Why games lag on fast internet: ping, jitter and packet loss

Your speed test looks excellent. The download finishes quickly. Then you join a match and another player slides across the screen, your character snaps backwards, or an action lands later than you expected.

A fast connection can still deliver a laggy game. Download speed tells you how much data a connection transfers per second. Online play also depends on how promptly and consistently updates arrive, and whether they arrive at all. Ping, jitter and packet loss describe different parts of that experience.

Start with the game's own performance display, compare a quiet connection with a busy one, and try a direct Ethernet connection if you normally use Wi-Fi. Those comparisons give you a more useful next step than buying a faster broadband plan on the strength of one speed-test result.

This guide is for games running on your own Windows 11 PC and communicating with an online server. Cloud gaming streams the picture as well, so its bandwidth requirements are a separate question. The method below combines official networking documentation with a worked example; the example's numbers are illustrative.

What ping, jitter and packet loss actually tell you

The numbers answer different questions. You can improve one without fixing the others.

Cloudflare's explanation of speed testing separates throughput from latency and explains why the test's destination and method affect the result. A good result to its server cannot establish the quality of your connection to a different game's server.

Jitter also needs context. Different tools calculate delay variation differently; the IETF's packet-delay variation standard explains why the term is ambiguous. Do not subtract the lowest ping from the highest and assume you have calculated the same jitter value shown by every game or test.

Why a stable average can hide an unpleasant match

Suppose most updates arrive promptly, but occasional ones arrive much later. An average can make that connection look better than those interruptions feel. Watch the graph through the moment of the problem, rather than taking one reading afterwards.

Riot's 2022 explanation of VALORANT's latency and buffering describes another complication: a game can hold updates briefly to handle uneven arrivals. That buffering adds delay beyond the network RTT. The ping number therefore does not represent every part of the time between your input and what appears on screen.

Why a video can play smoothly while a game feels awful

A video player can store some upcoming footage before showing it. A live game has less freedom to wait: it cannot know the action you are about to take.

In its technical explanation of VALORANT's netcode , Riot describes the trade-off. Buffering helps smooth uneven arrivals, but also delays the information shown to the player. Missing or late updates can lead to predictions and corrections, which can appear as characters jumping or sliding. Different games handle these situations differently; a particular visual symptom is a clue, not a diagnosis.

Conceptual packet arrivals, not a network test. Original diagram by Jitsumi.

First, separate network delay from PC stutter

“Lag” gets used for several problems. Before changing the connection, enable the game's network statistics and its FPS or frame-time display if available. Frame time is how long each frame takes to produce; uneven frame times can make motion look jerky even when an average FPS looks respectable.

Riot's instability guide makes the distinction explicit: high average ping, low client FPS and server problems are separate indicators. Look for equivalent statistics in your game, using its support documentation for the relevant controls.

Use what changes during the symptom to choose the next investigation:

• If the picture hitches while frame times worsen, examine local game performance as well as the connection.
• If movement corrections coincide with ping spikes or packet-loss warnings, investigate the network path.
• If both sets of readings worsen, keep both possibilities open. An active download, for example, may involve both network traffic and work on the PC.

Do not treat smooth FPS as proof that the connection is healthy, or a low ping reading as proof that the PC is rendering smoothly.

Compare four conditions before changing settings

Start by checking the publisher's service-status page for your region. If it reports an incident that matches your problem, that is useful context. If it reports nothing, continue collecting evidence; a clear status page does not prove every connection is working normally. Riot includes a regional status check in its basic connection troubleshooting guidance .

Then choose a repeatable online part of the game that actually connects to a server. A practice session or non-competitive mode is preferable to experimenting in a ranked match, provided it uses the online connection. Keep the same PC, game, server region and comparable observation period throughout.

1. Record a quiet baseline

Pause large transfers you recognise: a game update, an upload of recorded footage, or an active cloud backup. Where practical, account for transfers on other household devices too. Leave ordinary settings alone.

Play the chosen scenario and note the ping range, any loss indicator, frame-time behaviour and the visible symptom. A short screen recording can help connect a graph spike to what happened in the game.

2. Compare Wi-Fi with direct Ethernet

Connect the PC to the main router with an Ethernet cable, temporarily turn off the PC's Wi-Fi, and repeat the same scenario. Restore your usual connection afterwards if needed.

This comparison removes the PC's wireless link from the test. A cable connected to a mesh satellite with a wireless connection back to the main router still leaves a wireless segment in the route. TP-Link's explanation of mesh backhaul illustrates the distinction between wired and wireless links between mesh units. Record that distinction if you cannot test directly.

Riot recommends a wired comparison because Wi-Fi interference can cause connection problems. If the game repeatedly improves on the direct cable, investigate the wireless setup before buying a faster internet plan. If it does not, keep looking. Ethernet cannot remove every source of delay beyond your home.

3. Observe the connection while it is busy

On each connection type, repeat the scenario while the same known upload runs, then pause it and test again. Use an ordinary transfer you control. Record whether the game's delay increases and whether it settles after the transfer stops.

If practical, run the upload from another household device so it does not also add disk or processor work to the gaming PC. Otherwise, keep watching the frame-time display.

This is a comparison, not a stress-test contest. You are looking for a repeatable relationship between activity and symptoms. Keep the transfer type consistent: changing from a download on Wi-Fi to an upload on Ethernet would change two variables at once.

