High Ping or Lag Spikes in Games? How to Fix It
Rubber-banding, enemies teleporting, dying behind cover — that is rarely your average ping and almost always jitter: the packet-to-packet variance that Wi-Fi and bufferbloat create. The single biggest fix is to stop playing over Wi-Fi and plug in. The most overlooked one is bufferbloat, a queuing fault that no speed upgrade will touch. Measure ping and jitter first so you fix the cause you actually have.
Measure ping and jitter first — jitter is the tell
Ping is the round-trip delay between your input and the server registering it. Jitter is how much that delay varies from packet to packet — and it, not the average, is what produces lag spikes, because the game’s prediction can’t keep up when timing jumps around. A 20ms average with 15ms jitter plays worse than a rock-steady 25ms.
| Metric | Good | Playable | Bad |
|---|---|---|---|
| Ping (latency) | <20–30ms (competitive <15ms) | <60ms casual | >100ms sluggish |
| Jitter | <5ms | <10ms | >10–15ms rubber-banding |
| Packet loss | ~0% | <0.5% | 1%+ skipping/rewinding |
Before you rewire anything, get your numbers — run the Network Test and note both the ping and the jitter. A low, steady ping means the problem is a distant server or an in-game setting. A ping that jumps around, or high jitter, points straight at Wi-Fi or bufferbloat — the first two fixes below.
Fix 1 — Wire it: Wi-Fi to Ethernet is the biggest single change
If you do one thing, do this. A wired connection adds only about 1–5ms to the router and holds it steady. Wi-Fi typically adds 10–50ms of variable delay that swings with distance, walls, and every other device on the channel — and that variability is exactly the jitter that causes spikes. Switching a gaming PC or console to Ethernet removes both the latency add-on and the instability in one move.
No coax-free run to the room? Rank the alternatives honestly. If coax already reaches the room, aMoCA adapter gives near-Gigabit, consistently low latency and beats mesh Wi-Fi for stability (MoCA 2.5 is the mainstream pick; MoCA 3.0 is only just reaching retail around $120 a unit). A long Ethernet run or a tri-band mesh with dedicated backhaul is next. Powerlineis a budget last resort — its speed depends heavily on your wiring, and adapters on different breakers can lose half their throughput. It is not a first choice for competitive play.
Fix 2 — Kill bufferbloat with SQM
This is the most-overlooked cause of gaming lag. Bufferbloat is oversized, unmanaged queues in your router: when a big upload or download runs, your small game packets get stuck behind it and latency spikes even though the speed test still reads fast. It is a queuing problem, so upgrading your plan does nothing — you have to manage the queue.
The fix is SQM (Smart Queue Management), which uses the FQ-CoDel or CAKE algorithms to keep the queue short. It is built into eero, Ubiquiti, all Comcast DOCSIS 3.1 gateways, pfSense/OPNsense, and any OpenWrt router. When you set it, cap the SQM bandwidth at about 85% of your measured upload and download so your router — not the modem — controls the queue.
Your ping and jitter here give you the baseline; to prove bufferbloat specifically you need a latency-under-load test (the free Waveform Bufferbloat Test measures ping while saturating your line and grades it A+ to F — anything under a B is worth fixing).
Fix 3 — Escape 2.4GHz congestion
If you truly can’t wire in, at least get onto a clean band. The 2.4GHz band has only three non-overlapping channels — 1, 6, and 11 — and it is the most crowded because its longer range reaches every neighbor’s router too; many default to channel 6. Move gaming devices to5GHz, which is far less congested and whose shorter range is fine near the router, or use a Wi-Fi analyzer to pick the least-used of 1/6/11.
Fix 4 — Stop background uploads eating your headroom
Cloud backup (iCloud, OneDrive, Google Photos), game and OS updates downloading in the background, or another person streaming 4K or on a video call will spike everyone’s ping. Upload saturation is the worst offender because both your game traffic and the ACKs for any download need upload headroom. Pause backups and downloads while you play — or let SQM (above) hold latency down even when the line is maxed.
