How to Reduce Ping: Diagnose and Fix High Latency
Find out which of the four layers is costing you milliseconds, then fix that one. Covers bufferbloat, CGNAT, traceroute reading, and when a VPN or routing optimiser actually helps.
Almost every guide about reducing ping opens with the same list: close background apps, use an ethernet cable, restart your router. That advice is not wrong, but it skips the part that actually matters. Ping is not one number produced by one thing. It is the sum of four separate delays stacked on top of each other, and each one has a different fix. If your latency problem lives in the third layer and you keep applying fixes aimed at the first, you will spend a week changing settings and see nothing improve. This guide does it the other way around: work out where your milliseconds are actually being lost, then fix that specific layer.
The single most useful thing you can learn about ping is which part of it you can change and which part is fixed by geography. Nothing you install can make a signal travel faster than light through glass.
Ping, latency, jitter, and packet loss are four different problems
People use these words interchangeably, and that is a large part of why the wrong fixes get applied. They describe different failures and they feel different in a game.
Latency is the one-way time for a packet to travel from your machine to the server. Ping is the round trip: there and back. When a game shows you 40ms, it is almost always showing round-trip time, which means the one-way journey is roughly 20ms. This distinction matters when you start reading traceroute output, because those tools also report round trips.
Jitter is the variation in that number. A connection that sits at a steady 60ms is far more playable than one that swings between 20ms and 90ms, even though the second one has a better average. Games predict where objects will be based on recent timing. Steady latency is predictable and the game compensates well. Jitter breaks the prediction, and that is what produces rubber-banding, hits that do not register, and characters teleporting across your screen.
Packet loss is when data never arrives at all. The game has to either wait for a retransmission or carry on with a gap in its understanding of the world. Even one or two percent loss is far more destructive than an extra 30ms of stable ping. If you have to choose between a server with lower ping and one with no packet loss, take the one with no loss every time.
This is why "my ping is fine but the game feels terrible" is such a common complaint. The average is fine. The consistency is not. When you start diagnosing, look at the shape of the numbers over a minute of play, not the single figure in the corner of the screen.
The floor you cannot go below
Light travels through fibre optic cable at roughly two thirds of its speed in a vacuum. That works out to about 1ms of round-trip time for every 100km of cable, and real routes are never straight lines. Cables follow coastlines, railway lines, and existing infrastructure, so the actual distance travelled is commonly 1.5 to 2 times the straight-line distance between two cities.
Some practical numbers. If you are in London playing on a server in Frankfurt, roughly 640km apart, your theoretical floor is around 13ms and a realistic good result is 15 to 20ms. London to New York is about 5,500km, so the floor is around 55ms and a good real-world result is 70 to 80ms. London to Sydney is roughly 17,000km, giving a floor near 170ms and a realistic 250 to 300ms.
Every piece of equipment along the way adds to that floor. Each router that inspects and forwards your packet costs a fraction of a millisecond, and there are typically 10 to 20 of them on an international path. Your own connection type adds a baseline too: fibre adds almost nothing, cable adds a few milliseconds, and older DSL lines can add 10 to 40ms before your traffic has even left the neighbourhood.
The reason this matters is that it tells you when to stop optimising. If you are in Australia playing on a European server at 280ms, no software, no router setting, and no paid service is going to bring that to 60ms. The honest fix is to play on a closer server. Knowing your floor stops you wasting money on tools that cannot help with your particular problem.
The four layers where milliseconds disappear
Every millisecond of your ping is added in one of four places. Diagnosing properly means finding which one, because a fix aimed at the wrong layer does nothing at all.
Layer 1: your device
Delay added before your data reaches the network card at all. Background downloads, cloud sync, streaming in another window, game overlays, outdated network drivers, and aggressive security software all live here. This layer is usually the easiest to fix and it is where most guides stop.
Layer 2: your local network
Your WiFi signal, your router, and everything else in the house competing for the same connection. This layer is responsible for far more latency problems than most players realise, and one specific problem here, bufferbloat, is both very common and almost never mentioned.
Layer 3: your ISP and the path across the internet
Once your traffic leaves the house it crosses your ISP's network, then hands off to other networks until it reaches the game server. How good that handoff is depends on business relationships between networks called peering, and it varies enormously between ISPs in the same city. This is the layer nobody can see without looking, and it is where a routing optimiser can genuinely help.
Layer 4: distance
The floor described above. Fixed by physics and only changeable by choosing a different server.
Diagnose before you change anything
Do this in order. Each step narrows down which layer owns your problem, and the whole sequence takes about fifteen minutes.
