Bottom line: The map veto shapes a CS2 series before it starts. Across more than 19,000 professional veto series in our data (as of 16 July 2026), Inferno is the most-banned map by raw count, Mirage is the most-picked, and Ancient is left as the decider more than any other. One caveat up front: raw counts are confounded by how long a map has been in the pool, so read them as what the long-standing maps do, not a pure fear ranking. This is a data study built on the EsportsOdds CS2 data API.
How the veto works
Before a best-of-three, the two teams take turns removing and selecting maps from the active-duty pool until three remain: each team picks one to play, and the last unbanned map becomes the decider. What a team bans reveals what it fears; what it picks reveals its comfort; the decider is the pool's neutral common ground.
The pool changes, so lead with the data
One thing to fix before reading any veto table: the pool is not static. On 6 July 2026, Valve replaced Overpass with Cache, Cache's first appearance in years. Because Cache only returned to CS2 in April 2026, it has almost no competitive history, a few hundred bans against the tens of thousands the established maps carry, so nobody has a meaningful CS2-era read on it. Every ranking below is built on the maps with a real sample, and will keep shifting as the pool turns over.
That churn is why raw ban and pick totals are a trap: a map playable for four years racks up more of everything than one added last month. Train, back since 2024, carries about 4,100 bans; the retired old Cache sits at 171. The counts below describe the long-standing maps, not a clean emotional ranking.
The most-banned maps
Bans are the clearest signal in the veto: a team spends a ban on the map it least wants to play. Inferno leads by a wide margin:
Inferno, Nuke and Mirage are the three most-removed maps. That looks paradoxical for Mirage, also the most picked map, but it makes sense: heavily-practised maps get banned because everyone is dangerous on them, so a team removes one to deny a strong opponent rather than out of dislike. Newer maps get banned less simply because they appear in fewer series.
The most-picked maps
Picks reveal comfort, the map a team actively wants to play. Here Mirage is king:
Mirage remains the most-picked map in Counter-Strike years into CS2, the best-understood map in the game, where structured teams trust their defaults. Ancient and Nuke follow. Topping the pick list while near the top of the ban list makes Mirage the pool's single most contested map: everyone wants it, so everyone must decide whether to fight for it or ban it.
The decider maps
The map neither team bans nor picks, left standing as the decider, is the pool's neutral ground: comfortable enough that teams don't rush to remove it, but not a signature pick either. Ancient plays that role most often:
Which maps get avoided the most
Raw ban counts favour the veterans, so a fairer question is which maps get banned relative to how often they are picked. That ban-to-pick ratio partially cancels the tenure bias, since both grow the longer a map is playable. Rank by it and a different picture appears:
Train (about 4.6×) and Vertigo (3.9×) top the list, but both sit on small, recently-rotated samples, so treat them as noisy. The cleaner signal is among the heavily-played maps: Anubis (about 3.2×) and Overpass (3.3×, now rotated out) are banned far more often than they are picked. At the other end, Ancient, Mirage and Dust2 sit near 1.9×, picked almost as readily as they are banned, the pool's comfortable common ground. Trust this ratio over the raw totals: it corrects a tempting error, since Anubis is not uniquely situational once you account for how often each map is actually chosen.
What the veto reveals
Put together, the lists sketch the shape of the meta:
- Feared and contested: Inferno and Mirage, banned constantly because everyone is good on them.
- Neutral ground: Ancient, the map most likely to decide a series when the picks cancel out, and among the most comfortable on the ban-to-pick measure.
- Genuinely avoided: Anubis, banned far more than it is picked even after adjusting for tenure, the map teams least want to be forced onto.
A team's veto tendencies are one of the most predictive pieces of pre-match information there is. Know that a side almost always bans Nuke and picks Mirage, and you already know a great deal about how the series will be shaped before the first round.
Getting veto data from the API
Every ban, pick and decider in our dataset is available through the EsportsOdds CS2 data API as structured JSON, the raw material for a veto-tendency model, a pre-match preview, or a map-pool trend tracker, with no bracket-scraping. Our methodology page documents how series and veto steps are recorded.
Across 19,000+ pro Counter-Strike series, Inferno is the most-banned and Mirage the most-picked and contested. But raw totals favour long-standing maps: on a ban-to-pick basis Anubis is the most genuinely avoided, while Ancient, Mirage and Dust2 are the most comfortable.
Frequently asked questions
What is the most-banned map in CS2?
Inferno, across more than 19,000 pro veto series in our data, ahead of Nuke and Mirage. Teams ban the maps everyone is dangerous on, though raw counts also favour maps that have been in the pool longest.
What is the most-picked map in CS2?
Mirage, still the most-selected map in professional Counter-Strike years into CS2, because it is the most-practised and best-understood map in the pool.
Which map is most avoided in CS2 vetoes?
On a ban-to-pick basis, which corrects for how long a map has been playable, Anubis is the most avoided of the heavily-played maps, banned about three times as often as it is picked. Train and Vertigo score higher but on small, recent samples.
Why is Mirage both heavily banned and heavily picked?
Because it is the most contested map: teams pick it when confident and ban it to deny a strong opponent. Sitting near the top of both lists makes it the pool's most fought-over map.
Where can I get CS2 veto data via an API?
Through the EsportsOdds CS2 data API: every ban, pick and decider per series as JSON. Reconstruct any team's veto tendencies straight from the rows.