How Fast Do Elite Fights Actually End? We Measured TTK Across 2,458 Kills
How this was made (method & sample) Source: the PUBG official API. The sample is the high-rank lobbies of top squad-FPP players from the Asia (pc-as) official leaderboard — 30 matches in the current season, 1,828 player-appearances (every human in those lobbies, not just the ranked accounts), yielding 2,458 measured kills. TTK uses exactly the same definition as this site's Combat Dashboard: only engagements resolving in a Kill or Knockdown are counted (a pure-damage engagement that never ends has no knockdown moment to measure), and kills where the victim was a bot are excluded (bots don't dodge or react, so their numbers aren't representative). Captured 2026-07-27. No account is named and no individual ranking is published.
"How long does an elite gunfight actually last?" sounds like an obvious question, yet almost no site has answered it with per-engagement data. We measured all 2,458 kills from high-rank Asian squad-FPP lobbies in one pass — and found the number is messier than it looks, because "TTK" quietly bundles two completely different things.
First, what a "one-shot kill" is and why it has to be separated
Every kill is either a single decisive hit (a sniper headshot, a point-blank shotgun, a grenade) or something that took more than one hit to put down. In the first case the damage event and the death confirmation land at essentially the same instant — there is no "how long it took." The milliseconds you measure are just the engine's gap between two events, not the duration of a fight. Blend those into one median and the result describes neither: it can't tell you how fast a one-shot is (that should approach zero) nor how long a normal fight runs. Two phenomena, one number, both hidden.
So this article does it in two passes: first the overall TTK distribution, honestly including one-shots, then a second median computed only from kills with at least two hits — the real firefights. Both numbers are given; you decide which one answers your question.
Overall median is 0.34 seconds — but a quarter of it isn't a fight
Across 2,458 measured kills, the overall median TTK is 340 ms. The distribution:
| Band | Share | Note |
|---|---|---|
| < 0.3 s | 45% | mostly one-shots |
| 0.3–0.6 s | 18.6% | |
| 0.6–1 s | 10.3% | |
| 1–2 s | 8.8% | |
| 2–4 s | 6.2% | |
| 4 s+ | 11.1% | drawn-out slugfests |
Almost half of all kills land under 0.3 seconds, which at first glance says "elites just delete people." But 26.5% (651 kills) of the total are the single-hit kills defined above — one hit, no second. They all fall into that "< 0.3 s" bucket and drag the overall median down with them.
Strip the one-shots and a real fight costs 0.54 seconds
Remove that 26.5% and look only at the 1,807 kills that needed at least two hits: the median becomes 538 ms — nearly 60% longer than the overall 340 ms. That figure is much closer to what most people are actually asking: how long does an elite need to close out a fight where both sides are trading?
Push the range out and follow-up time jumps 7.7×
Group the two-hit-or-more kills by distance and the median rises almost monotonically:
| Distance | Share | One-shot rate at this range | Follow-up TTK median |
|---|---|---|---|
| Point-blank (< 10m) | 14.8% | 20% | 337 ms |
| Close (10–50m) | 44.8% | 17% | 435 ms |
| Mid (50–100m) | 18.8% | 29.1% | 736 ms |
| Long (100–200m) | 13.3% | 46.2% | 1,103 ms |
| Very long (200m+) | 8.3% | 52% | 2,581 ms |
Two things are true at once. The further out, the higher the one-shot rate (20% point-blank → 52% past 200m — long-range kills are dominated by one-shot sniping in the first place). But when a long-range kill isn't a one-shot and needs a second hit, the cost explodes: the median follow-up past 200m is 2,581 ms, 7.7× the 337 ms of a point-blank follow-up. Intuitive enough — landing a second precise hit at range is far harder than up close — but this is the first time the "how much harder" has been measured rather than assumed.
How much do weapons differ? Rifles are nearly tied; snipers and DMRs are not
Looking only at kills with two or more hits, weapons split into two clearly distinct groups.
