Issue 3: AK-47 takes 29% of kills, 1v2s won 21%
227 matches, 4,432 rounds and 31,511 kills: the AK-47 took 29% of kills, 1v2s were won 21% of the time, CT sides won 51% of rounds, 41% of kills were headshots and 33 aces were scored.
01Headline numbers
Issue 3 covers every Premier match played 12 – 19 Oct 2026 (UTC), up to its release on 19 Oct 2026.
- MATCHES
- 90
- ROUNDS
- 1,780
- KILLS
- 12,755
- HEADSHOT RATE
- 41%of 12,755 kills
- CT ROUND WINS
- 49%of 1,780 rounds
- T ROUND WINS
- 51%of 1,780 rounds
- CLUTCHES WON
- 18%of 2,142 clutches
- ACES SEEN
- 111 in 162 rounds
02Clutch odds
How often the last player alive on a side still wins the round, by how many enemies were alive when it started.
| Clutch | Overall | CT | T |
|---|---|---|---|
| 1v1 | 61%out of n=372 | 51%out of n=201 | 73%out of n=171 |
| 1v2 | 20%out of n=507 | 16%out of n=251 | 24%out of n=256 |
| 1v3 | 7.0%out of n=568 | 2.5%out of n=278 | 11%out of n=290 |
| 1v4 | 2.5%out of n=473 | 1.5%out of n=206 | 3.4%out of n=267 |
| 1v5 | 0.5%out of n=222 | 0.0%out of n=106 | 0.9%out of n=116 |
n = clutch situations; – = fewer than 30, too few for a number.
A clutch starts when a player becomes the last one alive on their side; N is how many enemies are alive at that moment, and 1v5 covers five or more. It counts as won when their side wins the round, by kills or by the bomb. Each side clutches at most once per round, so a 1v3 that ends as a 1v1 counts as a 1v3 for the clutcher and as a 1v1 for the last opponent standing.
03Weapons
The weapons with the most kills: each one's share of all kills, how many of its kills were headshots and the average distance between killer and victim.
| Weapon | Kill share | Headshots | Avg distance |
|---|---|---|---|
| AK-473,888 kills | 30% | 47% | 18.1 m |
| M4A1-S1,854 kills | 15% | 32% | 18.4 m |
| AWP1,576 kills | 12% | 12% | 27.9 m |
| M4A4719 kills | 5.6% | 39% | 17.3 m |
| USP-S711 kills | 5.6% | 70% | 18.0 m |
| Glock-18589 kills | 4.6% | 59% | 16.0 m |
| Desert Eagle426 kills | 3.3% | 66% | 19.9 m |
| Galil AR379 kills | 3.0% | 37% | 16.7 m |
| MP7344 kills | 2.7% | 30% | 13.9 m |
| MAC-10316 kills | 2.5% | 45% | 12.2 m |
| MP9288 kills | 2.3% | 56% | 12.3 m |
| SSG 08207 kills | 1.6% | 59% | 33.2 m |
| FAMAS184 kills | 1.4% | 28% | 17.4 m |
| XM1014139 kills | 1.1% | 40% | 11.1 m |
| P90119 kills | 0.9% | 35% | 12.1 m |
– = fewer than 100 kills with that weapon, too few for its headshots and distance.
04Aces and multi-kills
Counted per player per round: one player killing exactly three enemies in a round is one 3K, and an ace is all five. Two players who each get a 2K in the same round count as two.
| Kills | Per 100 rounds | Per match |
|---|---|---|
| 2K2,292 seen | 128.8more than one a round | 25.47more than one a match |
| 3K734 seen | 41.2one every 2.4 rounds | 8.16more than one a match |
| 4K147 seen | 8.26one every 12 rounds | 1.63more than one a match |
| Ace11 seen | 0.62one every 162 rounds | 0.12one every 8.2 matches |
05Knife and Zeus kills
- KNIFE KILLS
- 0.3%1 in 345 kills
- ZEUS KILLS
- <0.1%1 in 1,160 kills
Rare kills
| Kill | Share of kills |
|---|---|
| Noscope (sniper rifle, unscoped) | 0.3%1 in 290 kills |
| Wallbang (through a wall or object) | 2.6%1 in 38 kills |
| Through smoke | 5.0%1 in 20 kills |
| While blind (flashed) | 0.4%1 in 283 kills |
| Airborne (killer in the air) | 0.2%1 in 456 kills |
06Side win rates
The share of rounds each side won, and how often the bomb was planted (out of the rounds where we know whether it was).
