Upload a demo, choose three to five decisive rounds, watch the setup from player POV, then switch to free or top-down camera to check routes, spacing, sightlines and utility. Ten maps are supported for full demo review: Ancient, Anubis, Cache, Dust II, Inferno, Mirage, Nuke, Overpass, Train and Vertigo. On Nuke, Train and Vertigo, confirm the radar level before judging a route or a teammate's proximity.
Which CS2 maps can I review demos on?
The supported pool comes from the same map configuration the replay loader and radar use, so there is exactly one list and it cannot drift between features. A supported map has the extracted Source 2 geometry, the environment and lighting configuration, and the radar calibration needed to place recorded players and events in a reconstructed world. This is the current pool:
| Map | Demo map name | Review consideration |
|---|---|---|
| Ancient | de_ancient | Compare early utility with the space it actually creates before the first fight. |
| Anubis | de_anubis | Review route timing and whether teammates can support contact through connecting lanes. |
| Cache | de_cache | Check the setup before site contact and whether pressure elsewhere changes the defence. |
| Dust II | de_dust2 | Use free camera to separate information gained across long sightlines from hindsight. |
| Inferno | de_inferno | Track utility and teammate spacing through narrow approaches before the execute. |
| Mirage | de_mirage | Compare mid control, connector support and site pressure on the same timeline. |
| Nuke | de_nuke | Two-level radar: confirm whether players are on the upper or lower floor. |
| Overpass | de_overpass | Start the replay early enough to see rotation routes and when defenders lose information. |
| Train | de_train | Two-level radar: account for lower areas and use 3D camera around occluding trains. |
| Vertigo | de_vertigo | Two-level radar: verify elevation before interpreting nearby blips or rotations. |
This list describes viewer support, not Valve's current competitive map pool. The two can differ as maps enter or leave matchmaking. A demo on a name outside the pool is rejected consistently rather than rendered with incomplete geometry, because a half-built map produces confident conclusions about walls that are not there.
A map-review workflow that produces an action
- Choose one question. Ask about a specific site hold, route, rotation, execute or retake. "Why did our B defence fail?" beats "what went wrong?"
- Shortlist rounds. Use the rounds and events tabs to find the first breakdown, a lost advantage, and one successful comparison round on the same side.
- Watch the setup in POV. Start before contact and state only the information that player could have had.
- Rewatch in 3D. Use free, follow or top-down camera to inspect spacing, routes, sightlines and utility across the whole map.
- Check the sightline on foot. If your conclusion depends on what someone could see, press B and walk to that spot at player height rather than trusting a floating camera.
- Mark cause before outcome. The useful moment is usually an early gap, a late flash or an unsupported rotation, not the final death.
- Write one testable adjustment. Name the trigger, the position and the action you will try next time on that map.
For the controls behind these steps see the CS2 demo viewer controls reference. If you still need the file, start with how to find and download a CS2 demo, then upload it for analysis.
How do you review multi-level CS2 maps?
A radar is a flat projection of a three-dimensional map, so two players can appear adjacent while standing on different floors with no route between them. This is the single most common source of wrong conclusions in map review, and it produces confident nonsense: "he was right there, why didn't he trade?"
Nuke, Train and Vertigo each carry a separate lower-level radar configuration, with a height threshold that decides which level a player is drawn on. Use it for orientation, then confirm elevation, walls and reachable routes in the 3D scene before you judge anything.
- Nuke: follow the transitions between upper and lower. Pause before a rotation and identify the moment the destination became knowable, not the moment it became obvious to you.
- Train: inspect the lower passages, then move free camera around the trains before deciding support was available. Train is full of occluders that a top-down view flattens away.
- Vertigo: verify the floor first, every time. A short 2D distance is not proof that two players shared a sightline or a trade path.
Elevation matters on every map — Overpass and Mirage both punish flat thinking — but the explicit two-level radar makes it unavoidable on these three. Never infer travel time or visibility from radar distance alone.
Why does real map geometry matter for review?
Many demo tools draw players as dots on a radar image. That is enough to see rotations and nothing else. A reconstruction built from the map's actual extracted geometry lets you ask questions a 2D projection cannot answer: was that boost reachable, did the smoke actually cover the doorway or just the radar line between two points, was the head visible over the box from where the enemy stood, and how much of the lane was still open when the molotov burned out.
Lighting is part of this too. Each map carries its own environment configuration and baked lighting, so a dark corner in the replay is dark in roughly the way it was dark in the match. That is worth knowing before you write down that a player "obviously" should have spotted someone.
Useful review questions for each map
A map guide should help you interrogate evidence, not hand you a generic verdict. These prompts work with the recorded positions, events and utility in any demo:
- Ancient and Inferno: What utility was spent to contest a narrow area, and what space remained when it faded?
- Anubis and Mirage: Could the player making first contact be supported through the available connecting route?
- Cache and Dust II: What information justified the rotation, and which sightline or area was given up to make it?
- Nuke and Vertigo: Did the team correctly identify the floor and the practical route before committing to move?
- Overpass: When did each side gain or lose information across the longer rotation?
- Train: Was the apparent teammate proximity real once trains, elevation and reachable paths are considered?
Compare a failed round with a similar successful one on the same map and side. That controls for map context and makes the remaining difference — timing, positioning, equipment, utility order — describable without overclaiming.
Opening a map without a demo
Sometimes the question is about the map, not the match: is this boost reachable, does this smoke line have a clear path, how long is that rotation on foot. Standalone maps open the same geometry without loading a demo, and walk mode moves you through it at player height so you can check the answer directly. Both are PRO tools. Utility practice covers the throwing side, with the CS2 throw and physics behaviour ported rather than approximated.
What map review can and cannot prove
The viewer reconstructs recorded game state, so you can inspect positions, health, equipment, events, grenades and match context wherever the demo contains them. It can show that players were separated, that utility landed in a recorded place, that a route was taken. It cannot prove attention, intention, communications or why a player chose what they chose.
Player POV is a reconstruction, not a recording of the original screen. Keep the language careful: "the teammate was not on a reachable trade path when contact began" is supported by replay evidence; "the teammate refused to trade" assigns a motive that no demo records. See what a CS2 demo analyzer can and cannot tell you and how we handle replay evidence for the full boundary.
Frequently asked questions
Which CS2 maps can I review demos on?
Ten maps have the extracted geometry, lighting and radar calibration a reconstruction needs: Ancient, Anubis, Cache, Dust II, Inferno, Mirage, Nuke, Overpass, Train and Vertigo. A demo on any other map is rejected rather than displayed with missing geometry.
Why do two players look close on the radar but never trade?
A radar is a flat projection of a three-dimensional map, so two players can appear adjacent while standing on different floors with no route between them. Nuke, Train and Vertigo each carry a separate lower-level radar for exactly this reason, and elevation should be confirmed in the 3D scene before you judge a trade.
Can I open a CS2 map without loading a demo?
Yes. Standalone maps let you inspect geometry, routes, angles and lighting on their own, and walk mode moves you through the map at player height so you can check a sightline yourself. Standalone maps and walk mode are PRO tools.
Does the supported map list match the CS2 competitive pool?
Not necessarily. This list describes which maps the viewer can reconstruct, not Valve's current active duty pool. The two can differ whenever a map enters or leaves matchmaking.