What is the normal ping value for a game? How to reduce high latency? I pinged every segment of internet at home

the night I played Honor of Kings, my teammate kept saying three times in a row, "You're lagging and still not moving," causing peak latency to jump from 68ms all the way up to 240ms. When I exit, I test speed; 300Mbps broadband runs fully loaded, download speed is 29MB/s. At first, I couldn't figure it out either: the internet speed was clearly fast, so why was the game still laggy? Later, I realized that speed tests are about bandwidth, while gaming is about latency—these two metrics are completely different . Bandwidth is how wide the road is, while latency is how long it will take to travel that road. No matter how wide the road is, couriers can take a long route to deliver for three days, and you still won't get your goods.

to figure out where I spent that 240ms, I pinged my home network from computer to game server segment by segment, recording numbers in each segment. This article is my record sheet. To put the conclusion first: most families waste their delays on the indoor segment, which can be cut off in half without spending a single cent.

latency, jitter, and packet loss are three different things—first identify which type you are stuck

many people confuse "cards" together, but when broken down, they are actually three different diseases:

latency (ping value) : How long does it take for data to travel back and forth from your computer to the game server? When the road is long and there are more people lining up, the number of people gets bigger. This means the skill only appears after half a second of pressing it. jitter : Latency is unstable—20ms this time, 180ms next, sometimes fast, sometimes slow. This is shown as characters teleporting and being pulled back after taking two steps. packet loss : Data packets are lost halfway and the game server hasn't even received your actions. This manifests as being frozen for a few seconds and intermittent voice messages.

three types of diseases and three treatment methods. Most packet loss is due to line quality or the optical modem. I previously wrote about how to judge light degradation; The jitter and delay mostly come from inside your room. The way to tell is simple: check the latency numbers in the game—if it consistently shows around 150ms, it means latency is high; Repeatedly jumping between 40 and 200 indicates strong shaking; The numbers aren't high, but the character's teleport freezes—it's a packet loss.

my 240ms that night belongs to the first type of clip, the second type: base latency is relatively high, and it jumps back and forth during evening peak hours. Keep reading to see how I pieced it out piece by piece.

ping Value Grading Ruler: Normal Amount, Amount to Check

ping numbers need a reference to judge them. I combined my own measurements with seven or eight machines I had checked for friends, and compiled them into a gear splitter ruler:

for gaming, the feel can be roughly divided up: you can easily play within 60ms; 60~100ms is normal, and most people don't notice any difference; At 100~150ms, skills start to feel a bit sluggish; Movement drifts above 150ms, reaction is half a beat slower. So don't blindly believe in single-digit latency. Dropping from 80ms to 20ms won't help your rank; what really needs to be fixed is going from 150 to 80.

one more word about the cross-border channel. Optical fiber signals can only travel about 200 kilometers per millisecond, and a round trip to the US West Coast server takes over 100 milliseconds. Adding dozens of routers forwarding the message, 200ms is the line drawn by the laws of physics. Anyone who claims to reduce delays from China to North America to 50ms is just scamming money.

pinging segment by segment: three segments of positioning, which segment is really stuck on

High latency isn't scary; what's scary is not knowing which section is high. My approach is to divide the entire road into three sections, test each section separately, and repair any section that exceeds the standard.

first paragraph, inside the house: ping your router. hold Win+R and enter cmd, then type ping 192.168.31.1 -n 50 (change this IP to your own gateway address, check the sticker on the back of the router). My computer's network cable averages 2ms; Switching WiFi to the router in the same house averages 9ms; Walking to the master bedroom separated by a wall takes an average of 16 seconds. All of this is considered normal. If the network cable pings the gateway over 10ms, restart the router first. If it's still high, the machine itself can't handle it—my old AC2100 can ping up to 40ms when overloaded. This is the other side of old devices taking up the and not giving way—when the device is busy, the router's CPU is the first to give up.

second segment, line: ping operator node. I ping the telecom node on average, it's 13ms. Subtract it from the first segment: 13-2=11ms, which is the real overhead from my router to the operator's data center, healthy. If this section speeds above 50ms, everything at home would be useless. The problem lies in the line—first check the optical modem's power reception power ( light decay exceeds -28, you should report it for repair), then rule out whether it's a secondary broadband in the community. Late-peak traffic jams are structural, and complaining is useless.

third paragraph, remote end: ping game servers or major sites. ping baidu.com average is 19 ms, almost the same as the second segment, indicating a smooth route to the domestic major station. Game server IPs generally cannot be ping (they may be blocked), but the latency shown in the game can be compared: I show 88ms, the game server is in another province, 13ms within the province + tens of ms cross-province + server processing, accounts match. If all three stages go smoothly but the game is still laggy, then the game's own server has crashed. Wait until you play again.

here's a few tips: -n 50 averages out of 50 packets, much more accurate than the default 4. Just look at the line where "average = " is good; There's also a TTL value in the return. Around 64 means the other party is a Linux device (routers are all for them), and 128 is Windows. This is not used regularly, but if you make a mistake, it can help you confirm whether the ping is the machine you wanted.