Step 1: get a real baseline
Turn on your game's network display. Most competitive titles have one: a netgraph, a performance overlay, or a latency readout in the settings menu. Play for a full five minutes and watch the shape of the number rather than its value. Write down the lowest value you see, the highest, and how often it jumps. A connection that reads 45ms and never moves is a completely different problem from one that averages 45ms while spiking to 200ms every thirty seconds.
Note when it happens, too. Ping that is fine at 10am and terrible at 8pm points squarely at congestion, either in your own house or in your ISP's local network. Ping that is consistently bad at all hours points at routing or distance.
Step 2: separate local problems from everything else
Open a terminal and ping your own router. On Windows use ping 192.168.1.1 -n 50 and on macOS or Linux use ping -c 50 192.168.1.1. Your router address may be different; if that one does not respond, our guide to finding your default gateway covers how to find it on any platform.
This tells you the cost of getting from your device to your own router, before the internet is involved at all. On a wired connection you should see well under 1ms with zero loss. On good WiFi, 2 to 5ms is normal. If you are seeing 20ms, 100ms, or any packet loss at all on this test, stop here. Your problem is entirely inside your own house, and nothing you do to your ISP or your game server settings will fix it.
Step 3: look at the path to the server
Run a traceroute to something in the region where you play. On Windows, tracert 8.8.8.8; on macOS or Linux, traceroute 8.8.8.8. Better still, if you know your game server address, trace to that.
Read it as a story about your traffic. The first hop is your router. The next few are your ISP's local network. After that you will see the handoff to larger networks, and eventually the destination. What you are looking for is the point where the numbers jump sharply and stay high. A jump from 12ms to 90ms between two consecutive hops means those two routers are far apart, or the link between them is congested.
Two things commonly confuse people reading traceroute output. First, some routers deprioritise the replies that traceroute depends on, so a single hop showing a high number or asterisks while later hops are fine is not a problem: that router is just deprioritising your probe, not your game traffic. Only a jump that persists through every subsequent hop is real. Second, the last hop is the one that matters. Intermediate weirdness that does not affect the final number is noise.
Step 4: find out whether you are behind CGNAT
Carrier-grade NAT means your ISP has put you behind a shared public address instead of giving you your own. It adds a translation step to every connection, and more importantly it interferes with the peer-to-peer connections that many games use for matchmaking and voice chat. It is very common on mobile broadband and increasingly common on fixed lines.
Run our CGNAT test to check in a few seconds. If you are behind CGNAT, that alone can explain strict NAT types, failed party invites, longer matchmaking, and being placed on distant servers because direct connections could not be established. Our guide on why port forwarding fails under CGNAT covers what still works when you are in this situation.
Step 5: check your NAT type and ports
Strict NAT does not usually raise your raw ping, but it degrades matchmaking quality in ways that feel identical: longer queues, worse server selection, and more host migrations. Consoles report NAT type directly in their network settings, where Type 1 or Open is best, Type 2 or Moderate is normal and fine, and Type 3 or Strict causes problems.
On PC you can check whether the ports your game needs are actually reachable from outside using our open port checker. A filtered result there, combined with a positive CGNAT test, confirms that the block is upstream of your router rather than something you misconfigured.
Step 6: identify who actually carries your traffic
Two customers in the same city on different ISPs can have wildly different routes to the same game server. Use our ASN lookup to see which network announces your address, and browse the ISP directory for details about that provider. This is worth knowing before you conclude that your latency is unavoidable, because if your ISP routes badly to a particular region, the fix may be as simple as switching providers or as targeted as a routing optimiser.
Why your ping suddenly got worse
A connection that was fine last week and is bad today has a much shorter list of possible causes than one that has always been bad. Something changed, and the question is what.
Something new joined your network.A new television, a security camera, a housemate's console, or a phone set to back up photos over WiFi can all consume enough of your connection to trigger the bufferbloat problem described above. Check your router's connected devices list and see whether anything appeared recently.
Your router or a device updated itself. Firmware updates sometimes reset settings including QoS and port forwarding rules. Operating system updates can replace network drivers with generic versions and re-enable power saving on wireless adapters. If your problem started the morning after an update, that is your suspect.
Your ISP changed something. Providers reroute traffic, add or drop peering arrangements, and perform maintenance without announcing it to customers. A traceroute that now shows a different path than it used to is direct evidence. This is also the most common explanation when a whole neighbourhood reports the same problem at once, so it is worth checking a local outage map or community forum.
The game changed. Publishers move server infrastructure between providers and regions, sometimes without notice. If your ping to every other destination is unchanged but one game got worse, the game is the variable, not your connection.
