The Game Feels Sluggish, Enemies Are Too Fast, and You Die in a Millisecond

Everything about latency. This section is mainly user/consumer discussion. (Peer-reviewed scientific discussion should go in Laboratory section). Tips, mouse lag, display lag, game engine lag, network lag, whole input lag chain, VSYNC OFF vs VSYNC ON, and more! Input Lag Articles on Blur Busters.
rusihhh
Posts: 49
Joined: 28 Jun 2025, 04:57

Re: The Game Feels Sluggish, Enemies Are Too Fast, and You Die in a Millisecond

Post by rusihhh » 07 Jul 2025, 00:34

iwabik wrote:
06 Jul 2025, 12:08
...
iwabik, thanks for the real contribution — your graphs are actually impressive.
All 5 games fall into an unusually consistent range, only minor shifts.

May I ask:
– What was your average ping during these matches?
– Were all games played on the same server/region on FACEIT?

If none of these games had noticeable issues or advantages, perhaps your captured avg packet size range is close to a "baseline" for stable conditions?

Would you be willing to either:
– Switch to a different server/region (with noticeably higher/lower ping) and record the packets there,
**or**
– Record a session when you *do* clearly feel problems or advantages?

Also, it would be awesome to see a **percentage distribution** of packet sizes — for example:
what % of packets fall into ranges like 250–500, 500–750, etc.
That could reveal more subtle differences between "good" and "bad" sessions.

If there's any chance you're able to share the `.pcap` files (even a sample) —
I’d really appreciate it via private messages, *if you're comfortable with that*.

One question though — are you confident in this analysis method overall?
Achieving meaningful **byte-for-byte or packet-for-packet** similarity is extremely difficult even under controlled lab conditions or in custom simulations.
So any patterns we observe might be fragile or easily distorted unless we interpret them carefully.

User avatar
ChristophSmaul1337
Posts: 111
Joined: 11 Feb 2024, 21:01

Re: The Game Feels Sluggish, Enemies Are Too Fast, and You Die in a Millisecond

Post by ChristophSmaul1337 » 07 Jul 2025, 05:20

I've read through this as I've seen my videos being linked in here.
Slender wrote:
06 Jul 2025, 15:09
quiz!
Person solved this problem,
Person is me. It has indeed been solved, but not with a clear-cut answer like suggested here. I'm still not sure what exactly fixed it and why it did. PSU is definitely not the only factor, the "wire" you're talking about has since proven to be almost useless unless in very specific circumstances. PSU I can replace with any other one at the moment without any problems occuring. I'm still in the process of finding out what's happening. Please don't jump to conclusions for my case. Thanks.

I do however know for a fact it's not internet related and it's also highly unlikely that it is for other people who see 0% loss, normal ping and so on. So here's my 2 cents for the OP and the proposed issue.
rusihhh wrote:
28 Jun 2025, 09:47
Some games dynamically adjust the size of the packets they send to you
Correct. A multiplayer video game is a constantly varying environment, so how much information your client needs to receive changes from instant to instant. One frame, a lot might be going on in the world, meaning your client needs a lot of new information, so you're receiving this information from the server. At another time, there's maybe not much going on in the world, nobody is moving or shooting, so you don't require many updates, so the packet size is smaller. That's totally normal and how any game does it. Nothing to see here.
rusihhh wrote:
28 Jun 2025, 09:47
based on their evaluation of your connection quality.
There's no reason to do this. Servers don't care about the client's connection quality in the way you think. The server will blurp out packets for you just as it does for everyone else. If packets are lost along the way, that's a "you problem" in the eye of the server. Adjusting packet size according to the network quality has no benefits for the server whatsoever. All such a mechanism would do is to add another layer of computation to the server's workload, and yield no benefits to the server's side. So... no. Not at all how that works.
rusihhh wrote:
28 Jun 2025, 09:47
If the game thinks your connection is stable, it may send larger packets that contain more data (e.g. 10 gameplay frames worth).
Also not how that works. Each packet you're receiving from a game server doesn't contain multiple gameplay frames. Rather, each packet is one gameplay frame. Sure, each frame might contain more or less information, but as explained earlier, this is roughly proportaional to how much is going on in the game at any given time.

