- ReplayLab
- Features
CS2 positioning analysis: which of your deaths were positioning mistakes
Most deaths in CS2 are not lost on aim. They are lost a few seconds earlier — by standing where nobody can trade you, by peeking the same angle twice, or by being visible to two enemies at once. ReplayLab checks every death in your demo against four concrete positioning patterns and shows you the ones that match, with the evidence and a button that opens the exact moment in the replay.
The four positioning mistakes ReplayLab detects
Each detector is a fixed, measurable rule — not a guess about intent. A death can match several of them at once.
| Mistake | When a death is flagged |
|---|---|
| Isolation | You had living teammates, but the nearest one was more than 1,250 units away along the walkable path — about five seconds of running. |
| Multi-exposure | At the moment of death you were in the line of sight of at least two living enemies, with no smoke blocking either line. |
| Re-peek | A confirmed contact with the enemy broke for at least 125 ms and then resumed no more than 1.5 seconds before you died — you peeked the same fight again. |
| Untradeable position | Your death was not traded within 5 seconds, no teammate had contact or line of sight on your killer, and the nearest teammate needed more than 5 seconds to get there. |
Only kills by an enemy player are analysed. Deaths to the bomb, fall damage, suicides and team kills get no positioning label.
How positions are measured
The analysis uses the map itself, not just distances on the radar:
- Line of sight and field of view are checked against the map's collision geometry, including whether a smoke covered the line.
- Distance to teammates is measured along the navigation mesh — the walkable path — rather than in a straight line through walls.
- Contacts come from a timeline of who could see whom in the seconds before each death.
If the detailed geometry check is not available for a match, ReplayLab falls back to simpler data: a navigation graph for path distances and the game's own radar spotting for visibility. Episodes measured this way are marked as approximate, so you know which conclusions rest on less precise data.
What you see in the breakdown
The positioning results live in the Breakdown tab of each match:
- Highlights. Up to 20 mistake episodes and up to 8 strong moments, selected by code from the whole match, one card at a time.
- Tags and an explanation. Each episode shows which detectors fired, a short explanation and the key numbers — for example the distance to the nearest teammate.
- A radar schematic. A picture of the moment: players, where they were looking, lines of sight, smokes and fire.
- Watch the moment. One button opens the 2D replay at that second. From there you can switch to the 3D player-eye view to see the angle as the player saw it.
- Per round or per match. Browse the episodes of the current round or the full list for the match.
Around the positioning tags, the analysis adds context that often explains a death: whether you were reloading or your weapon was not ready, whether your footsteps gave you away, whether you were running while slowed by damage. Trades are analysed in stages — could a teammate trade you, did they try, did they succeed — so you can tell an unlucky death from a position nobody could support.
How to use it to improve your positioning
- Look for repeats, not single deaths. One isolated death is a moment; five in one match is a habit. Start with the label that shows up most often.
- Watch every flagged death in the replay. Check the five seconds before it: where were your teammates, and what information did you have? If the demo recorded comms, turn on voice chat and listen to the calls.
- Decide what you would change. For isolation, the fix is usually timing — waiting for a teammate or rotating together. For multi-exposure, it is choosing a spot where only one angle can see you. For re-peeks, it is repositioning instead of swinging the same angle again.
- Compare across matches. ReplayLab keeps metrics across your uploaded matches, so you can see whether a pattern is fading.
Limitations
- Rules, not judgement. The detectors flag measurable situations. A planned lurk or a deliberate late-round save can match “isolation” and still be the right call. Treat a tag as a question to check in the replay, not a verdict.
- Map versions. Geometry is the current version of each map. For old demos on maps that have changed since, line-of-sight checks can be off.
- Feedback. Every finding can be marked “correct”, “incorrect” or “disputed”; these ratings are used to measure the accuracy of the analysis.
FAQ
What positioning mistakes does ReplayLab detect?
Four: isolation (nearest teammate more than about 5 seconds away along the path), multi-exposure (visible to two or more enemies at death), re-peek (the same contact resumed within 1.5 seconds of death) and untradeable position (not traded within 5 seconds and no teammate able to reach or see the killer).
Does it use real line of sight or just the radar?
Where the detailed analysis is available, line of sight is checked against the map geometry, including smokes, and distances use the walkable navigation mesh. If it is not available, the analysis falls back to radar spotting and a navigation graph, and those episodes are marked as approximate.
Does every flagged death mean I played it wrong?
No. The detectors flag measurable situations. A planned lurk can match isolation and still be a good decision. Use the tag as a pointer to a moment worth reviewing in the replay.
Which deaths are analysed?
Only kills by an enemy player. Bomb deaths, fall damage, suicides and team kills get no positioning label.
Can I see the mistake in the replay?
Yes. Every episode has a button that opens the 2D replay at that moment, and you can switch to the 3D player-eye view from there.
Is positioning analysis free?
It is part of the standard demo analysis. Your first own demo is analysed free on the Free plan ($0); further demos require Pro.