How to Lower Ping in Games (Measured, Not Guessed)

Most advice on lowering ping is written by people who have never measured it. You get a list — close background apps, restart your router, buy a gaming VPN, upgrade your internet — with no indication of which of those does anything.

So I measured all of it. Ten-minute packet captures, three connections, before and after each change. Here is what moved the number and what did not.

The three numbers that matter

Before anything else, understand what you are trying to improve, because "ping" is used to mean three different things.

A flat 70 ms is a genuinely fine gaming connection. A 30 ms average with 45 ms of jitter and 1.5% loss is a miserable one, and the person with it will swear their ping is great. If you only remember one thing from this article, make it that.

Our ping quality checker grades all three against the genre you actually play, which is more useful than a single generic threshold.

Step 1: Measure before you change anything

You cannot tell whether a fix worked without a baseline. This takes ten minutes and it is not optional.

Open an elevated Command Prompt and run a sustained ping to a stable host:

ping -n 600 1.1.1.1

That is 600 packets, roughly ten minutes. Let it run while you do something normal — have a game open, let Windows do its thing. A ping test on an idle connection tells you nothing about how it behaves under load.

When it finishes, read the summary:

• Average is your baseline latency.
• Maximum minus Minimum gives you a rough sense of jitter. If the gap is more than about 30 ms, you have a jitter problem.
• Lost should be 0%. Anything above 1% on a wired connection is a genuine line fault.

Step 2: Get off Wi-Fi

I know. Everyone says this and nobody wants to hear it.

The reason it matters is not raw speed — modern Wi-Fi is plenty fast. It is that wireless is a shared, contended medium. Your card has to wait for a clear moment to transmit, retransmit anything that collides, and share airtime with every other device and every neighbouring network on the same channel. All of that is variable, and variable is exactly what games hate.

Here is what I measured on the test laptop, two rooms from the router, 5 GHz, nothing else heavy on the network:

The average difference between Ethernet and 5 GHz is 8 ms, which you would probably not notice. The jitter difference is 6x and the maximum is 7x, which you absolutely would — those are the moments where your character teleports.

If you genuinely cannot run a cable:

• Use 5 GHz, never 2.4 GHz. The 2.4 GHz band is shared with microwaves, Bluetooth, baby monitors and every neighbour within range.
• Set a fixed, clean channel rather than Auto. Use a Wi-Fi analyser app to find the least crowded one.
• Get the PC in line of sight of the router if at all possible. Walls, and especially anything with foil-backed insulation or plumbing in it, are brutal.
• Powerline adapters are usually worse than Wi-Fi. MoCA over existing coax, if your house has it, is genuinely close to Ethernet.

Step 3: Kill bufferbloat — the biggest hidden win

This is the fix almost nobody mentions and it is the one that transforms connections.

Your router has a buffer. When more data arrives than the line can send, the router queues it instead of dropping it. That sounds sensible, and for file transfers it is. But when that queue fills with a big upload, your game's packets go to the back of the queue and wait — adding hundreds of milliseconds of delay that has nothing to do with distance.

This is why your ping is fine until someone in the house starts a video call, and why it spikes when a game is patching in the background.

The fix is Smart Queue Management (SQM), which uses an algorithm called fq_codel or CAKE to keep the queue short and prioritise small, latency-sensitive packets over bulk transfers.

How to enable it:

1. Run a speed test and write down your actual download and upload figures.
2. Log into your router. Look for SQM, Smart Queue, QoS, Bufferbloat or Adaptive QoS depending on the brand.
3. Enable it and choose fq_codel or CAKE if you are offered a choice.
4. Set the download and upload limits to about 90% of your measured speeds. This is the part people get wrong — you must leave headroom, otherwise the queue forms upstream at your ISP where your router cannot manage it.
5. Re-run the bufferbloat test.

If your ISP router does not support SQM — many do not — a cheap router running OpenWrt in front of it, or a Ubiquiti/MikroTik box, will. It is the best money you can spend on your gaming connection, and it costs less than a year of a gaming VPN.

Step 4: Choose your server region manually

No amount of tuning beats physics. A packet travelling to a data centre 3,000 km away cannot arrive in under about 30 ms, and in practice fibre routing makes it 45 to 60 ms.

Most games auto-select your region, and auto-select is frequently wrong — it picks based on a one-off measurement at login, or on a geolocation database entry that is out of date, or on population balancing rather than latency.

Go into the network or matchmaking settings of your game and check what it chose. If there is a closer data centre listed with a lower ping, pin it manually. On our VDSL connection, VALORANT had auto-selected a region 900 km further away than the nearest one, for a 19 ms penalty we did not need to be paying.

Step 5: Remove local interference

Now the smaller stuff. Individually these are worth single-digit milliseconds; together they are worth measuring.

Pause anything uploading. Cloud backup, Steam downloads, OneDrive sync, a torrent client, a phone backing up photos. Uploads hurt latency far more than downloads because the upstream link is narrower. If you did Step 3 properly, this matters much less — that is the point of SQM.

Turn off the overlays you do not use. Discord, Steam, NVIDIA App and Xbox Game Bar overlays each hook the game's render path. They do not affect network latency, but they add input-to-photon latency, which people commonly report as "lag". Kill the ones you never open.

Disable Windows Update delivery optimisation. Settings → Windows Update → Advanced → Delivery Optimisation → turn off "Allow downloads from other PCs". By default Windows uploads updates to strangers using your upstream bandwidth.

Check for a metered or throttled adapter. Some laptops have power-saving settings that throttle the network adapter on battery or under thermal load. In Device Manager, open your adapter's Properties → Power Management and disable "Allow the computer to turn off this device to save power".

What does not work

I tested these because readers keep asking about them.

The DNS one surprises people. Public DNS like Cloudflare or Google genuinely helps with connection errors and initial lookups — it is one of the first fixes for Roblox connection codes, for instance. But once you are in a match, DNS is not in the path any more. It cannot affect your ping.

When the problem is your ISP

If you have done all of the above and you still have high ping or loss, run a traceroute and find out where it breaks:

tracert -h 20 <game server IP or a host in that region>

Look at where the latency jumps. If hop 2 or 3 — inside your ISP's network — is where the number balloons and never recovers, that is their fault, not yours. If loss appears at one hop but not the ones after it, ignore it: many routers deprioritise responding to ping and it is not real loss.

When you call your ISP, do not say "my game lags". Say: "I have sustained packet loss of X% and latency to Y measured over ten minutes, here is the traceroute." That gets escalated past the script.

Realistic expectations

Here is roughly where you should end up once everything above is done properly:

If you are inside those ranges and the game still feels bad, the problem is not your connection any more — it is your frametimes. Steady frame delivery and low input latency account for a lot of what people describe as lag, and those are a different investigation entirely.