A game server usually operates on fixed tickrates, for example in CS:GO the server would send out 64 updates every second. The server itself might run the game simulation (locally on the physical hardware) at much higher framerates, sometimes with thousands of FPS, but the game state gets sent out to the clients only once every 1/64th of a second. That's why a game's tickrate is so important. Imagine each packet you receive like one "server frame", like the server is "live streaming" its view of the game to you at 64 FPS.

This narrative of "the server pushes more data when conditions are right" is reminiscent of how some people think PC PSU's work. Like, they're worried pairing their hardware with a 2000W unit because that'll "push too much power!!!". However, that's not how that works. Current isn't being pushed, it's pulled and your hardware will only draw as much current as it needs. Similarly, data from a game server isn't increased with a variable. Each tick the server sends out a data packet with exactly all the required information. You can't cram more data inside of a packet, because there isn't any more data to cram.
rusihhh wrote:
28 Jun 2025, 09:47
If it suspects issues — even in the absence of real packet loss — it may start sending smaller packets with fewer frames of data.
How can the game "think" that your connection isn't stable if it can't detect anything wrong like packet loss, increased ping or similar? What aspect of the network would have to go "bad" in order for the game to adjust the size of packets? What parts of the connection are evaluated by that algorithm?
rusihhh wrote:
28 Jun 2025, 09:47
The server continues sending data, so you see zero loss.
This directly contradicts another statement from you:
rusihhh wrote:
28 Jun 2025, 09:47
But your client is now forced to interpolate or wait for more updates
When the server sends you data and diagnostics show 0% loss, you have the necessary data to display the correct game state without guesswork. All the "gameplay frames" are there. If data is missing and your client needs to interpolate, that's the literal definition of packet loss. A data packet with no legible information inside is considered lost, even if it technically arrived. Servers always run on fixed send rates and every player will receive the exact same amount of packets each second. In CS:GO, this was either 64 or 128 packets per second. Varying packet size is normal, it correlates with how much is going on in the game.
rusihhh wrote:
28 Jun 2025, 09:47
Analogy:
Imagine a mailman is delivering 10 frames of a movie every second.
If the network is perceived unstable, he starts delivering only 5 frames per second.
The movie keeps playing, but now it’s missing half the real information.
That's a textboox example and analogy for packet loss. If a movie theater sends out 10 frames each second and only 5 of those are delivered, 50% of the 10 frames that were originally sent out are lost. Basic stuff.
rusihhh wrote:
28 Jun 2025, 09:47
It’s technically a “delivery success,” but you’re no longer seeing the full picture.
Not for the other 5 frames each second that weren't delievered. You're once again directly contradicting yourself.
rusihhh wrote:
28 Jun 2025, 09:47
It’s technically a “delivery success,”
VS
rusihhh wrote:
28 Jun 2025, 09:47
it’s missing half the real information.
You can only have one of the two.
rusihhh wrote:
28 Jun 2025, 09:47
What do you think? Does this sound plausible?
No. A network connection isn't some sort of black magic hole where data is sucked into and spat out the other end from the mouth of a unicorn. Networks are well understood and can be accurately diagnosed. There's only a handful of variables that would have an influence on the quality of a game, and we can measure and analyze all of them. The three major ones are ping, packet loss and jitter. For all of these 3, basically every modern game comes with bulit-in diagnostic tools to show you exactly what's going on. It's not like the game presents you with these diagnostics and hides from you the vital signs of said unicorn or something. In other words: If the game isn't showing something to be amiss, the network connection is fine and problems are of a different origin.

iwabik
Posts: 9
Joined: 14 Jun 2024, 19:28

Re: The Game Feels Sluggish, Enemies Are Too Fast, and You Die in a Millisecond

Post by iwabik » 07 Jul 2025, 05:25

rusihhh wrote:
07 Jul 2025, 00:34
– What was your average ping during these matches?
– Were all games played on the same server/region on FACEIT?

