What we learned from VALORANT agent pick rates during closed beta

The VALORANT closed beta has come and gone, leaving with it no small amount of hype and anticipation over its upcoming June 2 release. Players have been defecting to VALORANT before it officially has an esports scene.

While these defections are arguably premature, it’s difficult to ignore the impact that VALORANT has already had on the esports landscape. With the expectation that VALORANT is aiming to have a large esports footprint, there have already been numerous tournaments throughout the closed beta from which we can draw rudimentary analysis. It’s a rare look at an esport in its infancy prior to its official launch.

A month ago, I took a look at how the agent selection meta evolved over ESPN Esports’ own VALORANT invitational. Here is a cursory look at agent selection in subsequent tournaments, how it has evolved since the ESPN invitational and some lingering thoughts on what VALORANT’s future will hold regarding agents and abilities.

Read more: What we thought of VALORANT’s beta | Who pioneered each VALORANT agent

A note about data collection

Due to the unavailability of entire VOD libraries and a lack of API scraping — the former will hopefully come with time and the latter should begin upon the game’s release — the data collected from each tournament is only a small slice of the games played during those tournaments. Data is from the T1 Invitational semifinals and grand finals, the Fnatic Proving Grounds open semifinal and grand finals, the Mandatory Cup third-place match and grand finals, and the Solary Cup grand finals.

The data collected is but a small sample of what was available from these four tournaments, yet I do think it is a microcosm for agent selection and team composition building in VALORANT.

VALORANT agent pick-rate data

Despite the small sample size, most of the agent selection data across all tournaments was consistent, with small deviations depending on players, teams and maps.

It should come as no surprise to anyone that Sage and Cypher were picked 100% of the time by all teams regardless of map. Cypher is VALORANT’s most versatile vision-giving agent, and Sage is (thus far, pending subsequent agent releases like Reyna) the game’s only healer. Additionally, she provides utility that no other agent comes close to having for their team with her abilities, including the wall, slows and resurrection ultimate.

In the first tournament I analyzed, the ESPN Esports Invitational, Sage was picked 94% of the time (all but two times), and both were losses for the Sage-less team. I’m certain that if I had catalogued every single match from each of these four tournaments there would be a few outlier games where one team chose not to pick up Sage, but these would be few and far between. During the semifinals and finals, Sage was, and remains, a must have.

Vision is another key component in the game, and while Cypher is the primary vision agent, it’s also interesting to take a look at how many teams opted for both Cypher and Sova for extra scouting abilities. Sova was picked 39% of the time in the games I collected data from, and due to Cypher’s 100% pick rate, this 39% pick rate also applies to the Cypher-Sova combination. A large percentage of those picks came from the T1 Invitational matches between Gen.G and Team Brax, both of which selected Sova and Cypher in their grand finals matches.

In the ESPN Invitational, Brimstone (84% pick rate) was chosen more than Breach (50% pick rate), but as tournaments and teams have evolved, Breach has become a far more popular pick. In my data from the T1 Invitational and Mandatory Cup, Breach was 100% picked, alongside Cypher and Sage.

ESPN Daily newsletter: Sign up now!

With all of the data from these four tournaments combined, Breach (89% pick rate) was selected more often than Brimstone (69% pick rate), although the two are both popular, especially together. While teams at the ESPN Invitational were still figuring things out, the shift to Breach priority and continued focus on the Breach-Brimstone combination has made for a meta staple. The smokes and flashbangs that the two provide are invaluable tools in VALORANT and are a cut above similar abilities in the game from Phoenix, Viper, Omen and Jett.

Speaking of Phoenix (14% pick rate) and Jett (11% pick rate), it might come as a surprise that neither of these two faces of the game was the most popular entry fragger. That award goes to Raze, who was picked 57% of the time. In the Fnatic Proving Grounds Open semi between Royals and 2G4L and finals between 2G4L and Prodigy, Raze had a 100% pick rate. The initial Raze nerfs took effect mid-tournament during the ESPN Esports Invitational, and she was not picked by teams in the tournament after her abilities were tuned down. Yet in subsequent tournaments she has occupied a strong identity as one of the game’s most effective entry fraggers.

A study in abilities

Now that I’ve bored you to death with pick-rate statistics, it’s time to talk a bit about why these pick rates make sense, and how they tie into abilities generally.

Due to the nature of abilities in VALORANT, certain agents just aren’t as useful as others. Consider Omen for a moment and what his kit is designed to do. Although he, like Jett, gives you a certain amount of mobility and verticality over someone like Brimstone, why would you play Omen over Brimstone when Brimstone has similar smoke bombs and offers more zone control over sites with his ultimate? The comparison between Breach and Phoenix (who also has a flashbang) is a bit more specious. Since Phoenix is more entry and Breach isn’t, there’s no reason that you would pick a Phoenix to have a flashbang over a Breach, given Breach’s other abilities, which can allow entire teams to follow up and take over a point.

Ultimately, agent abilities are what differentiate VALORANT from Counter-Strike and other games. As agent metas evolve, I’ll be keeping an eye on agent selection and how Riot handles the high pick rates for the likes of Sage and Cypher in competitive compositions.