WiFi is the big delay seller: the account of the blank queue

After

three-stage test, I realized one fact: I couldn't change those 11ms outside the province—the real bulk was inside the house. The network cable pings the gateway at 2ms, the WiFi wall at 16ms, and just the WiFi segment alone takes 14ms, and the evening rush hour is even worse—once at 9 PM on Tuesday, I was ping the gateway in the master bedroom, averaging 34ms, with heavy shaking.

reason is not complicated. WiFi is a "walkie-talkie" rule: everyone on the same channel takes turns speaking, and before speaking, they listen to see if someone is speaking; if someone is speaking, they wait. Your router, your neighbor's router, or the camera downstairs—as long as they're crowded on the same channel, they're all in line. 2.4G has only 13 channels, and the only ones that truly do not interfere with each other are 1, 6, and 11. With dozens of routers in a building squeezed into three lanes, how could it not be congested during evening rush hour? That's why latency during evening rush hours doubles: it's not that your home internet is slow, but that the entire "empty road" is jammed.

to deal with the queue for the empty port, I arranged the queue based on the effect: First, plug in the network cable, and desktop devices bypass the queue. After plugging in the desktop, the ping gateway stays steady at 2ms—one step; Second, phones and tablets connect to the 5G band. The many 5G channels and weak wall penetration actually become advantages—the signal can't get through your wall, neighbors can't get in, and the specific choice between 4G and 5G 2 already discussed; Third, manually nailing a clean router channel, scanning a loop with WiFi Analyzer to check for wave valleys. Android can view it, but Apple doesn't have permission and can only experiment. The method is in the of the channel selection; Fourth, don't force the signal in very poor corners. Queueing for retransmission below -75dBm doubles the cost. First, use blind spot self-check to confirm, then consider moving the router or adding a Mesh node. The wireless backhaul sub-router will add a few milliseconds on its own, but it's much better than the dead zones at -75.

shake hurts games more than high latency: sometimes fast, sometimes slow, the real culprit

This is the biggest takeaway from this investigation. Helping my cousin check latency: his dorm ping game server averages 120ms, which I find too high, but he actually manages to play; My average was 88ms back then, but my teammates complained about the lag. After comparing the 50 packages from both brands, I realized the difference: his 50 bags were neatly arranged in 115~125ms, while mine fluctuated between 35~210ms.

this is the difference in jitter. Average 88 but jitter ± 90, much more uncomfortable than the average 120 jitter ±5. game logic follows a fixed rhythm, and the stable 120ms can delay your operation by 120ms, still running smoothly; If you keep moving fast and slow, your rhythm won't match, and the server will just wait for you. The show is teleportation, pullback, and skill key swallowing. People are adaptable to stable high latency and have zero tolerance for instability.

Jitter Diagnostics and latency are different in direction, focusing on three main points: first, WiFi uninterrupted port congestion (using the system above); Second, when the router is fully loaded, with too many devices or someone downloading while running, the CPU is overwhelmed and forwards packets one by one. Try first using QoS to hold the of the big downloader . When someone at home is watching a live stream, I don't turn on QoS and it shakes; when I turn it on, it stabilizes immediately. Third, poor network cable contact. Oxidized connectors can cause scattered packet loss and delayed spikes. The cable in my study that was caught by the door was one that disappeared after replacing the connector.

order to reduce delays: start with free things, then spend less

organize what I learned this time into a sequence table and follow the instructions from top to bottom, testing one step at a time:

Finally, let's talk about accelerators. Its principle is to have you take the dedicated lines it rents and bypass congested public routes, which has an immediate effect on cross-border games (foreign server alliances, Steam international servers), reducing speed from 200ms to 120ms; But most domestic game servers are in your own province, and if you use a normal route, it's usually only 20~30ms. The savings the accelerator can make are all within the margin of error. So first, clean up the house according to the zero-cost action in the table, then consider accelerators for cross-border demand—don't reverse the order.

240ms that night, I found out two things had combined: weak signal from the master bedroom WiFi wall, plus the neighbor's router was squeezed into the same channel as me. The network cable solves the former, and the channel is moved to 11 to solve the latter. Now the master bedroom game is stable at 71ms, with ±8 jitter, which is enough.