Automatic rifles and SMGs — when a follow-up is needed, almost all resolve around half a second:
| Weapon | Sample | One-shot rate | Follow-up TTK median |
|---|---|---|---|
| AKM | 97 | 16.5% | 303 ms |
| Mk47 Mutant | 28 | 17.9% | 247 ms |
| ACE32 | 94 | 13.8% | 423 ms |
| Beryl M762 | 264 | 18.6% | 426 ms |
| M249 | 53 | 24.5% | 425 ms |
| M416 | 214 | 10.3% | 436 ms |
| MP5K | 72 | 13.9% | 468 ms |
| SCAR-L | 23 | 0% | 472 ms |
| AUG | 532 | 13.3% | 502 ms |
| MG3 | 24 | 4.2% | 552 ms |
The AUG, with by far the largest sample (532), sits at 502 ms — not the fastest. The AKM (303 ms), widely considered harder to control, and the Mk47 Mutant (247 ms) both close faster. But the spread isn't dramatic: everything lands between 250 and 550 ms, which says your choice among the popular rifles matters much less than people assume.
Snipers, DMRs and grenades — high one-shot rate, but a long wait when the first one doesn't finish it:
| Weapon | Sample | One-shot rate | Follow-up TTK median |
|---|---|---|---|
| Dragunov | 55 | 40% | 1,003 ms |
| SLR | 33 | 24.2% | 1,337 ms |
| Mini 14 | 110 | 17.3% | 1,404 ms |
| Mk12 | 131 | 24.4% | 1,505 ms |
| Kar98k | 102 | 90.2% | 3,446 ms |
| M24 | 98 | 86.7% | 4,117 ms |
| Frag Grenade | 127 | 79.5% | 6,759 ms |
Bolt-action rifles (Kar98k, M24) resolve around 90% of their kills in a single shot — that is their design purpose. But for the remaining tenth that needs a second shot, the median runs 3.4–4.1 seconds, because cycling the bolt and reacquiring both take time. Grenades are more extreme still: nearly 80% land in one throw, and when a second attack is genuinely needed the median is 6.8 seconds (a grenade has its own cooldown and arc to pay for).
Being honest: what this didn't verify
- Limited sample scope: one region (Asia, pc-as), one season, high-rank squad-FPP lobbies only. Other regions and modes (solo, TPP) may not produce the same numbers.
- "Two hits" isn't always two trigger pulls for shotguns: a pump like the S1897 registers multiple pellets as separate hit events from a single shot, which gets misread as a follow-up and distorts the measurement (this weapon's follow-up median came out implausibly low). We excluded it from the weapon table rather than force in a possibly misleading number.
- Weapons need 20+ kills to appear: anything below that threshold (rare pistols and sidearms) is left out, so small-sample noise doesn't get mistaken for a conclusion.
- These players are the participants of high-rank lobbies, not just the ranked accounts themselves: same approach as our other recipes — lock onto recent matches of top leaderboard players, then analyse every human in those matches.
Conclusion
There is no single answer to "how long does an elite fight last" — it depends which kill you mean. The overall median of 340 ms sounds instant, but a quarter of it is one-shot kills with no process at all. Strip those out and the real median for a fight needing a follow-up is 538 ms, and it climbs sharply with every distance band: 2.6 seconds past 200m, 7.7× the point-blank figure. Between weapons, the mainstream automatic rifles are nearly tied (all around half a second); the real gap comes from whether a gun kills outright — the higher a sniper or DMR's one-shot rate, the more its design is about never giving you the chance to trade back.
Want to know how long your own fights last? Look up one of your matches on L!NCHPIN — the Combat Dashboard lists the real TTK of every engagement, so you can see how far you sit from these high-rank numbers.
Data is a single snapshot taken 2026-07-27, a 30-match cross-section of high-rank Asian squad-FPP lobbies. It does not represent all matches or other regions and modes. No account is named and no individual ranking or account information is disclosed.