High ping and lag: where the delay actually comes from

Buying a faster internet plan almost never fixes lag. This surprises people, and it is the reason so much money gets spent on the wrong upgrade. Bandwidth is how much data you can move per second. Latency is how long a single packet takes to get there and back. A gigabit connection with 120ms of latency will feel worse in a game than a 50 megabit connection with 15ms, every time.

Lag is a latency problem, and latency has a small number of causes that you can work through in order. Some of them you can fix in five minutes. Some of them are physics and no amount of money will help.

What counts as good ping

Round trip time	How it feels
Under 20ms	Excellent. Competitive gaming territory.
20ms to 50ms	Good. Nobody will complain.
50ms to 100ms	Fine for most things, noticeable in fast games.
100ms to 150ms	Calls feel slightly delayed, games feel sluggish.
Over 150ms	Conversations start talking over each other. Games become guesswork.

Two numbers matter more than the average, and most speed tests hide both. The first is the maximum: an average of 30ms with occasional 400ms spikes feels far worse than a flat 60ms. The second is jitter, the variation between consecutive packets. Real time applications hate variation more than they hate delay.

The part you cannot fix

Light in fibre travels roughly 200,000 kilometres per second, and real routes are never straight lines. That gives every connection a hard floor based on distance alone:

• Same city: 5ms to 15ms
• Same country: 15ms to 40ms
• Across a continent: 40ms to 80ms
• Across an ocean: 80ms to 150ms
• Satellite in geostationary orbit: 500ms or more, no matter what

If your game server is on another continent, 130ms is not a fault. It is geography. Before you spend an evening tuning a router, check where the server actually is. A public IP and geo lookup on the server's address will tell you roughly where it sits and which network runs it, which is often the fastest way to discover that you are connected to a region on the other side of the world.

The practical move here is to change servers, not hardware.

Steady high ping vs ping spikes

These are different problems with different causes, and lumping them together is why people chase the wrong fix.

Steady high ping means every packet takes the same long route. Distance, a slow last mile technology, a VPN sending traffic the long way around, or routing that takes an inefficient path. It is consistent and predictable.

Ping spikes mean something is periodically getting in the way. A queue filling up, Wi-Fi retransmissions, another device saturating the link, a background update. The average may look fine while the experience is terrible.

Telling them apart takes one test: run a continuous ping with count set to 0, leave it going for a few minutes, and watch the per packet numbers rather than the summary. A flat line of 90ms is a routing and distance story. A line that sits at 25ms and jumps to 300ms every 20 seconds is a congestion story.

# Leave this running and watch the shape, not just the average
https://netdebug.app/open/ping?target=1.1.1.1&count=0

Finding where the delay is added

Latency accumulates hop by hop, so a traceroute shows you where it gets added. Read it as a series of steps and look for the jump.

A healthy trace climbs gradually: 2ms to your router, 12ms to the ISP, 14ms, 18ms, 22ms across the backbone, 24ms at the destination. When you see 2ms, 12ms, 14ms, then suddenly 95ms and everything after that stays around 95ms, the delay was added at that hop and everything before it is fine.

Two things to keep in mind while reading:

• A single slow hop that does not carry forward is noise. If hop 6 answers in 200ms but hops 7 through 12 answer in 30ms, hop 6 is just slow at replying to probes. Your traffic passed straight through it.
• Asterisks are usually not failures. Plenty of routers simply do not reply to traceroute probes. If the hops after them answer normally, nothing is wrong.

What you are looking for is a step change that persists to the end. That is where the time is really going.

The fixes, in order of how often they work

1. Get off Wi-Fi, or fix it

   Wi-Fi adds latency and, worse, adds variable latency. Retransmissions, channel contention with neighbours, and power saving on the client all show up as spikes. A cable eliminates the variable in one move. If a cable is not possible, move to 5GHz, get closer to the access point, and check how many networks are fighting for your channel.

2. Find out what else is using the connection

   A cloud backup, a console downloading an update, someone else's video upload. Uploads are the usual offender because home upload capacity is small. Run a continuous ping, then start a large upload deliberately and watch what happens to the numbers. If latency jumps from 25ms to 400ms the moment the upload starts, you have bufferbloat.

3. Turn on smart queue management

   Bufferbloat is a solved problem. Routers with SQM, fq_codel or similar keep latency low even when the link is full. If your router has the option, enable it. If it does not, that is a genuinely good reason to replace it, and a far better upgrade than more bandwidth.

4. Check whether a VPN is in the path

   A VPN endpoint in another country adds the round trip to that country onto everything you do. Sometimes it also improves routing and lowers ping, which is why gaming VPNs exist, but the default outcome is worse. Test with it off, then with it on, against the same target.

5. Pick a closer server

   Games, voice platforms and cloud services almost always let you choose a region. This is the highest leverage change available and it takes ten seconds.

6. Look at the last mile technology

   Satellite and older DSL have latency floors that no configuration can remove. Cable and fibre are much better. If you are on geostationary satellite, sub 100ms is not available to you at any price.

Jitter, and why calls break before games do

Jitter is the variation between consecutive packet arrivals. If packets take 30ms, 32ms, 29ms, 31ms, your jitter is tiny and everything feels smooth. If they take 30ms, 95ms, 28ms, 140ms, the average is still acceptable and every voice call will sound like a robot.

Voice and video applications compensate with a jitter buffer: they hold audio for a few tens of milliseconds so late packets have time to arrive. When jitter exceeds the buffer, the packet is treated as lost and you hear a gap. That is why a connection can pass every speed test and still fail a Zoom call.

Under 5ms of jitter is excellent, under 20ms is workable, and above 30ms you will hear it. The speed and latency baseline reports jitter alongside throughput, which is the number to quote when someone insists the connection is fine because the download speed is high.

Build a baseline before you need one

The most useful thing you can do is measure your network while it is behaving. Run a latency and speed baseline once a day for a week. When something goes wrong later, you will know whether 60ms is normal for you or double what it should be, and you will be able to show a support agent a chart instead of an anecdote.

The short version

Lag is latency, not bandwidth, so a faster plan rarely helps. Check the distance to the server first, because that sets a floor you cannot beat. Then decide whether you have steady high ping (routing and distance) or spikes (congestion and Wi-Fi), because the fixes are completely different. And watch jitter, not just the average.