Seasonal and time-of-day congestion. Evening peaks are normal, but they get worse as more households join a segment. A connection that coped last year may not cope now, which usually shows up as good daytime performance and poor evening performance.
Ping spikes every few seconds
Regular, rhythmic spikes are a distinct problem from generally high latency, and they have their own short list of causes. If your ping sits at 30ms and jumps to 200ms every few seconds like clockwork, work through these.
Wireless adapter power saving. The single most common cause on laptops. The adapter sleeps between transmissions and takes time to wake. Open your network adapter properties, find the power management settings, and disable power saving for that adapter. On Windows, also check that the power plan is not set to a battery-saving mode.
WiFi scanning. Devices periodically scan for other networks, and on some drivers this interrupts the active connection. Connecting by cable eliminates it entirely, which is also a fast way to confirm the diagnosis.
Interference on a busy channel. Microwaves, cordless phones, baby monitors, and neighbouring networks all compete on the same frequencies, and 2.4GHz is far more crowded than 5GHz. If the spikes correlate with someone using the kitchen, that is not a coincidence.
Background scanners. Security software that performs periodic scans, and backup tools that wake on a schedule, produce spikes at regular intervals. The timing is the clue: a spike exactly every 60 seconds points at software, not at radio.
A failing cable or port. On wired connections, intermittent spikes with occasional loss can mean a damaged cable or a dying switch port. Swapping the cable costs nothing and rules it out.
Fix layer 1: your device
Start here because it is free and fast, but do not expect miracles unless your baseline test pointed at this layer.
Check what is using your connection. On Windows, open Task Manager, go to the Performance tab, click Open Resource Monitor, and look at the Network tab to see exactly which processes are consuming bandwidth. Game launchers updating other titles in the background are the single most common culprit, followed by cloud storage sync and system updates. Set launchers to not auto-update while you play.
Update your network adapter driver from the manufacturer's site rather than relying on the operating system, which often keeps an older version. If you use a WiFi adapter, check whether power saving is enabled for it in the adapter properties, because that setting causes periodic latency spikes that look exactly like network problems.
Turn off overlays you do not use. Recording software, chat overlays, and performance monitors all hook into the game and can add input latency that feels like network lag. This will not change your ping number, but it can change how the game feels, and it is worth ruling out before you chase network fixes for a problem that is not on the network.
Fix layer 2: your local network
This is where the biggest free wins usually are.
Use a cable if you possibly can
This advice is repeated everywhere because it is genuinely the highest impact change available to most players. WiFi adds latency and jitter even when the signal looks strong, because the airwaves are a shared medium and your device has to wait its turn. A wired connection typically removes 5 to 20ms and, more importantly, removes most of the variance.
If running a cable is impossible, powerline adapters are usually better than WiFi for latency, and a mesh node with a wired backhaul is better than one relaying over the air. If you must use WiFi, use the 5GHz band rather than 2.4GHz, sit closer to the router, and check for channel congestion using your router's admin page or a WiFi analyser app.
Fix bufferbloat, the problem nobody mentions
This one deserves its own section because it is extremely common, it produces dramatic symptoms, and almost no gaming guide covers it.
Bufferbloat happens when network equipment has oversized buffers. When your connection is saturated, say because someone is uploading a video or a console is downloading a patch, the router queues packets rather than dropping them. Your game packets end up sitting in a long queue behind that bulk traffic. The result is a connection that tests fine when idle and becomes unplayable the moment anything else uses the line.
The tell is unmistakable: your ping is 30ms normally and jumps to 300ms or higher whenever someone else in the house starts a download or a video call. If that describes your situation, you have bufferbloat, and no amount of closing background apps on your own machine will fix it because the problem is other devices.
The fix is Smart Queue Management, usually listed in router settings as SQM, fq_codel, cake, or sometimes just Bufferbloat or Adaptive QoS. If your router supports it, enable it and set your bandwidth limits to about 85 to 90 percent of your actual measured speed. That deliberate sacrifice of peak throughput keeps the queue short and can take a connection from 300ms under load to under 40ms. If your router does not support it, this is one of the few cases where buying better hardware genuinely fixes a latency problem.
Ordinary QoS that simply prioritises gaming traffic is less effective than SQM but better than nothing. Prioritise your gaming device by its address, and remember to give it a static assignment so the rule keeps matching.
Look at the router itself
Old or ISP-supplied routers are frequently the bottleneck. A router with an underpowered processor will add latency under load regardless of your connection speed. Update the firmware, and if the unit is more than five years old, consider that replacing it may do more for your latency than any subscription service.
