04: On Peeker’s Advantage & Ranked
New week, new us. From the last week to this one, our family and community of VALORANT has grown exponentially, and we’re happy to be able to see players in Latin America, Brazil, and Korea getting into the closed beta! Welcome! And now that you’ve all been playing VALORANT, we assume you also have endless questions that we’re here to answer.
We’re going to keep our intros short because there are three topics we’d like to cover. The first is one we’re passionate about: peeker’s advantage. Specifically, in the times when an enemy player comes around the corner and appears to be running and gunning while still being accurate. Here’s David Straily, VALORANT’s technical engineering lead to explain:
Hi all! Let’s get right to it:
---- PEEKER'S ADVANTAGE ----
Until humanity learns to move faster than the speed of light, peeker's advantage is never going away. It’s something that exists for all games. That being said, our goal is to minimize it as much as possible in VALORANT. Philosophically, we believe the less peeker's advantage there is, the more tactical the game's meta becomes.
How do you measure it?
It's <ENEMY'S CLIENT FRAMERATE> + <ENEMY'S 1-WAY NETWORK LAG> + <SERVER FRAMERATE> + <YOUR 1-WAY NETWORK LAG> + <NETWORK INTERP DELAY>
Some of the values are constant:
NETWORK INTERP DELAY = 7.8125ms (this is "buffering" to smooth out enemy player positions on your screen. It is tuneable with the "Network Buffering" setting in the options menu, if you have poor internet or bandwidth)
SERVER FRAMERATE = 7.8125ms (1/128th of a second, 128-tick servers)
Others are not:
1-WAY NETWORK LAG is based on game server locations, your (and your opponents) geographic locations, and Riot Direct routing paths. We are targeting <17.5ms for 70% of the player population. Closed beta is a time for testing and improvement, and we are on our way to hitting this target.
YOUR CLIENT FRAMERATE is based on how fast your computer runs; we are trying to optimize framerate as much as we are able to on our side, as a dev team.
The sum of all the above values is a player's advantage when rolling to you from an angle. It is the amount of time that a player sees you, without you seeing them.
Average human reaction time is about 247ms, no joke . Peeker's Advantage in VALORANT is an average of 40-70ms, but even a minor head start is still meaningful.
---- STRAFE SHOOTING ----
It's likely you've come across a situation where you are killed by a player who appears to be mid-run. In actuality, on the server and from their point of view, they are standing still.
We devs are aware of this issue, and are thinking about some ways to improve it.
First, some context:
Network Interpolation Delay (see my comments in the first block above) causes a 7.8125ms desync between movement data and damage data. Otherwise said, when you get killed, the player's move position is always shown 7.8125ms behind from where they actually fired.
This can make it appear as if the player was moving when they shot you, when they might have actually been standing still. This is a more prevalent issue when players are using counter-strafing effectively, as the typical muscle memory is to shoot in the frames after a counter-strafe.
So what are we as devs doing about it?
A) We're investigating all parts of moving/shooting accuracy to understand what's underscoring the issue—whether it's players that are able to become too accurate too quickly from full movement, or that shots on-the-move in certain scenarios are actually more accurate than intended. When we do have changes, you’ll see it in the patch notes.
B) We are working on animation blending updates. When players come to a stop, we'd like to speed up the transition of their animation from “running” to “standing.” The animation is sometimes lagging versus what has “actually” happened.
C) On the frame you die, your corpse (typically) blocks full visibility of the enemy, so you can't see what the opponent did immediately after you died. We are going to fix “corpse-blocking” in the next patch or two, so you can have constant vision of the enemy. This will help to see if they actually did come to a standstill while moving.
D) The “death” message sent to your computer usually arrives ahead of the move simulation timeline. We can “delay” your death by a few milliseconds such that you see the player kill you as they actually did (ex: standing still). Downside here is, during that “delay” time... you’re in purgatory: Dead but given a fake extension of life to sync timelines.
E) We might consider a player setting for “no” remote interp delay; it will get rid of that 7.8125ms desync, but at the expense of you seeing a lot of visual “player pops” when you have packet loss. Packet loss on average varies, but it is not uncommon to have 1-3%. With those numbers, you might see pops of up to once-a-second. Not ideal. We're going to test with options A and B (and maybe some other small design tweaks) to see if we can get to a good place, before we consider this extreme third option.
At risk of writing a novel, I’ll also address...