Fix 5 — Pick the nearest server or region
Accepting a default matchmaking region can quietly route you to a far server. Manually selecting the nearest one in the game or its matchmaking settings can cut 30–100ms or more, depending on how far the assigned server actually was. This costs nothing and is worth checking any time your ping is high but steady.
Fix 6 — Set up QoS (with one caveat)
QoS prioritizes gaming traffic over bulk transfers. On the router admin page, openAdvanced/QoS or Traffic Management → enable QoS → add a highest-priority rule for the gaming device’s IP or MAC; on ASUS routers, drag "Gaming" to the top of the traffic list ahead of Streaming and File Transfer. Reserve headroom rather than trying to use the whole pipe (on a 100 Mbps plan, guides suggest holding back 20–30 Mbps). The caveat: QoS only helps when the bottleneck is your own household’s traffic — for latency spikes under load, SQM matters more.
The evening-only lag: cable peak-hour congestion
If spikes cluster in the 7–11pm window and your tests are clean earlier in the day, this is shared-node congestion. On cable (DOCSIS), latency commonly rises from around 14–17ms off-peak to 22–28ms at peak, and speed can drop 15–40% on a loaded node. Fiber barely moves — roughly a 3–5% speed reduction and 1–10ms latency regardless of the hour. Short of switching to fiber or timing sessions around peak, this one is largely out of your hands.
What does not work
"A faster plan will fix my ping." It won’t. Speed determines how fast files download, not how responsive input feels — game packets are tiny and frequent, not bulk. Upgrades routinely leave ping unchanged. Connection type and topology set your latency, not the advertised Mbps.
"More Wi-Fi bars means lower ping." Signal strength says nothing about congestion, jitter, or bufferbloat — the things that actually cause spikes.
"A gaming VPN always lowers ping." Sometimes it finds a better-routed path to a distant server, but it also adds a hop and encryption overhead and can just as easily raise your ping. It is not a universal fix — test with it off first.
FAQ
What is a good ping for gaming?
Under 20–30ms is excellent and competitive-grade; under 60ms is fine for casual play; 50–100ms is noticeable in fast games but okay for browsing; above 100ms feels sluggish for anything real-time. Competitive FPS players in CS2 or Valorant often target under 15ms. Just as important is jitter — keep it under 5ms ideally, under 10ms at worst.
Why does my ping spike randomly while gaming?
Random spikes are almost always jitter, not a high average ping — and the usual cause is Wi-Fi or bufferbloat. Wi-Fi adds variable delay that jumps with distance, walls, and interference. Bufferbloat delays your game packets behind a big upload or download. A connection with 20ms average but 15ms jitter plays worse than 25ms average with 1ms jitter.
Does Ethernet really lower ping vs Wi-Fi?
Yes. Ethernet adds roughly 1–5ms to the router and stays steady; Wi-Fi typically adds 10–50ms of extra, variable delay. The variability is the real win — it is what removes the jitter and random spikes, not just a lower average. It is the single highest-impact change for gaming latency.
What is bufferbloat and how do I fix it?
Bufferbloat is oversized, unmanaged router queues that hold your small game packets behind large bulk transfers, so latency balloons under load even when raw speed looks fine. It is a queuing problem, not a bandwidth one, so a faster plan will not fix it. Enable SQM (Smart Queue Management) on your router — it uses FQ-CoDel or CAKE — and set its bandwidth limit to about 85% of your measured speed.
Does upgrading my internet speed reduce lag?
Almost never. Download speed decides how fast a 100GB game installs, not how responsive input feels — game traffic is tiny packets sent many times a second, not a bulk transfer. People report going from a few Mbps to many times that with ping essentially unchanged. What lowers ping is connection type and topology: fiber vs cable, wired vs Wi-Fi, and killing bufferbloat.