What is ADR in CS2? It stands for average damage per round: the damage you deal to enemies, divided by the number of rounds played. ADR sits next to K/D ratio, KAST and the HLTV Rating on scoreboards and stats pages, and each one measures something different. This guide explains all four, with HLTV's own account of how its rating works.
What is ADR in CS2?
ADR measures damage output. HLTV added it to its stats pages in early 2016, and Liquipedia's glossary notes that it counts only effectively dealt damage, capped at 100 per opponent. A headshot that would do more than 100 damage to a full-health enemy still counts as 100, and finishing off an enemy on 20 health adds 20.
CS2 tracks it too: the game's text files include an ADR column for the scoreboard, with the tooltip "Average Damage per Round", alongside KDR (kill-death ratio), HS% (headshot percentage), utility damage and enemies flashed, plus a button that cycles through the stat columns.
ADR credits damage that didn't become your kill. HLTV has said that adding damage to its rating in 2017 rewarded supportive players who dealt lots of damage, got assists or had their deaths traded without finishing kills.
K/D ratio and its limits
K/D ratio is your kills divided by your deaths, and was the usual way to judge a performance before ratings. When HLTV introduced its first rating in April 2010, it called K/D "a solid indicator in around 75% of situations" but listed two problems. It favours players who play fewer rounds and so die less, and it has no upper limit, which makes single matches hard to compare.
In HLTV's example from one 18-round match, one player went 22-4 (a K/D of 5.5) and another 20-9 (2.2). K/D makes the first look two and a half times better, while HLTV's rating gave them 1.87 and 1.56.
Assists don't count in K/D at all. CS2 awards an assist for less damage than CS:GO did, and with Rating 3.0 HLTV went back to counting an assist only at 40 damage, the old CS:GO value, "for all purposes".
What is KAST?
KAST stands for Kill, Assist, Survived, Traded. It is the percentage of rounds in which you did at least one of those things: got a kill, got an assist, stayed alive, or died and had a teammate kill your killer straight away, which is called a trade. HLTV added KAST in May 2017 and describes it as round-to-round consistency. It picks out players who rarely post big numbers but contribute something in most rounds.
Since Rating 2.1 in October 2024, the KAST that goes into HLTV's rating only counts survival if your team won the round, so saving your weapon in a lost round no longer earns the point. HLTV noted that the KAST percentage shown on its pages still includes saves.
How the HLTV Rating is built
The HLTV Rating is a single number that compares a player's stats with expected values, with 1.00 as the average. It has gone through four versions:
| Version | Introduced | What goes in |
|---|---|---|
| 1.0 | April 2010 | Kills, survival and multi-kill rounds, per round |
| 2.0 | June 2017 | Kill, Survival, KAST, Impact and Damage ratings, per side |
| 2.1 | October 2024 | 2.0 retuned for CS2, with less reward for saving |
| 3.0 | August 2025 | Eco-adjusted Kills, Damage, Survival, KAST and Multi-Kills, plus Round Swing |
Rating 1.0 was simple enough to publish. Each stat was divided by the average for Counter-Strike as a whole (0.679 kills per round, for example), survival counted at 0.7, and the total was divided by 2.7 so that an average game scored 1.0. HLTV said it ran from 0 to 3 but rarely went above 2.
Rating 2.0 kept the principle but worked out expected values separately for each side and measured how many standard deviations a player sat above or below them. It used ten components, five per side, with Impact covering multi-kills, opening kills and clutches won. HLTV stopped publishing the exact formula at this point, and in 2024 it said the formula remained private.
Rating 2.1 came a year into CS2. Its averages were rebuilt from CS2 data to fit MR12 matches, first to 13 rounds as the game modes guide explains, and the lower assist threshold. HLTV said the average rating had crept up to about 1.06 before the update brought it back to 1.00 over an event. The five sub-ratings got equal weight, and survival in lost rounds was rewarded less.
Rating 3.0, launched on August 20, 2025, was the biggest change since 2017. Two ideas drive it:
- Eco-adjustment. Each kill is weighed by how hard the duel was, based on both players' armour and best weapon. On the T side, a rifle-against-rifle kill, won 48% of the time, is worth about 1.10 kill points, while killing a player with a starter pistol, won 75% of the time, earns about 0.54. That cuts the value of kills against a team on an eco, a round where it skips buying to save money, as the economy guide describes.
- Round Swing. Every kill is scored by how much it changed the team's chance of winning the round, given the economy, the bomb, the players alive and the map. HLTV put an opening kill at around 20 percentage points and a won one-on-one clutch at around 50. Credit is shared out by who did the final damage, each player's share of the damage, flash assists and trades.
HLTV's example: in one match EliGE's 133 ADR and 23 kills gave him a 2.06 under Rating 2.1. Nine of those kills came against full ecos and three more against half buys, so Rating 3.0 rated the same map 1.40.
HLTV has tuned 3.0 twice since. On October 29, 2025 it gave kills more weight and Round Swing less, and added eco-adjusted K-D, ADR and KAST figures to its scoreboards. On October 1, 2026 it updated its win probabilities with CS2 matches up to September 29, added map-specific data for Cache, and stopped kills from adding Round Swing when fewer than 10 seconds remain on the bomb, or 5 if the CTs have a kit, unless a CT is already defusing. Across maps played from July to September 2026, single-map ratings moved by 0.021 on average.
Reading the numbers
Each stat answers a different question. ADR shows how much damage you put out, KAST how often you contributed to a round, K/D how your fights went, and HLTV's rating tries to fold all of it, plus context, into one figure. That context is why fewer kills can still mean a higher rating: HLTV wrote that a player with an 8-8 K-D can have more impact than one with a 12-4.
HLTV calculates its rating for the matches in its database, with averages drawn from LAN games, and it underpins its event MVP awards and Top 20 players of the year. It is unrelated to Valve's CS Rating, the number Premier gives you in matchmaking, which the Premier ranks guide covers.
Sources
- What is that rating thing in stats? — HLTV
- Introducing Rating 2.0 — HLTV
- HLTV introduces Rating 2.1 — Dust2.in
- Introducing Rating 3.0 — HLTV
- Rating 3.0 adjustments go live — HLTV
- Rating 3.0 receives small accuracy improvements — HLTV
- Definitions — Liquipedia Counter-Strike Wiki
- csgo_english.txt (CS2 text strings) — GameTracking-CS2 on GitHub