---- DO HIGH PING PLAYERS CAUSE WORSE PEEKER’S ADVANTAGE? ----
Yes. This is the unfortunate nature of the internet.
---- DO HIGH PING PLAYERS HAVE A GAMEPLAY ADVANTAGE? ----
In short, no. Low latency is near always superior, as your player actions get to the game server faster than your opponents' actions.
The only tricky bit with a high ping player is that it does cause worse peeker's advantage, for everyone. But you as a low-ping player have just as much peek advantage on your high-ping opponent as they have against you.
Consider things a level playing-field, even though the game experience is degraded for everyone. (** - See exception notes below)
PRO TIP: Play aggressive in high-ping disparity games.
Playing with high ping is not the ideal, but it is the short-term reality considering the smaller CB populations & ongoing infrastructure build-out. We as Riot want everyone to have good ping. We have some work underway to improve, so stay tuned for more in a future update.
** - NOTE: Peeker's Advantage is only equal if the high ping player has symmetric lag (the same latency from CLIENT->SERVER, as SERVER->CLIENT).
If lag is asymmetric (high latency up to the server, low latency down to the client), that individual will have higher levels of peeker's advantage.
This is something we as a dev team find unacceptable. Some of the asymmetry is Riot's fault (we need to further optimize routing to our game servers), and some of this is player abuse (lag-switches). We will be tackling both topics in the near future.
Still with us? Here’s Ian Fielding, product manager of the VALORANT competitive team, to give some insight into ranked in VALORANT, as well as answer a question that’s… pretty familiar to us at Riot Games.
Why does VALORANT allow any premade size for ranked? Why not a true solo queue?
It's important to us that competitive matchmaking in VALORANT is focused on teamplay, because we believe that playing as a team is a major component in your overall mastery of the game. If players have a group of teammates they perform well with, we don't want to discourage them and set a precedent that the real test of skill is in solo play.
Having a solo queue can easily lead to that becoming the definitive test of someone's skills and the primary way to play competitively. We've opted instead to allow players to play at any team size they prefer. We also think it’s valuable for players to search for good teammates now for competitive play, so when higher-stakes competition becomes available, they already have teammates they can rely on.
If you are playing solo or with smaller group sizes, we’ve worked to make it so our matchmaking will favor placing players against similar premade team sizes. Also, outside of playing at the highest rank, we’ve included measurement on a small degree of ‘performance’ in determining how your rank increases or decreases. So if you perform exceptionally well and have an outsized contribution to your team’s win, you‘ll rank up faster.
We’ve also heard some concerns from players who feel they need to maximize their KDA (kill-death ratio) to boost their performance evaluation, and how doing so is not always aligned with their team winning.
Let’s be clear: Winning or losing a match has the biggest effect on your rank. If you aim for something other than winning (like KDA), and you lose matches because of it, your rank will trend downward.
Now that Competitive mode is out, what’s next?
Part of the reason we wanted to get competitive matchmaking out early in VALORANT Closed Beta is so that we could have a conversation with all of you and begin building the mode together. We’ve heard a lot of great feedback already and we’re actively looking into some changes for our full launch of competitive mode this summer.
A couple items high on our list in the short term are making it easier to queue with your friends during your placement matches as well as better clarity for rank icons. Long term, we’re exploring adding a rank progression along with other features based on your feedback. More to come on that in future updates!
And finally, because Paul Chamberlain, VALORANT’s anti-cheat lead, enjoys talking about Vanguard, we let him add in another update this week on how the team’s listened to feedback:
We’ve heard feedback from players that it felt like Vanguard was being overbearing in its approach to security so we’re taking another approach. As of last week, Vanguard’s compatibility with vulnerable software packages has been expanded. While a small number of drivers are still blocked, from now on almost all of the software that Vanguard was previously blocking will work properly on Vanguard protected computers.
In the future we’re going to prefer non-invasive solutions to security issues where possible, finding ways for Vanguard, VALORANT and that one weird RGB driver to coexist. If we can’t find a way for them to all work together, we’ll disable the game rather than disable your other software. Only in the most egregious cases will we block any software from running on your computer.
We’re continuing to improve Vanguard and we need your feedback to make good decisions. So please let us know how you feel about our anti-cheat efforts and we’ll do our best to keep everyone (except cheaters) happy.
… We’re going to skip the known issues because… this whole post has been about them (we’re still actively working on client FPS drops!). See you next week! :)