If none of these games had noticeable issues or advantages, perhaps your captured avg packet size range is close to a "baseline" for stable conditions?
My ping was around 20-30 ms, and all matches were on Faceit's German servers. However, having five 'normal' games in a row is quite an anomaly. I usually play on these servers, but the feeling can differ substantially from server to server, even if they are hosted in the same region. I agree that this might suggest some kind of baseline for stable conditions, where the game feels fair.
rusihhh wrote:
07 Jul 2025, 00:34
Would you be willing to either:
– Switch to a different server/region (with noticeably higher/lower ping) and record the packets there,
or
– Record a session when you do clearly feel problems or advantages?
I don't think switching regions is possible, as they are assigned automatically, and I have a stable ping (20-40 ms) to most servers around the EU. The most popular German, Dutch, or French servers are all within that range. The only exception would be the Swedish servers. I believe there is something fundamentally wrong with the connection between Poland and Sweden because whenever my friends or I play on them, it's always the worst possible experience—and that has been true for at least a decade now.

I will definitely record a session that isn't 'normal' when it happens. My gut tells me it will be soon, because getting six normal games on Faceit in a row would be quite an achievement :lol:
rusihhh wrote:
07 Jul 2025, 00:34
Also, it would be awesome to see a percentage distribution of packet sizes — for example:
what % of packets fall into ranges like 250–500, 500–750, etc.
Here you go. I think the distribution is mostly consistent across the games.

Game 1:
0-249: 6.0%
250-499: 24.2%
500-749: 24.1%
750-999: 18.6%
1000-1249: 18.8%
1250+: 8.3%

Game 2
0-249: 5.3%
250-499: 28.4%
500-749: 28.8%
750-999: 26.0%
1000-1249: 18.2%
1250+: 6.6%

Game 3:
0-249: 4.1%
250-499: 22.6%
500-749: 25.7%
750-999: 24.4%
1000-1249: 17.0%
1250+: 6.2%

Game 4:
0-249: 9.4%
250-499: 26.3%
500-749: 21.6%
750-999: 20.2%
1000-1249: 16.6%
1250+: 6.0%

Game 5:
0-249: 4.6%
250-499: 35.1%
500-749: 30.9%
750-999: 33.7%
1000-1249: 16.9%
1250+: 5.1%
rusihhh wrote:
07 Jul 2025, 00:34
If there's any chance you're able to share the .pcap files (even a sample) —
I’d really appreciate it via private messages, if you're comfortable with that.
Sadly, I don't have the .pcap files saved. I only have .csv exports—I don't know if they would be helpful.
rusihhh wrote:
07 Jul 2025, 00:34
One question though — are you confident in this analysis method overall?
Honestly, I'm not sure. As I've stated, I don't know anything about network engineering, but I strongly believe that servers play a huge role in how CS2, in particular, feels. It just doesn't make sense for hardware or power-related issues to be the main culprit. How likely is it that you'd get 'bad' electricity for exactly one game, only for the problem to vanish for the next one? It makes no sense to me. Obviously, those are real and serious issues, but I believe they would have a more permanent effect instead of differing from match to match. As for the analysis method, I can't say for sure. I haven't recorded an 'abnormal' session yet, so I guess we'll learn more when that happens.

rusihhh
Posts: 49
Joined: 28 Jun 2025, 04:57

Re: The Game Feels Sluggish, Enemies Are Too Fast, and You Die in a Millisecond

Post by rusihhh » 07 Jul 2025, 06:07

ChristophSmaul1337 wrote:
07 Jul 2025, 05:20
...
Honestly? I'm too lazy to go line by line, but let's be clear:

Your claim that "servers don't care about the client's connection quality" is **completely false**. That's simply not how modern games work — and it's easily verifiable.
For example, in **Rainbow Six Siege**, the **developers themselves have stated** that **additional connection checks** are triggered when a player's ping or stability becomes questionable.
Yes, it adds computation, but that doesn’t mean the mechanism doesn’t exist.

