Ping is the round-trip time, in milliseconds, between your input and the game server's response. You lower it by removing every device and process between your PC and the server: use Ethernet instead of Wi-Fi, close bandwidth-hogging apps, switch to the nearest server region, and get your router's bufferbloat under control. If your base latency is already 25β40 ms to a nearby server, no tweak will push you to 5 ms β you're at the speed-of-light floor for that distance.
Most guides tell you to "port forward for lower ping." That advice is wrong, and it's the first thing worth clearing up. Port forwarding opens inbound pathways so hosting and peer connections work; it doesn't shorten the hops between you and a server. Worse, it exposes ports to the wider internet, so applying it blindly is a security cost with zero latency benefit. What actually moves the number is physical distance, local congestion, and a router queue that balloons under load.
This article walks through the fixes in the order they actually help, names the settings you'll see on screen, and flags the places where chasing a lower number costs more than it returns.
Measure Your Real Baseline Before Changing Anything
You need a true baseline before you touch a setting, and the game's own scoreboard isn't one. Game menus often show a smoothed or averaged figure, and some titles display "ping" that includes server tick processing. Use Windows' built-in measure: open Command Prompt and run ping -n 20 against your game's server region, if you can find the IP from a community server list. The average of 20 packets, not one reading, is your baseline.
Steam's server browser gives you a free reference point without entering a match. It lists regional servers with a ping column you can screenshot for comparison before and after changes. Write down three numbers: idle ping to a nearby server, ping while a download runs, and ping while another household member streams video. The gap between the first and the other two is your local congestion and bufferbloat problem in plain view.
A traceroute narrows where the delay happens. Run tracert in Command Prompt and look for the hop where latency jumps by more than 30 ms and stays high. If the jump happens at the second or third hop, the problem is your router or ISP edge. If it happens at hop eight, it's the wider internet and no local change fixes it.
Test at the same time of day for two or three days. Evening congestion between 7 p.m. and 11 p.m. can add 15β40 ms in dense residential areas, and that pattern tells you the fix is scheduling or QoS rather than hardware. A single test at 2 a.m. will flatter your connection and mislead you.
Pick the Closest Game Server Region, Not the Fullest One
Distance is the tax you can't negotiate away. Light in fiber travels at roughly two-thirds of its vacuum speed, around 200,000 km per second, so 1,000 km of straight-line distance costs about 5 ms one way and 10 ms round trip β before routing detours, which typically double the physical path. That's why a player in New York connecting to a Dallas server sees roughly 40β50 ms while the same player connecting to Frankfurt sees 90β110 ms.
Matchmaking sometimes puts you in a fuller server across an ocean because the algorithm optimizes queue time. In games with a server browser or region selector, force the nearest region manually. In titles that auto-assign, check the in-game network menu for a region or datacenter display and confirm it's the one you'd choose.
The counterintuitive part: a slightly higher-ping server in your own region feels better than a lower-ping server far away, because packet loss and jitter scale with path length, not just latency. A 35 ms connection with 1% packet loss will feel worse than a 55 ms connection with clean delivery.
If the game offers crossplay with console players on a wider pool, note that the pool and the server region are separate settings. Restricting the pool can shrink the player base without moving you closer to a datacenter.
Switch to Ethernet, Because Wi-Fi Is the Most Common Cause
Wi-Fi is the single biggest avoidable latency source in home gaming, and the numbers are concrete. Wi-Fi 6 (802.11ax) adds its own retransmission and channel-sharing delay on top of the airtime; typical residential Wi-Fi adds 3β15 ms of jitter versus a wired link, and in a crowded 2.4 GHz band with neighbors' networks, spikes of 50β100 ms are routine. A $10 Cat 5e or Cat 6 cable from your router to your PC removes all of it.
Here's the physical part most guides skip: run the cable along the baseboard and behind furniture rather than across a doorway, and if the run exceeds about 90 meters, you need a switch or a different plan. For rooms on another floor, a pair of powerline adapters or a MoCA adapter over existing coaxial cable beats Wi-Fi in most homes, though powerline performance depends heavily on your home's wiring age and circuit layout β test before committing.
If you absolutely can't run cable, the next best option is a Wi-Fi 6 router on the 5 GHz band with the PC in the same room, line of sight, and the channel set manually to a non-overlapping one. Sit closer. Turn off 2.4 GHz for the gaming device if the router lets you split bands.
One failure case worth knowing: some powerline adapters go to sleep to save energy, adding a 100β200 ms wake-up spike on the first packet of a burst. If your ping looks fine in menus and terrible at match start, suspect the adapter, not the game.
Kill Local Bandwidth Hogs: Windows, Steam, and Cloud Sync
Bandwidth saturation raises ping because your packets queue behind other traffic at your router. The fix is finding what's uploading, not what's downloading β upload saturation hurts latency more per megabit because residential upload caps are smaller. Windows Update's delivery optimization and peer-to-peer sharing uploads to other PCs, OneDrive and Google Drive sync upload continuously, and Steam's client downloads game updates in the background even when you aren't playing.
