Field note · Updated · Jay Fan
A resolver clock and a game-path clock answer different complaints.
Changing DNS can speed name lookups and app updates; it does not rewrite the path a live game already has to its server. A slow resolve is the wait to learn an address. A large delay figure, or a loss mark, is what the game shows after some address is already in use. Separate those clocks before you edit a resolver field. This desk has not timed a resolver, has not timed a game path before and after a resolver swap, and has not kept a capture that pairs the two.
Symptom
What does a slow name lookup look like before a match exists?
A lookup is the short exchange that turns a hostname into an address. It runs when a launcher, a store, a patch client, or a console asks for a name, and again when a cached answer expires. The device usually asks one recursive resolver: the address configured on the adapter, or the address the router handed out. That resolver may have to ask further, and any cache along the way can answer instead. You rarely see that chain from inside a game. You can see whether the resolver you think you are using answered, failed, or never got a fresh question because a previous answer was still stored.
The complaint that fits this clock shows up before a match is on screen. The launcher sits on a connecting or updating line and never draws a server list. A patch listing stalls before any of the patch itself moves. A console system update or a game-store catalog spins on the name and has not started moving bytes. A hostname fails while an address you were already allowed to try still answers. Those are reasons to read the resolver. They are not, by themselves, a reading of round-trip time on a live game flow.
An update belongs on this clock only while the client is still asking who to download from. After an address is in hand, the bytes travel on the path to that host. A quicker answer can start the transfer sooner. It does not add capacity to the path the bytes then take. If the client is handed a different address than last time, write that as a different host. A different host is a different fact from a shorter trip to the old one. “The update felt snappier” is not a finished row until you know which of those two things moved, and this desk has no row of either kind to lend you.
Stored answers hide the swap. If the device, the router, or the resolver already holds an answer, editing the resolver field does nothing until that answer expires or you clear it on purpose. A failure can be stored too. A pass that never checked the cache can credit a resolver for a calm minute that was only a remembered address, or blame a resolver that was never asked. If you did not look, leave the cache cell blank. Do not write that the lookup was instant, and do not write zero.
Browsers and some system features can send lookups to a resolver of their own, over an encrypted channel, while the game keeps using the adapter’s resolver. Changing one of those does not change the other. The device that is about to launch the game is the one whose field matters. A phone, a PC, and a console in the same house can be pointed at three different resolvers, especially when one of them has a manual entry and the others still take what the router offers. Read the screen on the machine that will play. A change made on a laptop does not describe the console.
What stays on the game path once the client already has an address?
Inside a match the client is already sending UDP or TCP to an address the session gave it. Often that address arrived from a matchmaker over a connection that was already open, not from a fresh question to your recursive resolver on every input. The delay figure is the round trip toward that game server, or the client’s own estimate of it. The loss mark is data that never arrived. Neither counter is the time a resolver spent answering a name.
Swapping the resolver while that address stays the same does not redraw the route. Game packets do not detour through the resolver on their way to the server. A resolver answers questions. It becomes part of the game flow only if you also turned on a tunnel, a proxy, or a product whose own page says it forwards game traffic. Read that page before you call the product a DNS change. A setting that only chooses who answers name questions is still a resolver, and this experiment stops at that boundary. If the product moves the packets, you are on a different path, and the row has to say so.
The one way a resolver change can touch a later path is by handing the client a different destination. Some names are answered with an address that depends on where the question was asked. A new resolver can return a different address, including an address in the other family, IPv4 or IPv6, which the client may then prefer. Confirm that by reading the address, the region, or the host the client shows. You have a new server selection. You do not have a measurement that the old path got shorter. If the address did not change, the path was not rewritten, and a calmer next match is a different session until you can show otherwise. Keep the server, realm, or host name in the row when you judge the delay figure.
A loss mark inside a match that already connected sits on the same side of the fence. Missing game packets are not a slow answer from a resolver you are no longer asking. The delay-and-loss note is where a flat number, a jump, and a hole get their own rows. This page does not repeat those shapes. If you change the resolver and the server picker in one sitting, both rows are void. Write the void down. A clean story that drops the spoiled pass is how a resolver gets credit for a matchmaker.
This desk has not measured any of that split. There is no timed query from here to an ISP resolver, to a router-forwarded resolver, or to a public resolver address. There is no before-and-after delay figure. There is no before-and-after loss mark. There is no capture that puts a lookup beside a live game flow. The August 19 web-edge record timed ICMP and TLS toward public and publisher websites. A DNS hostname appears in that file as a destination of those probes. The file did not time a recursive lookup, and it did not time a game server. Citing it here would borrow the wrong clock.