I also mentioned **similar mechanisms in other engines and games** in earlier posts. You didn’t even attempt to verify, just assumed it’s “too much work” for a server — which doesn’t invalidate that it *happens*.

As for packet sizes:
Sure, packet size fluctuates with activity — but that **doesn’t explain** when the entire packet range **shifts from 500–900 bytes to 200–400 bytes** steadily. Even if there’s more or less going on, that range shift is **too significant** to be explained by in-game movement alone.

Now answer this:

If it's "not the internet" —
**Why can I play two matches in a row**, and one feels *perfect* (10–0), and the next feels *completely broken* (0–10)?
Did my **PSU die between rounds**?
Did I **hot-swap my GPU mid-lobbie**?

No.
And it's not a "skill issue" either — I've played long enough to know the difference between being outplayed and the **game literally feeling like a different engine**.

**IMHO**, trying to understand these dynamics and packet behavior is way more productive than chasing the myth of a "golden power outlet".

So again:
If the connection is "fine" — **what exactly changed between games**?

Think before jumping to oversimplified answers.

rusihhh
Posts: 49
Joined: 28 Jun 2025, 04:57

Re: The Game Feels Sluggish, Enemies Are Too Fast, and You Die in a Millisecond

Post by rusihhh » 07 Jul 2025, 06:12

iwabik wrote:
07 Jul 2025, 05:25
...
Thanks a lot for the detailed response and the distribution stats — very helpful.

I completely agree with your take:
I'm also not swapping my CPU mid-match or switching generators from diesel to gasoline between rounds

If power supply was really a core factor in this kind of issue, the entire competitive scene would've already moved to **online double-conversion UPS systems** — and it would be common knowledge by now.
But that's not the case. So....

There’s clearly *something* affecting how the game feels from one match to another, and the fact that hardware or power issues would come and go on a per-match basis makes absolutely no sense.

Really appreciate your insight. Looking forward to any “abnormal” session data you might share later!

User avatar
ChristophSmaul1337
Posts: 111
Joined: 11 Feb 2024, 21:01

Re: The Game Feels Sluggish, Enemies Are Too Fast, and You Die in a Millisecond

Post by ChristophSmaul1337 » 07 Jul 2025, 06:51

rusihhh wrote:
07 Jul 2025, 06:07
For example, in **Rainbow Six Siege**, the **developers themselves have stated** that **additional connection checks** are triggered when a player's ping or stability becomes questionable.
You explicitly stated that what you're describing happens when connection diagnostics do not show abnormailites.
rusihhh wrote:
28 Jun 2025, 09:47
- You die instantly, even with perfect ping and no packet loss
But now you're arguing with checks that are being performed once the player's ping becomes questionable. Which in turn would mean the player doesn't have "perfect ping" and a warning would be triggered by the game's diagnostic overlay. At this point, we're talking about an actual problem with the network which has nothing to do with the premise of the original post.
rusihhh wrote:
07 Jul 2025, 06:07
You didn’t even attempt to verify
Slow down, friend. Careful how you word things and what your message insinuates.
rusihhh wrote:
07 Jul 2025, 06:07
that range shift is **too significant** to be explained by in-game movement alone.
According to who? Clearly you are involved in the development of some game here, judging by the assertiveness in the wording. What game is it, and what is a "significant" shift in packet size range in that game? How is that game similar to others, and how can the changes in that one game be translated to other games?