Open Task Manager (Ctrl+Shift+Esc), go to the Performance tab, open Resource Monitor from the bottom link, and check the Network tab's "Send (Bytes/sec)" column sorted descending. That shows you the process uploading at the moment. Windows Update's delivery optimization can be limited in Settings > Windows Update > Advanced options > Delivery Optimization > Advanced options, where you can cap the percentage of upload bandwidth.
On Steam, go to Settings > Downloads and enable "Limit auto-updating schedule" so patches only apply during hours you're not playing. Set the download speed limit too β it's under the same panel β because an unthrottled Steam download will flood your router's queue and push in-game ping from 30 ms to 200 ms or worse.
Console and phone on the same network count. A single 4K stream from Netflix runs at roughly 15 Mbps, which is enough to saturate a modest DSL or fixed-wireless upload and cause ping spikes even if your download looks fine.
Fix Bufferbloat With Router QoS and SQM
Bufferbloat is the specific reason your ping spikes to 300 ms the moment someone uploads a photo. It happens because routers have large buffers, and when the line is saturated, game packets sit in a queue behind bulk traffic instead of being dropped and retried. The buffer that protects downloads wrecks latency.
The fix is Smart Queue Management (SQM) or a well-configured QoS setting. The key number: set your upload limit to about 85β95% of your measured upload speed, not 100%. Leaving headroom keeps the buffer from filling, which is where the latency comes from. On routers running OpenWrt or the open-source cake algorithm, SQM is a built-in package; on consumer routers, look for "QoS," "Bandwidth Control," or "Adaptive QoS" in the admin panel and prioritize the gaming device or the "Gaming" category.
Test before and after with the Waveform Bufferbloat Test or DSLReports' speed test, both of which grade latency under load on an AβF scale. A connection that grades "F" under load with 250 ms added latency can typically be brought to "A" or "B" with SQM at 90% of upload speed. That single change often does more for in-game feel than any ping "booster."
One caveat: on very fast connections, some consumer routers' QoS processors can't keep up with the throughput, so you'll cap your download to protect your ping. That's a real trade β decide whether you want 500 Mbps or consistently low latency, because on that hardware you may not get both.
Compare the Fixes Before You Spend Money
Not every fix costs the same or returns the same. Work down this table in order; the top rows are free and usually decisive.
| Fix | Typical cost | Ping reduction (typical) | Effort | When it fails |
|---|---|---|---|---|
| Ethernet cable | $8β15 for Cat 6 | 10β40 ms, removes most jitter | Low | Room has no route to router |
| Close background uploads | Free | 5β200 ms during spikes | Low | You don't own the network |
| Nearest server region | Free | 20β80 ms vs. distant region | Low | Matchmaking forces region |
| Router QoS/SQM | Free (built in) or $60β150 router | 50β250 ms under load | Medium | Router lacks the feature |
| Powerline or MoCA adapter | $40β90 per pair | 5β20 ms vs. Wi-Fi | Medium | Old wiring, long runs |
| Gaming VPN | $5β12/month | Often negative; occasionally 10β20 ms better if ISP route is bad | Low | Your ISP route is already good |
| New router/gateway | $80β250 | Varies; matters mainly for bufferbloat and Wi-Fi | Medium | Base line is already clean |
The verdict is split by situation. If your ping is fine when idle and terrible under load, SQM is your fix, not new hardware. If your idle ping is consistently 100 ms+ to a nearby server, the problem is your ISP's routing and a VPN might reroute around it β test a free trial before paying. If idle ping is 25 ms and you just want it steady, spend the $10 on the cable.
The recommended pick for most people is the wired connection plus firmware-level SQM, because together they address both the average and the spikes. The real drawback: SQM at 90% of upload caps your sustained upload speed, and some router firmware applies QoS to all devices, so a large backup you start will now take a bit longer.
Turn Off the Software That Adds Hops: VPNs, Overlays, and "Boosters"
Gaming VPNs and "ping booster" services promise to route your traffic through optimized paths. Sometimes they help: if your ISP's route to a game server crosses a congested peering point in another city, a VPN's alternate path can cut 10β20 ms. More often they add a hop and slow you down. The honest way to evaluate one is a return window and a head-to-head test.
Discord's screen share, game overlays from GeForce Experience or the Xbox Game Bar, and live-streaming software all consume bandwidth and CPU. Discord's overlay, in particular, is fine at idle but its screen share pushes upload traffic through the same queue as your game packets. Disable the overlay in Discord's settings and close the app entirely if you aren't using voice.
Antivirus and firewall software can inspect game traffic packet by packet. That's usually negligible, but some suites' "game mode" or web shields cause periodic 100 ms+ stalls while they scan. If your ping is smooth and then hiccups at regular intervals, the scan interval is the suspect β whitelist your game's executable rather than disabling protection.