- CT ROUND WINS
- 49%of 1,780 rounds
- T ROUND WINS
- 51%of 1,780 rounds
- BOMB PLANTED
- 53%of 1,780 rounds
07Compared with Issue 2
The same numbers in Issue 2 and in this issue. Rates change in percentage points (pts); aces in aces per 100 rounds. Metrics either issue has too few matches for are left out.
| Metric | Issue 2 | Issue 3 | Change | Significant? |
|---|---|---|---|---|
| Matches | 129 | 90 | −39 | Not tested |
| Headshot rateSample counts KILLS | 40.9%n=17,982 | 41.2%n=12,755 | +0.3 pts | No |
| CT round win rateSample counts ROUNDS | 54.7%n=2,523 | 49.3%n=1,780 | −5.4 pts | Yes |
| Clutch win rate (all 1vN)Sample counts CLUTCHES | 17.3%n=3,024 | 17.7%n=2,142 | +0.4 pts | No |
| Aces per 100 roundsSample counts ROUNDS | 1.11n=2,523 | 0.62n=1,780 | −0.49 | No |
| AK-47 kill shareSample counts KILLS | 29.6%n=17,982 | 30.5%n=12,755 | +0.9 pts | No |
| M4A1-S kill shareSample counts KILLS | 14.6%n=17,982 | 14.5%n=12,755 | −0.1 pts | No |
| AWP kill shareSample counts KILLS | 11.6%n=17,982 | 12.4%n=12,755 | +0.8 pts | Yes |
| M4A4 kill shareSample counts KILLS | 6.8%n=17,982 | 5.6%n=12,755 | −1.2 pts | Yes |
| USP-S kill shareSample counts KILLS | 5.5%n=17,982 | 5.6%n=12,755 | +0.1 pts | No |
| Glock-18 kill shareSample counts KILLS | 4.9%n=17,982 | 4.6%n=12,755 | −0.3 pts | No |
Significant = bigger than chance alone would likely produce with samples this size (two-proportion test, 95%). “No” means the change is within normal week-to-week noise.
08Methodology
Issue 3: 90 matches, 1,780 rounds and 12,755 kills, played 12 – 19 Oct 2026 (UTC). The default view counts Premier and Competitive matches together; the switch at the top shows Premier, Competitive or Wingman on its own. Every player in each demo counts, and only totals are published: no names, no profiles.
A weekly issue covers every match played since the issue before it (Issue 1: since launch), up to its release: every Monday at 06:00 UTC, or earlier when we close a week by hand. An issue doesn't change after its release: matches parsed later for its period stay out of it (they still count in the all-time numbers). Times are UTC.
- Left out: other game modes, warmup, rounds that never finished and rounds ended by a surrender vote. Valve matchmaking has no knife rounds.
- A kill is a death dealt by an enemy. Team kills, suicides and deaths to the world (a fall, for example) count as deaths only.
- Live numbers refresh hourly; new demos join as they're parsed. Positions and bomb plants come from demos parsed since gatherer 0.3.4; older matches were backfilled from their stored kills and rounds.
A number needs a minimum sample before it shows; below it, the cell shows a dash and the section says so once. The minimums are the same for an issue as for all time, so a quiet week shows more dashes.
- Clutch odds (each cell)
- 30 situations
- Side win rates, bomb plants
- 300 rounds
- Multi-kills and aces
- 1,000 rounds
- Kill shares, headshot rate
- 2,000 kills
- A weapon's headshots and distance
- 100 kills with it
- Heatmaps
- 300 positioned kills
- A map in the by-map tables
- 20 matches
09Cite this issue
Quoting these numbers? Please credit killcam.gg with a link to this issue: Issue 3 is released, so its numbers won't change.
Source: killcam.gg CS2 stats, Issue 3 (12 – 19 Oct 2026), https://killcam.gg/stats/issue/3<a href="https://killcam.gg/stats/issue/3">CS2 stats by killcam.gg, Issue 3</a>