4. Read the pattern, not just the biggest number

Illustrative example: the following values are invented to demonstrate the reasoning. They are not Jitsumi measurements or universal targets.

Invented values illustrate a diagnostic pattern, not a benchmark: both connections worsen during the same upload. Original diagram by Jitsumi.

In this example, quiet Wi-Fi and quiet Ethernet both have a narrow range. Both become much slower during the upload. The useful lead is the activity they share, so the next comparison is to pause or limit that transfer and repeat the test.

The example does not establish that the router is faulty, identify the exact place where delay occurs, or prove that a faster plan will solve it. Repeat the quiet and busy comparison before spending money. The result you want is a pattern you can reproduce.

Why uploads can make a fast connection feel slow

Network devices queue packets when data arrives faster than a link can send it onwards. Some buffering is normal. Excessive queues add waiting time, a problem often called bufferbloat. The IETF's recommendations on active queue management describe the relationship between queues and latency.

That explains why an upload is worth checking even when a speed test shows excellent download throughput. Sending and receiving are separate directions, and either direction can become a bottleneck. A game can have enough capacity to exchange its updates while still suffering from the time those updates spend waiting.

If pausing a transfer repeatedly improves play, a useful immediate option is to schedule it outside gaming time or use the app's transfer limit. For persistent problems, consult your router maker's documentation for supported queue-management features. Active queue management can reduce queueing delay, but a feature labelled “gaming QoS” is not a guarantee of a particular result.

A speed test that reports idle and loaded latency offers another comparison: delay when the connection is quiet versus delay during transfers. Cloudflare documents both measurements. Treat them as evidence about that test's destination and conditions, then compare them with the actual game.

Use Windows ping to add a local check

If the game comparisons leave you unsure, a small Windows test can help. It uses built-in tools and does not reset your network settings.

Find the gateway for your active connection

Open Command Prompt and run:

ipconfig

Under the adapter you are using, find Default Gateway. Ignore disconnected adapters and unrelated virtual connections. Microsoft's ipconfig documentation confirms that the plain command displays addresses and gateway information. For a typical home connection, the gateway is your router.

Send a short series of probes

If your actual IPv4 gateway is 192.168.1.1, run:

ping /n 30 192.168.1.1

Replace 192.168.1.1 with the gateway shown on your PC. The /n 30 option requests 30 probes; it does not promise an exactly 30-second test. Press Ctrl + C if you need to stop.

Microsoft's ping reference describes these as ICMP Echo requests. The replies show round-trip times to the selected destination. Compare the gateway results over Wi-Fi and direct Ethernet, ideally while observing the same symptom in the game.

For an outside comparison, you can use a destination specified by the game's support team. An optional public target is Cloudflare's 1.1.1.1 address :

ping /n 30 1.1.1.1

This probes that address; it does not change your DNS settings or measure your game server's ping.

Interpret missing replies carefully

A timeout says that no reply arrived within the allowed time. It does not automatically mean the same percentage of your game's traffic is being lost. Some devices limit responses addressed to themselves while still forwarding other traffic; Microsoft's pathping example illustrates that distinction.

If a target does not answer, leave the result inconclusive. If a short test shows no loss, keep in mind that an intermittent problem may simply not have occurred during that sample.

Use the results to choose a follow-up:

What to send your ISP or the game's support team

A useful report tells support what changed and what stayed the same. Include the game, server region, dates and times with your time zone, connection type, and whether the issue appears in other games. Add the quiet-versus-busy and Wi-Fi-versus-Ethernet comparisons, plus a short recording or relevant statistics if requested.

For example: “On the same PC and game region, the problem appeared over both Wi-Fi and direct Ethernet while an upload was running. It cleared when I paused the upload and returned when I resumed it.” That is more actionable than “my internet is fast but the game lags”.

If only one game or region is affected, say so. That narrows the investigation, but the game, its route and your local setup still need to be considered together. Send diagnostic logs through the provider's official support channel.

Should you buy faster broadband?

Extra capacity can help when your household regularly saturates the connection. It cannot guarantee a shorter route to a game server, steadier Wi-Fi, or a solution to a server-side problem. Use the comparisons above to decide what needs improving first.

The most useful outcome is specific: the game improves on a cable, the problem follows an upload, or one destination remains troublesome while other tests are steady. Once you can describe that pattern, you have a practical next step and a way to check whether a change actually helped.

Sources and further reading

• Riot Games: Basic Connections Troubleshooting Guide , 2 June 2026.
• TP-Link: Deco Ethernet Backhaul: Setup, Wiring Rules, and Troubleshooting , updated 17 April 2026.
• Cloudflare: How does Cloudflare's Speed Test really work? , 27 October 2025.
• Microsoft: ping , updated 1 November 2024.
• Microsoft: ipconfig , updated 3 February 2023.
• Microsoft: pathping , updated 1 November 2024.
• Riot Games: VALORANT Gameplay Consistency Update 2 , 24 May 2022.
• Riot Games: VALORANT Game and Network Instability Basics , 2 December 2021.
• Riot Games: Peeking into VALORANT's Netcode , 28 July 2020.
• Cloudflare: 1.1.1.1 launch announcement , 1 April 2018.
• IETF: RFC 7567, Recommendations Regarding Active Queue Management , July 2015.
• IETF: RFC 3393, IP Packet Delay Variation Metric , November 2002.