The failure case to watch: a "gaming accelerator" that claims to lower ping is often a VPN that routes through its own servers, and if its nearest node is in another city, you've just moved your packets the wrong direction. Test with the free tier before paying.
Tune In-Game and Windows Network Settings That Actually Matter
Most in-game "network smoothing" or interpolation settings don't lower ping β they change how the game hides it. Raising interpolation buffer makes other players' movement smoother at the cost of slightly delayed visible positions, and lowering it makes hit registration feel sharper but exposes jitter. On Counter-Strike 2, the community-standard practice is to keep interpolation at defaults and fix the underlying connection; the launch option -tickrate does nothing on official 128-tick servers.
Server maximum tick rate matters more for feel than measurable latency: a 128-tick server registers actions about twice as often as a 64-tick one, and the difference reads as "trading kills" rather than as a ping number. If a game lets you pick community servers, favor higher tick rate and nearby location together.
On Windows, there's one setting worth checking: disable "Delivery Optimization" peer uploads as described earlier, and make sure your network adapter's driver is current β old drivers occasionally negotiate at a capped link speed. Don't disable Nagle's algorithm or apply "network throttling" registry tweaks from old forum posts; those were written for Windows Vista-era code and can cause more harm than good on current builds.
One more edge case: if you share a router with a housemate who runs a torrent client, the connection table can fill up and slow everything, and no game-side setting fixes it. That's a conversation, not a configuration.
Know When the Problem Isn't Ping at All
Many players chasing lower ping are actually fighting packet loss, jitter, or a bad server. Packet loss shows up as teleporting, hits that don't register, and rubber-banding even when your ping reads 30 ms. Jitter β the variance between packets β matters more for smoothness than the average, and a stable 60 ms beats an unstable 35 ms.
Run a ping test over several minutes and look at the min, max, and average. A connection with a 30 ms min, 30 ms average, and 200 ms max has a jitter problem, and the fix is usually removing a Wi-Fi link or bufferbloat. Packet loss of even 1% is visible in fast-paced games.
The server matters too. If one teammate with great ping still gets hit behind walls, the server may be under load or running low tick rate. Check server status pages for the game you play, and if a specific match feels off while a new match feels fine, the problem was the server, not you.
Finally, some hardware's network stack bottlenecks before the cable does. USB Wi-Fi adapters and cheap USB-Ethernet dongles can cap throughput and add latency under load; internal PCIe or motherboard Ethernet doesn't. If you're using a dongle, that's the next thing to swap.
FAQ
Does port forwarding lower ping?
No β port forwarding changes which inbound connections are allowed, not how fast packets travel, and it adds exposure you don't need. The common advice to forward game ports for ping is a misconception carried over from hosting-focused guides. If a game needs a port for hosting, forward it for that reason only.
Why is my ping high only in the evening?
Evening congestion between roughly 7 p.m. and 11 p.m. is residential internet peak time, when your neighborhood's shared ISP capacity fills. Added latency of 15β40 ms is typical on cable and fixed wireless. SQM on your router helps with your in-home queue, but neighborhood congestion requires your ISP or a different access technology.
Can a VPN really lower my ping?
Occasionally, and only when your ISP's route to the server crosses a congested peering point that the VPN's path avoids. A 10β20 ms improvement is possible on bad routes; a 20β50 ms penalty is just as common on good ones. Use a free trial and measure with the game's own server, not the VPN's marketing page.
Is 20 ms ping good enough?
Yes β 20 ms is excellent for any game, and going lower rarely changes the outcome. At 20 ms round trip, the visible difference versus 10 ms is below the threshold most players can perceive, and matchmaking opponents typically connect at 30β80 ms. Chasing single-digit ping is mostly placebo beyond that point.
Why does my ping spike when someone streams video?
Video streaming consumes download bandwidth, and downloads leave small acknowledgement packets going upstream that fill the upload queue. That's bufferbloat, and it pushes game packets behind a full buffer. Setting router QoS/SQM at about 90% of upload speed usually caps the spike to single digits.
Does a higher-tick server lower my ping number?
No β tick rate doesn't change round-trip time, but it does change how often the server processes input, which reads as faster hit registration. A 128-tick server at 60 ms often feels sharper than a 64-tick server at 40 ms. Ping and tick rate are separate measurements, and both show in different parts of the scoreboard.
What's the realistic floor for my connection?
The physical floor is set by distance and the speed of light in fiber, roughly 5 ms per 1,000 km one way before routing detours. Routers and peering usually double the physical path, so a 500 km game server realistically floors around 10β20 ms round trip. Anything below that figure from a distant server is measurement error.
Does upgrading my router reduce ping?
Only if your current router is the bottleneck β typically through bufferbloat, weak Wi-Fi, or an old CPU that can't handle QoS. A modern router with SQM helps most homes, but a player already wired to a decent router at 30 ms idle won't see improvement. Measure under load first.