Reading guide
Which clock is the complaint actually using?
Use one row. These cells are a reading guide, not a timed session from this site. If your own copy of the log has an empty cell, you did not read that clock. An empty cell is not a zero, and it is not a pass.
| Complaint you can point at | Clock that matches | A resolver swap may explain | A resolver swap is a weak story |
|---|---|---|---|
| The name never resolves, and a permitted address still answers | The query to the resolver, including a stored failure. | The old resolver failed, was blocked, or was never asked because a cached answer or a cached failure was still in the way. | A delay figure on a match that already had an address. |
| The launcher or patch is slow to start, then the transfer itself drags | Lookup first. Then the path to whatever address came back. | How long the name took, and whether the address you were handed changed. | The transfer’s own pace after the address stayed the same. |
| The in-match delay figure is large or jumpy, and the session address did not change | Round trip to the game server already chosen. | Nothing on that path. | Treating the HUD number as if it were lookup time. |
| A loss mark, a voice hole, or a position snap inside a connected match | Missing game packets on the path already in use. | Nothing, while that path stays put. | A resolver that had already answered before the match. |
| The answer changed, and so did the region or host | A new destination, not a rewritten old one. | Which address you were handed, including a change of address family. | A claim that the previous path was repaired. |
Two clocks can be ugly in the same hour. A launcher that cannot resolve a patch host can share an evening with a match that is simply far away. Write both. Clearing the name does not clear the delay figure, and a screenshot of the delay figure does not say whether the name was slow.
What do you leave blank until that session exists?
Copy the header into your own notes. Fill a row only after the pass. The first body row is a label so the columns have a meaning. It is not a household test. Pass 1 and pass 2 start empty on purpose. A cell you leave empty is unread. Do not turn it into zero later, and do not let a spreadsheet turn it into zero for you. A zero on a loss mark would mean you looked and nothing was missing. A blank means you do not know. Those sentences point at different next checks.
| Pass | What had not connected yet | Resolver you actually read | Name asked | Answer | In-match delay figure | Loss mark | Address unchanged? | Void? |
|---|---|---|---|---|---|---|---|---|
| Example shape | Launcher, update, or already in a match | The address in the field, or unread | The hostname, or none yet | Answered, failed, cached, or unread | Still large, still jumpy, calmer, or not in a match | Quiet, ticking, or not shown | Yes, no, or not checked | Yes if two things changed |
| Pass 1 | ||||||||
| Pass 2 |
What this desk has not measured, so the blank rows cannot be mistaken for our results:
- No timed query to an ISP resolver, to a router-forwarded resolver, or to any public resolver address.
- No before-and-after in-game delay figure around a resolver swap.
- No before-and-after loss mark around a resolver swap.
- No capture that places a lookup beside a live UDP or TCP game flow.
- No ranking of resolver operators, and no claim that a particular address is the right one for your line.
Troubleshooting steps
How do you swap one resolver without spoiling the row?
- Write the complaint in one sentence before you open the setting. The name will not resolve. The update will not start. Or the match number is the complaint. Only the first two put the resolver in the lead. The third belongs on the delay-and-loss note, unless the only thing you are checking is whether the address changed.
- Read the resolver the playing device is using now. A router can hand out an address you never typed. A browser can use a different resolver from the game. Copy what the field on that device shows. If the screen will not show it, leave the cell blank and say so. Inventing a vendor is how a log becomes fiction.
- Hold the game, the account region, and the link still. Change only the resolver, in the field the system or the console already offers. Menu names move between updates, so trust the screen in front of you rather than a screenshot from another firmware year. Do not also restart the modem, move to a cable, and start a tunnel. A restart clears caches and drops sessions, which is a second change even when you meant it as housekeeping. The cable comparison is the Wi-Fi note. A tunnel replaces the path and often replaces the resolver too, so it voids this row on contact.
- Ask the same name again and write the answer. If the complaint was the launcher or the update, see whether that step proceeds, and keep the transfer’s own pace in a separate sentence from the lookup. If the name still fails, you have a lookup result. You do not yet have a comment on the match. If the complaint was an in-match number, this pass did not speak to it. Stop.