I really hope you're a dev, because otherwise, your statement is just an assumption, which would make you a turbo hypocrite. Oh wait, your OP mentions that you are indeed
rusihhh wrote:
28 Jun 2025, 09:47
not from gamedev
Well... talk about shooting yourself in the foot.
rusihhh wrote:
07 Jul 2025, 06:07
**Why can I play two matches in a row**, and one feels *perfect* (10–0), and the next feels *completely broken* (0–10)?
You think if I had the answer to that question I'd keep it for myself? Nobody knows this, that's what we're here to find out.
rusihhh wrote:
07 Jul 2025, 06:07
Did my **PSU die between rounds**?
Nobody said it has something to do with your PSU.
rusihhh wrote:
07 Jul 2025, 06:07
Did I **hot-swap my GPU mid-lobbie**?
Nobody said it has something to do with your GPU.
rusihhh wrote:
07 Jul 2025, 06:07
And it's not a "skill issue" either
Nobody said that either.
rusihhh wrote:
07 Jul 2025, 06:07
chasing the myth of a "golden power outlet"
Nobody does that and the notion of something like that even existing is nonsensical jibberish.
rusihhh wrote:
07 Jul 2025, 06:07
If the connection is "fine" — **what exactly changed between games**?
You could conduct experiments in order to find that out, instead of typing some nonsensical manure like
rusihhh wrote:
07 Jul 2025, 06:07
chasing the myth of a "golden power outlet"
Well, I thought I could give some insights here but apparently I'm being talked down to by the OP, so there you go, I'm out of this one. Hypocrites like you I'm not interesting to have a conversation with.

A well-intentioned advice: Don't type in your original post that you're not a game dev, just to get angry at someone who points out flaws in the logic. This whole idea is based on the assumption that multiple "gameplay frames" are contained within a data packet, and that's simply not true. One packet equals one "gameplay frame", and if you have 0% loss, all the information is there. That's the end of that train of thought.

Also, try to keep to your own advice and stop assuming things you don't know. Goodbye forever.

rusihhh
Posts: 49
Joined: 28 Jun 2025, 04:57

Re: The Game Feels Sluggish, Enemies Are Too Fast, and You Die in a Millisecond

Post by rusihhh » 07 Jul 2025, 07:50

ChristophSmaul1337 wrote:
07 Jul 2025, 06:51
...
Let’s clarify a few things.

First, I never claimed I was a game dev — quite the opposite, I said I *wasn’t*. But that doesn’t mean I can’t think critically or observe real behavior over hundreds of hours. And I’m not asserting things blindly — I **referenced known mechanisms** that exist in modern games, like Rainbow Six Siege, where **developers explicitly stated** that **connection-based checks** are triggered when a player’s ping or stability is questionable.

You argue it would add server load — sure. But **why assume that means it doesn’t exist?** Did you verify? Why present it as a *fact* that such systems aren't used? That’s the core issue here.

You call me a hypocrite for “assuming,” but your whole argument basically comes down to:
> “Why would anyone design it that way? It’s inefficient.”
That’s not logic — that’s just wishful thinking.

---

About packet ranges:
You don’t need to be a developer to notice that in Game 1 the packets are consistently 200–300 bytes, and in Game 2 they’re 500–600 bytes — throughout the entire match.
That’s not just "more action in-game". That’s a **baseline shift**, which points to something systemic — not tied to player behavior.

---

Now about your videos — did you ever check the packets during those deaths?

Here's what I believe happened:

- The server was supposed to send you the enemy’s position, aim, and shot in **Tick1**.
- But instead, the system chose to send **irrelevant or incomplete data first** — something not directly related to your combat situation.
- The game had **part of the picture**, but not enough to render the enemy or your own death.
- So it waited — and only by **Tick4**, when the missing pieces arrived, it rendered everything at once: enemy appears + you die.

This isn’t about ping or packet loss.
It’s about the **game giving you the wrong or partial data**, which still needs to be **waited on and reassembled** before anything can be shown.

And here’s the key point:
Your **computer and hardware did their job 120%**.
Your system **instantly, accurately rendered everything** the moment it had enough data.
The issue isn’t what your PC showed — the issue is **what it was given to work with**.
And that’s entirely on the game engine and how that **network logic module** decided what to send first.


---

Now let me explain the **core of my theory**:

I believe modern games (especially ones like CS2 or Siege) use **dynamic packet splitting**, but the implementation **was not carefully designed**.
The way packets are split and prioritized **doesn’t align with gameplay-critical priorities**.

And frankly — I doubt the developer who wrote that system even considered fairness or concurrency.
They were likely tasked with writing a **module to assist players on unstable connections**, or **reduce server load**.