Restarting the router genuinely does help sometimes, though not for the reasons usually given. It clears the NAT translation table, which can become full on busy connections, and it can trigger a fresh connection to your ISP that may take a different path. It is worth trying once. It is not a maintenance routine.
Fix layer 3: the path across the internet
If your local tests are clean, your device is quiet, your connection is wired, and your ping is still bad, the problem is between your ISP and the game server. This is the layer where the interesting decisions are.
Understand what is actually going wrong
Networks connect to each other through peering and transit arrangements. When two networks peer directly, traffic between them takes a short path. When they do not, traffic has to travel through an intermediate network, which can mean a long detour. A player in one city might reach a game server in 25ms while their neighbour on a different ISP takes 90ms for the same destination, purely because of who peers with whom.
Congestion is the second path problem. A peering link that is running at capacity during peak hours will queue traffic, adding latency and sometimes loss. This is the classic pattern of good ping in the morning and bad ping at night when nothing at home has changed.
Your traceroute from step 3 is the evidence. If you see the big jump happening several hops out, past your ISP's own network, and it gets worse in the evening, you are looking at a congested or inefficient path rather than anything you control.
The free options first
Change server region in the game if there is a closer option. Many games pick a server automatically and pick badly, especially if CGNAT interfered with the connection test.
Contact your ISP with your traceroute. This works more often than people expect, particularly if you can show that the problem starts at a specific hop inside their network. Support scripts will push you towards speed tests, so be specific: state that throughput is fine and latency to a named destination is the problem, and provide the trace.
Changing DNS provider is frequently recommended for lower ping. Be clear about what it does: DNS affects how quickly a name is resolved into an address, which affects connection setup and page loads, not the latency of an established game connection. It is a worthwhile change for other reasons but it will not lower your in-game ping.
When changing ISP is the right answer
Sometimes the honest conclusion is that your provider is the problem and no product will fix it. Three signals point that way together.
First, your local tests are clean: wired, low latency to your own router, no loss, queue management working. Second, your traceroute shows the jump happening inside your ISP's own network or at its first handoff, rather than close to the destination. Third, someone else in your area on a different provider gets substantially better results to the same game server. That last check is the most persuasive, and gaming communities for your region are usually the fastest way to get it.
Before switching, work out what you would actually be buying. Use our ASN lookup to identify the network behind each candidate provider, since several retail brands often resell the same underlying network and switching between them changes nothing about your route. Check the connection technology too: moving from DSL to fibre changes your baseline in a way that no configuration can, while moving between two fibre providers may only change the peering.
Ask specifically whether the provider issues a genuine public address or places customers behind CGNAT, because that affects matchmaking and port forwarding as much as raw latency. Some providers offer a public address on request or as a small add-on, which can be a cheaper fix than changing provider entirely.
The paid options, and what each one actually does
Two categories of product claim to help here, and they work differently.
A VPNencrypts your traffic and sends it through a server you choose. By default it adds a hop, so on an already-good connection it will usually add a few milliseconds rather than remove them. It can genuinely reduce ping when your ISP throttles gaming traffic or routes inefficiently, because the VPN provider's network may take a better path. It also hides your home address, which matters if you are being targeted. Our guide to VPNs for gaming covers the picks and the honest limits, and gaming IP security covers the DDoS angle specifically.
A routing optimiser, sometimes called a game booster, is a different category. It is not a VPN and does not primarily exist to provide privacy. It routes your game traffic over paths its operator has selected, aiming to avoid the congested or inefficient hops your ISP would otherwise use. GearUP is one option in this category. GearUP describes its approach as adaptive routing across multiple platforms including Windows, mobile, and consoles, and states that it supports over 3,000 games with a free trial available.
Be clear about when this category can and cannot help, because the marketing around it is often vague. A routing optimiser can help when your traceroute shows an inefficient or congested path, when you play regularly on servers in another region, or when your ISP's peering to a particular game network is poor. It cannot help when your problem is at layer 1 or 2, when your route is already direct and efficient, or when the distance itself is the limit. If your baseline test showed the problem was your WiFi or bufferbloat, this category of product will change nothing, and you should fix the actual cause instead.
The right way to evaluate any of these is with the free trial and your own numbers. Record your ping and jitter over five minutes of real play before you install anything, then repeat the identical test with the product active on the same game, same server, and same time of day. Trust that comparison over any advertised figure, including the vendor's own statistics.
If your own before-and-after numbers do not improve, cancel and go back to the layer that is actually costing you the milliseconds. There is no shame in discovering that the answer was a cable.
Console-specific notes
Consoles hide most network detail, which makes diagnosis harder, but the same layers apply.
