Valorant Agent Pick Rates – Pro vs. Ranked

Pro meta vs. ranked: why the pick rates differ so much

Anyone looking at pro statistics for the first time is often surprised: in tournaments Omen appears in almost every team (73% presence at VCT 2026 Americas Stage 2), while in ranked he sits mid-pack at around 10.5% pick rate. The other way round, ranked is ruled by Chamber at 27.6% ahead of Jett (25.1%) and Clove (23.5%) – three agents played far less, or almost never, in pro play. The reason isn't different truths but two completely different playing environments. You can click through that exact contrast above, agent by agent and map by map – this page puts pro presence and ranked pick rate directly side by side.

In pro play every team coordinates five fixed roles: at least one controller for smokes, an initiator for info and flashes, a sentinel to lock down, and one or two duelists for the entry. A controller like Omen is therefore almost mandatory – which is why top agents reach presence values up to 100% on certain maps. Pick rate here is really a presence: the share of maps on which the agent was in the team. Because a pro team deliberately spreads its five slots across tasks, a tight, stable meta forms in which the same twelve to fifteen agents play almost every game.

In ranked (here Platinum+) every player freely picks their agent – usually their own main. Picks spread out far more, and self-sufficient agents with high solo value dominate: Reyna (no team needed), Chamber (strong self-utility and repositioning), Jett (escape potential and entry freedom). That's why even the most-picked agent sits at just under 28% instead of 70%+. Ranked is also an environment without fixed calls: many teams run two or three duelists and no real controller – a pattern that would be punished instantly in pro play.

The biggest pro-vs-ranked gaps in detail

Omen is the prime example: 73% presence in pro play, but only 10.5% in ranked. He is the most flexible controller in the game – his smokes re-position instantly, his blind and teleports give a coordinated team enormous control. Solo queue often lacks the communication to use that, so many reach for Clove instead (23.5% ranked, 0% pro): Clove can still smoke and resurrect after dying – perfect for uncoordinated rounds, but inferior to the clean, planned smokes of Omen and Viper in structured pro play.

Sova shows the same gap from the initiator side: 60% pro presence, but only 15.3% in ranked. Information is half the battle in pro play – a recon dart that reveals the site prevents blind pushes. In ranked info is undervalued because the team often doesn't act on the calls. Viper is the most extreme case: up to 100% presence on certain maps in pro, but effectively non-existent in ranked at 0.3% – her value only unlocks with team timing and post-plant coordination. Understanding these patterns reveals which agents are underrated in ranked: controllers and info initiators win games even when they are rarely picked.

Pick rate is not the same as strength

A high pick rate first of all only means: popular or meta-mandatory. It doesn't automatically mean the agent wins. A good example is Chamber: the most-picked ranked agent (27.6%), but with a win rate around 49.8% he's rather average. Conversely Clove has the highest win rate at Platinum+ (52.6%) and is also very popular – a rare combination that explains the S-tier status. Killjoy too (52.0% win rate at just 1.6% pick rate) and Cypher (51.6%) are quiet winners: rarely played, but above-average successful. That's why the ranked values always also show win rate and tier, so you can tell popularity and actual strength apart. Rule of thumb: an agent with a high win rate and low pick rate is often an underrated pick for climbing.

Pick rate per map: the pro meta map by map

In pro play agent choice depends heavily on the map – every map has its own sightlines, chokes and rotation paths. Use the map selector above to see the pick-rate-sorted ranking for each map instantly. Here is the full meta breakdown of all eight maps from VCT 2026 Stage 2, each with the reasoning behind why certain agents dominate:

Ascent is the classic two-site map where mid control decides the round. Omen and Sova are near-mandatory at 100% presence each: Omen splits mid and the sites with his smokes, Sova gathers the information that is everything on Ascent. Behind them come Phoenix (57%) and Neon (53%) as entry duelists, Cypher (50%) locking B-main and flank, with Chamber and Jett at 37% each. Ranked looks completely different – there Chamber, Jett and Reyna dominate while Sova sits at only ~15%.

Haven is the only three-site map and rewards information and rotation timing. Here Sova (91%) leads Neon (89%) and Phoenix (78%) – three sites force maximum recon and fast entry. Omen (74%) and Cypher (50%) provide smoke and flank cover, Chamber (43%) and Astra (26%) round it out. Haven is the map where info initiators reach their highest value.

