measure fairness in a match?
  • Are there any underlying issues in core systems that may cause combat to feel different from one match to the next?
  • What tools or data can we provide you to help distinguish between your performance and fairness issues in your games?

We’re auditing our systems for any issues that might lead to match-to-match inconsistencies, making a few improvements, and then adding some extra graphs for you all to help diagnose (and report back!) any issues that you’re seeing in your games moving forward.

Our investigation has three-prongs: Latency, Correctness, and Peekers with High Ping.

LATENCY

This has been our main focus! Most of the themes we continue to hear could be explained by variations in end-to-end latency that aren’t captured in existing measurements.

We already provide graphs for most of the inputs to your total latency: round-trip network travel time to server, input latency, render latency, and client/server tick rates. Yet, reports still come in of an inconsistent feel even when these measurements all appear stable.

So what does that leave? Network buffering and network jitter.

Let’s go over some basics. The server receives a constant stream of input data from your client, which it then validates and plays back into its authoritative simulation of the world. It needs to process your inputs at a smooth 128x/second, but variable network conditions and client framerates mean that it often receives player inputs in uneven, unpredictable bursts.

Network buffering attempts to smooth gameplay by introducing a small amount of delay before processing player input data to hide inconsistencies in how packets arrive.

Network jitter refers to the variance in how long it takes sent packets to reach their destination. If jitter is low, the server receives packets from you at reliable intervals. It can keep buffer times extremely low, since it knows when the next packet will arrive. If jitter is high, the server needs to delay your inputs by a larger amount on average, because there are frequent, unpredictable gaps between one packet and the next. The same goes for how your client receives enemy movement data and world updates from the server.

Here’s what we’re reviewing:

  • Is network buffering behaving consistently from one match to the next (relative to network conditions)?
  • How well does VALORANT’s buffering handle long-term jitter? Does it effectively minimize processing delays while providing smooth movement?
  • How quickly does buffering recover from short-term disruptions (bursts of packet loss, framerate throttling from alt+tabbing, etc)?

So far, we've identified some minor improvements that we can make to help with latency consistency in non-ideal conditions:

  • Currently, when network buffering jumps in size due to a network or performance hiccup, the system takes longer than we'd like to recover. During that short recovery window, you may experience higher end-to-end latency. This is most noticeable when alt+tabbing with background framerate throttling set to a low value. We're on our way to implement improvements to address this.
  • Also, VALORANT minimizes net buffering to keep input latency and your view of other players as close to real time as possible. However, for high jitter networks, we've noticed that this can cause a cyclic pattern of latency adjustments every few seconds as we prioritize low latency first, but then account for network jitter. We're working on modifying buffering behavior to better account for this scenario.

Investigating in this area is a work-in-progress, and we’ll give another update later this year on our findings.

CORRECTNESS

This part of the investigation is a sanity-check on whether combat rewards consistent player performance with consistent results. Our goal here is to review gameplay systems that could contribute to inconsistencies in individual player performance from one match to the next.

We built diagnostics and internal tooling to answer the following questions:

  • Can network conditions impact the way you need to play to land accurate shots?
    • If so, this would mean that you need to slightly change your behaviors from one match to the next (counterstafe timings, weapon recovery times, etc).
  • Can your client and the server ever disagree on how shots are processed?
    • If so, this would mean that even though you’re playing well on your client, the server might not always agree on the results.
    • Can they disagree on movement accuracy or recoil states at the time of shots?
    • Can they disagree on the vector that the shot travels along?
    • Can they disagree on the target's animation pose when shots are processed?

Based on our investigation so far, everything's looking good on the correctness front and our combat systems are performing up to standard. We'll continue to monitor these systems moving forward in case issues arise in the future.

PEEKERS WITH HIGH PING

Lastly, we’ve been hearing more feedback from some of you that it feels like trying to hold angles against high ping peekers is a losing endeavor. In VALORANT, we don’t want your experience in the game to be affected by the latency of others so this sentiment is something we take quite seriously. For a refresher on how we think about peeker’s advantage, take a look at this previous article.

So far, we’ve mainly been focusing on investigating the previous two areas mentioned. We wanted to make sure the core systems were functioning as expected before looking elsewhere in case some fundamental issues could explain the troubles. Once we wrap up the core systems audits, we’re going to be doing some work to help us investigate why players are feeling this way about high ping players.

We’ll upgrade our internal tooling tailored specifically to measure the exact peeker’s advantage. Using that tooling, we will test various networking scenarios to exactly measure whether an aggressor’s high ping can have a noticeable effect on these values. In addition to gathering internal data with this tooling, we are also reviewing data from live matches to evaluate trends in how high ping impacts your performance and match outcomes.

NEXT STEPS

Thanks for taking the time to read the update here and for your patience as we continue this work.

Another update detailing any findings will be published when we’re approaching a PBE release for any potential changes. If you've noticed inconsistency issues in your games and live in the North America region, please sign up for PBE to try out any fixes early and let us know whether they make a difference.

Beyond this investigation, we have future work planned to continue improving, monitoring, and internal testing capabilities for these systems. We’re also interested in adding extra tools to help you all understand and improve your own performance within and across matches.

—VALORANT Gameplay Team

0