ADR (Average Damage per Round) is a metric showing the average damage dealt per round. The calculation is simple: all damage a player deals to enemy players over the course of a match is divided by the number of rounds played. The metric is available in FACEIT statistics (a platform for ranked matches with an ELO system), in the CS2 client statistics, and in most match trackers.
Unlike K/D (kill-to-death ratio), ADR shows not just "killed or didn't kill," but the player's overall contribution to the round: wounding an opponent, helping a teammate finish them off, dealing damage to walls with grenades. A player can end a round without a frag but with high damage — and this is also reflected in ADR.
Rough benchmarks for the metric are as follows: an ADR around 70–80 is considered normal for an average player, figures of 90 and above indicate strong performance, and values around 100 and above sustained across many matches are typically seen among high-level players.
This is exactly why ADR often comes up in disputes about fair play. As in the case of a player whose consistent stats — a rating of around 1.28, ADR near 100, K/D above one with almost no variance over many matches — raised suspicions within the FACEIT community. The logic behind such accusations is that this kind of consistency over a long stretch of matches is atypical even for strong players, whose numbers usually fluctuate from match to match depending on form, opponents, and maps.
That said, a high ADR alone does not prove cheating. The metric depends on the player's role: for example, an entry fragger (a player who is first to make contact with the enemy) or a player in a position with heavy utility (grenade) usage can consistently rack up damage through frequent gunfights, even without winning rounds. ADR also tends to rise on maps and in lineups with a lot of trading (exchanging frags between teams) and prolonged firefights.
As a metric, ADR is most valuable in combination with other indicators — K/D, platform rating, headshot percentage. On its own, none of them gives a complete picture of a player's skill or fairness, which is why analysts and the community usually look at the statistics as a whole rather than at a single number.