Lotus is the second three-site map, with rotating doors and heavy post-plant play. The double controller stands out: Omen (100%) and Viper (85%) split the wide angles, joined by Fade (79%) for info, Raze (65%) as entry and – unusually – the sentinel Vyse (59%), whose traps seal the many entrances. Lotus shows how map-specific the pro meta is: Vyse is near-mandatory here, barely seen elsewhere.

Split is a tight, vertical map with many chokes – Viper is near-mandatory at 100% presence because her toxic wall splits the narrow corridors perfectly. Omen (94%) as second controller, Neon (75%) for fast executes, Fade (59%) and Phoenix (50%) plus Skye (41%) complete it. Split is a utility and close-quarters map: pure aim duels lose to well-timed abilities.

Sunset is a compact two-site map with a large mid. Omen (100%) and Neon (91%) dominate, alongside Sova (57%), Cypher (54%) and Chamber (46%). Phoenix (39%), Fade (34%) and – typically for this map – Sage (27%), whose wall blocks the tight mid entrances, round it out. Sunset rewards fast mid control with flash entry.

Breeze is a large, open map with long sightlines – the double controller shines here most. Sova (100%) is set, Harbor (73%) jumps up as a second controller because his water walls split the long lines, joined by Neon (67%) and Viper (63%). On no other map is Harbor this valuable. Cypher and Phoenix (40% each) plus Jett (37%) complete it.

Abyss is the oddity: a map without kill barriers at the edges (fall death!), where positioning is vital. Sova (100%) and Harbor (64%) lead, followed by a broad field of Phoenix, Astra and Waylay (50% each). The many open angles demand plenty of info and smoke – pure duelists struggle here.

Summit is the new three-lane map with reactive walls, where mid control is everything. Neon (81%) is the clear top pick for fast lane control, with Fade (69%) for info and Omen (67%) for smoke. Chamber (39%), Cypher (36%) plus KAY/O and Harbor (28% each) complete it. As the newest map, Summit's meta is still shifting the most.

Ranked pick rate, by contrast, barely varies by map: a Reyna main plays Reyna on every map. That's why we show the ranked number across ranks as a stable comparison next to the map-specific pro presence – so you can see at a glance where solo queue and the pro meta diverge most.

The four roles compared

Duelists rule ranked (Jett 25.1%, Reyna 19.9%, Neon 11.1%) but are usually limited to one or two slots in pro. The big exception is Neon: at 72% pro presence she is by far the most-played duelist in tournaments, because her speed and stun fit coordinated executes perfectly. Jett, by contrast, drops to 15% in pro, Reyna all the way to 0% – she lacks team value.

Initiators are valued far more in pro than in ranked: Sova (60% pro vs. 15.3% ranked) and Fade (31% vs. 8.8%) deliver the information without which no site hit works. In ranked this role is badly neglected – one reason a single good initiator can swing so much there.

Controllers show the biggest gap of all. Omen (73% pro / 10.5% ranked), Viper (up to 100% on some maps / 0.3% ranked) and Harbor (19% / 0.5%) are indispensable in pro, fringe picks in solo queue. Only Clove flips the picture (0% pro / 23.5% ranked). Consistently playing a controller in ranked does your team one of the biggest favours possible.

Sentinels are mixed: Cypher (37% pro / 6.0% ranked) and Killjoy are info and lockdown anchors in pro, while Chamber is the only sentinel high in both worlds (31% pro / 27.6% ranked) – he combines sentinel utility with duelist-like aim potential.

How to use pick rates to climb

Don't blindly copy pro comps – without communication they only half work. Take three lessons instead: first, make sure your team has at least one controller and one initiator; those are exactly the roles ranked lacks most and they win rounds. Second, check the map before locking: on open maps like Breeze and Abyss a second controller (Harbor/Viper) pays off, on info maps like Haven and Ascent a recon initiator (Sova/Fade). Third, watch the win rate, not just the pick rate: agents like Killjoy, Cypher or Clove win above average even though (or because) they are picked less. One disciplined utility player lifts a solo-queue round more than a fourth duelist.

Underrated agents & hidden gems

The most interesting picks often aren't at the very top of the pick rate, but of the win rate. Killjoy is played by only 1.6% at Platinum+, yet wins 52.0% of her rounds – her traps and lockdown ult are hugely valuable when nobody else secures the site. Similarly Cypher (51.6% win rate, 6.0% pick rate) and Sage (50.8% / 3.4%), whose wall and heal break whole executes. Map-specifically the upside is even clearer: Vyse is near-mandatory on Lotus at 59% pro presence, yet barely shows in the general ranked list – a clear sign she is underrated there. Playing such gems early banks RR before the crowd catches on.