- Only for the “different address” case, play a comparable moment and read the region or host again. Address unchanged means the game path was not rewritten. Mark the row as a lookup test. A calmer HUD on a new host did not measure the resolver’s effect on the old path. If you cannot see the address, write “not shown” and do not upgrade that into “unchanged.”
Void the row if a game update ran during the match comparison, if the matchmaker moved you, or if two settings changed. Keep the void row. The next pass starts clean. This page does not name a resolver to install. Whatever address you are about to write in, the log is the point. We have not timed the common public ones, and we have not timed the one your router already offers.
Command lines for a Windows gateway echo and a public control already sit on the methodology page. This note does not repeat them. An echo aimed at a resolver’s address times that echo. It does not time a DNS query. Microsoft’s ping command page treats connectivity and name resolution as related troubleshooting, and the times a completed ping prints are round trips of the echo. A line that cannot find the host is the lookup failing. Those are different lines of output, and neither line is your match. A shorter statement of the same fence is on the symptom page. This note is the worksheet.
When to switch measurement method
When do you leave the resolver and read the game path instead?
Leave as soon as the match exists and the HUD is the complaint. Open the delay-and-loss note and name the shape you can point at: flat and large, a jump that returns, or a hole. The resolver field will still accept a new address if you type one. That edit is not the instrument for a number that is already on screen, and it is not the instrument for a loss mark on a session that already connected.
Leave for the evening-queue note when the delay figure jumps only while a backup, a call, or an install is sending. A resolver does not sit in that queue. Pausing the transfer is the matching change, and it should not share a pass with a resolver swap. If the install is still stuck on the name, you are not in that note yet. Finish the lookup row first, then decide whether any bytes moved.
Leave for the Wi-Fi note when the only question left is the last hop inside the house. A cable comparison mixed into the resolver pass makes both notes useless. Plenty of console sessions have no cable at all. Write that limit in the row rather than inventing a wired pass you did not run.
Leave for the console NAT note when the invite fails, voice is one-way, or a strict label is the sentence you can defend, and names are already resolving. A resolver that answered is not an inbound filter. The NAT letter will not explain a delay figure either. If you are still on the “which clock” table above, the party-join row is not on it. That is deliberate.
Stay on this page when the name itself fails, or the launcher never gets far enough to show a delay figure, and you have not yet done one resolver change with the answer written down. Stay when you suspect the new resolver handed you a different host, until both addresses are in the log. Stay when the playing device and the router disagree about which resolver is in use. That disagreement is still a lookup fact.
Leave network checks entirely when the picture hitches and every network mark the game shows stays still. A resolver cannot draw a frame. Valorant, Tarkov, WoW, and Genshin are where a specific HUD word matters. Use the guide for the label on that client. Do not treat those pages as a second copy of this fence between a lookup and a path.
Do not cite the August 19 file as a resolver timing or as a game-path timing. It remains one web-edge record. A private first sort, before you spend the evening in adapter settings, is still Diagnose lag. Pick the symptom that matches the sentence you wrote. The tool does not read your resolver, and it does not store the answer.
Which questions keep getting filed under ping?
Does a faster resolver lower the delay figure inside a match?
No. The delay figure times the path to the game server the session is already using. The resolver is not on that path. A quicker lookup can matter later only when it hands the client a different address, which shows up as a different host rather than a shorter trip to the old one.
Can changing DNS help a launcher or a game update?
It can, when the stall is the lookup itself or a resolver that never answers the name. Once the download host is known, the bytes use the path to that host. Time the lookup and the transfer as two rows. A quicker start is not a thicker path.
Why can a hostname ping succeed while the match still feels late?
A completed ping prints the echo round trip after the name has already become an address. That column is not the lookup, and an echo to a public address is not the game server. A line that says the host could not be found is the lookup. Microsoft documents ping as a connectivity and name-resolution check, and the times it prints are those round trips.
If the new resolver returns a different address, was the old route fixed?
You changed destinations. Write both addresses and the region or host the client shows. A calmer match on the new address is a server change. It is not a measurement of the old path, and this desk has not published either timing.
Has this desk timed a resolver against a game path?
No. We have not published a lookup time, a before-and-after delay figure, or a capture that pairs the two. The tables are reading guides. An empty cell in your copy is not a zero. The August 19 file is a web-edge record, not a resolver trial.
Editorial log
November 5, 2026. First publication by Jay Fan. No resolver timing and no game-path capture are attached. The tables are reading guides. Corrections: [email protected].