Check the NAT type in network settings first. On PlayStation this appears as Type 1, 2, or 3; on Xbox as Open, Moderate, or Strict. Moderate or Type 2 is normal for most households and is not worth chasing. Strict or Type 3 is worth fixing, and the usual causes are CGNAT, double NAT from having two routers in series, or UPnP being disabled.
Double NAT is common and easy to miss. It happens when an ISP-supplied modem-router is running in router mode and you have added your own router behind it. The fix is to put the ISP device into bridge or modem mode so only one device is performing translation.
Consoles also download updates aggressively in the background, including for games you are not playing. Set downloads to pause during gameplay if the option exists, and check whether another console or PC in the house is doing the same thing.
Mobile and cloud gaming
On mobile networks, latency varies far more than on fixed lines, and CGNAT is close to universal. Signal strength matters more than the advertised generation of the network: a strong 4G signal will usually beat a weak 5G one for latency and consistency. Avoid playing while moving between cells if you can, because handovers cause spikes.
Cloud gaming has a different profile because your inputs travel to the server and video comes back. That makes it far more sensitive to jitter and loss than to raw ping, and it makes a wired connection or a strong 5GHz signal much more important than for a locally rendered game.
Myths worth discarding
A faster plan lowers ping. Bandwidth and latency are different properties. Upgrading from 100Mbps to 1Gbps will not reduce your ping unless you were genuinely saturating the slower connection. The exception is switching connection technology, because moving from DSL to fibre changes the baseline.
Ping boosters make your internet faster. Nothing makes your packets travel faster than the medium allows. Products in this category change which path packets take. That is a real and sometimes valuable thing, but it only helps if the path was the problem.
Closing every background app is essential. Modern systems handle background processes well. Unless something is actively using the network or pinning your processor, closing it changes nothing measurable.
Registry tweaks and TCP optimisers help. Most of these target settings that stopped mattering many years ago, and games typically use UDP rather than TCP anyway. The risk of breaking your network configuration outweighs any benefit.
What counts as good ping
Under 20ms is excellent and only realistic on a server near you. Twenty to 50ms is very good and competitive for any genre. Fifty to 80ms is fine for the large majority of games including most shooters. Eighty to 120ms is noticeable in fast competitive play but perfectly workable for strategy, role-playing, and co-operative games. Above 150ms, fast-paced competitive games become genuinely difficult.
Weigh consistency alongside the number. A steady 80ms with no loss beats a 40ms connection that spikes and drops packets, and it is worth choosing servers on that basis rather than picking whichever shows the lowest figure at the moment you check.
One more variable explains why a good ping sometimes still feels wrong. Servers update the game world at a fixed rate, called the tick rate, and a server running at 20 ticks per second only refreshes every 50ms no matter how fast your connection is. On a low tick rate server, the difference between 20ms and 40ms of ping is largely invisible because the server itself is the bottleneck. If your connection tests clean and the game still feels unresponsive, the server may simply be the limiting factor, and that is not something any change on your end will fix.
How to test a change properly
Most people change three things at once, feel that the game is better, and never know which change did it or whether anything changed at all. Latency varies enough hour to hour that a casual impression is close to worthless. A few minutes of discipline will save you from paying for something that did nothing.
Change one thing at a time. If you enable queue management, switch to a cable, and start a trial of a routing product in the same evening, you have learned nothing about any of them. Worse, if the result is better you will likely keep paying for the one that costs money when the cable was doing the work.
Control the conditions. Test the same game, the same server region, and roughly the same time of day. Evening and daytime results are not comparable, because congestion is the thing you are trying to measure around.
Measure for long enough. Five minutes of real play, not thirty seconds in a menu. Record the lowest ping, the highest, and how often it spiked. If you want a cleaner number, run a sustained ping to a fixed destination in the background using ping -n 300 8.8.8.8 on Windows or ping -c 300 8.8.8.8 elsewhere, and read the summary line at the end, which reports minimum, maximum, and average together with any loss.
Judge jitter and loss, not just the average. A change that lowers your average by 5ms but doubles your spikes has made the game worse. The summary line gives you both, which is exactly why it is more useful than the number in the corner of the screen.
Test the reverse too.If something helped, turn it off and confirm the problem returns. That one step separates a real improvement from a quiet evening on your ISP's network.
The short version
Measure first and note whether your problem is a high average or an unstable one. Ping your own router to separate house problems from everything else. Run a traceroute and find where the numbers jump. Check whether you are behind CGNAT and what your ports look like. Fix your own layer first, because a cable and working queue management beat any subscription. If everything local is clean and the path itself is the problem, then a VPN or a routing optimiser becomes a reasonable thing to trial, judged on your own before-and-after numbers rather than on anyone's marketing.