Conversely, the »safe« picks deserve a critical eye. Reyna and Jett are extremely popular but deliver only average win rates (50.2% and 50.1%) – they reward individual aim but add little to the team. In a round without a controller, the fourth duelist is rarely the answer. This is exactly where the biggest free improvement lies for most ranked players: filling a role nobody else takes.

How the meta shifts over a season

Pick rates aren't a law of nature – they move with every patch. Riot ships an update roughly every two weeks that buffs or nerfs agents, tweaks weapons or changes the map rotation. A buff can lift a fringe agent into the meta overnight, a nerf can drop a staple out of comps. New maps shift everything too: Summit only entered the competitive pool with patch 13.00, and its meta is still moving the most. Big tournaments act as a catalyst: when a top team shows a new comp, other pros copy it first and ranked lobbies soon after. That's why it's worth checking this page regularly – we pull the numbers each patch so you always see the current state instead of outdated rules of thumb.

Methodology & sources

The pro numbers come from VLR.gg and cover VCT 2026 Americas Stage 2 (150 maps played, Jul 16 – Sep 7, 2026). Pick rate here means presence: the share of maps on which an agent was played by a team (0–100%). The ranked numbers come from valking.gg for rank Platinum+ and give pick rate, win rate and tier. Both datasets are on patch 13.05 (as of Sep 11, 2026). Because the meta shifts with every patch, we pull the numbers each patch and update the date. Champions 2026 (Sep 24 – Oct 18, Shanghai) had not started at the time of collection – once the tournament runs, the new international data will be included.

Valorant agent pick rates: the facts at a glance

• Most-picked agent in pro play (VCT 2026 Americas Stage 2): Omen at 73% presence, followed by Neon (72%) and Sova (60%).
• Most-picked agent in ranked (Platinum+, patch 13.05): Chamber at 27.6% pick rate, ahead of Jett (25.1%) and Clove (23.5%).
• Highest win rate in ranked: Clove at 52.6% (the only S-tier), ahead of Killjoy (52.0%) and Cypher (51.6%).
• Biggest pro↔ranked gap: Omen (73% pro vs. 10.5% ranked) plus Reyna and Clove, who are top 3 in ranked but barely appear in pro play.
• Pick rate measures popularity/meta necessity, not automatically strength – that's why win rate is always shown alongside.

Sources: VLR.gg (VCT 2026 Americas Stage 2) & valking.gg (Platinum+) · As of: September 11, 2026 · Patch 13.05

Why is the pro pick rate so much higher than the ranked pick rate?

Because the two numbers are measured differently. In a pro match each team fills five fixed roles – controllers like Omen are almost mandatory, so top agents reach presence values up to 100% on a map. In ranked every player freely picks their agent; picks spread out far more, which is why even the most-played agent (Chamber) reaches only around 28%. Both values are shown with their source.

Where do the numbers come from?

The pro numbers come from VLR.gg (VCT 2026 Americas Stage 2, 150 maps played), the ranked numbers from valking.gg (rank Platinum+). Both are on patch 13.05 (as of Sep 11, 2026) and are refreshed each patch. No estimated or invented values.

Can I see the pick rate per map?

Yes. Use the map selector to switch between "Overall" and single maps (Ascent, Haven, Lotus, Split, Sunset, Breeze, Abyss, Summit). The pro bars then show presence on exactly that map. Ranked pick rate is shown across ranks because it barely shifts by map.

Why are Reyna and Clove strong in ranked but rarely seen in pro play?

Reyna and Clove are solo-queue agents: Reyna rewards individual aim with no team dependency, Clove as a controller can still make an impact after dying. Coordinated pro teams instead favour structured controllers (Omen, Viper) and info initiators – so Reyna and Clove almost never appear there.

Does a high pick rate mean the agent is the best?

No. Pick rate shows popularity or meta necessity, not automatically win rate. An agent can be picked a lot and still be only average (e.g. Chamber: very high pick rate, but win rate around 50%). That's why the ranked values also show win rate and tier.

How often is the data updated?

With every major patch. Since Riot patches roughly every two weeks and the meta shifts, we pull the numbers accordingly and update the date shown at the top of the page.