Did they think about what happens in a 64-tick, latency-sensitive match when the wrong info arrives first?
Probably not.
Their only concern was:
> “Keep the connection alive. Avoid disconnections.”
And they succeeded.
But at the cost of **delivering the wrong data first**, or **delaying critical information** — like your enemy’s actions — for no good reason.

---

So in your video example:

- The client didn’t get what it needed at Tick1.
- It got **partial or low-priority info**, and had to **wait for additional packets**.
- Rendering only started at Tick4, by which point the server had already processed your death.
- You saw the enemy *and* your death appear at once — but only *after* those skipped ticks — and had **no chance to respond**.

---

**TL;DR (if you skipped the above):**

> The game gave you either **the wrong** or **incomplete** data.
> It made you **wait for the rest** before rendering anything meaningful.
> Your PC and hardware responded perfectly — it rendered exactly what it received, as fast as possible.
> But what it received came late and incomplete — and **that’s the real problem**.

It’s not hardware.
It’s not power.
It’s not skill.
It’s **bad logic in how the game decides what to send, when, and in what order**.

**IMHO, this explains far more than magical “voltage ghosts” or golden power outlets.**

Again — I’m here for respectful discussion.
*Sorry if anything sounds off — I’m using a translator to write this in English.*

User avatar
dervu
Posts: 379
Joined: 17 Apr 2020, 18:09

Re: The Game Feels Sluggish, Enemies Are Too Fast, and You Die in a Millisecond

Post by dervu » 07 Jul 2025, 09:31

I see the main culprit is question: What data would engine use to determine if your connection is bad, other than ping, jitter and loss?
If those are perfect, why would it be bad?
Unless real data is hidden from user, it doesn't make sense to me. Anyways you could make your own server and measure it yourself to see what exactly happens with that data.

Only possible thing in that case could be adjusting packets send based on your ping, jitter history. That could be easily tested with your own server.
Ryzen 7950X3D / MSI GeForce RTX 4090 Gaming X Trio / ASUS TUF GAMING X670E-PLUS / 2x16GB DDR5@6000 G.Skill Trident Z5 RGB / Dell Alienware AW3225QF / Logitech G PRO X SUPERLIGHT / SkyPAD Glass 3.0 / Wooting 60HE / DT 700 PRO X || EMI Input lag issue survivor (source removed) 8-)

iwabik
Posts: 9
Joined: 14 Jun 2024, 19:28

Re: The Game Feels Sluggish, Enemies Are Too Fast, and You Die in a Millisecond

Post by iwabik » 07 Jul 2025, 10:17

I don't understand why people would choose to argue for 15 pages over a theory they don't like instead of just downloading free software, letting it run in the background, and actually providing some contradictory evidence. Or, you know, just ignore the topic if they deem it stupid and useless. It's not like OP is providing any solutions here anyway; it's just a discussion, really.

Anyway, back to the topic. Just a thought from my end: averages are usually very poor metrics, statistically speaking. I have one more observation because I found a game I had labeled as "bad," but sadly, I have no other info and no graph from Wireshark for it.

"Bad game"
Packets sent/s: 63.95
Packets received/s: 66.54
AVG packet size (received): 714.99
StdDev packet size (received):304.89
AVG packet size (sent): 737.99
StdDev packet size (sent): 175.84

Counterintuitively, the packets seem to be bigger here, but again, averages are pretty bad for statistical analysis.

I also have a bonus graphs and stats because another idea came to me: what about the timing between each packet sent and received? If the game is 64-tick, then technically the time between two consecutive packets sent or received should be exactly that: 1/64 of a second, i.e., 0.015625s. So, I created a window of +/- 10% of that value and checked how many packets were sent or received within that margin (i.e., the difference in time between one packet sent/received and the next one in the log).

Here are the percentages of packets within that +/- 10% time window:
  • Game 1: 95% of packets received, 94% of packets sent
  • Game 2: 92% of packets received, 78% of packets sent
  • Game 3: 98% of packets received, 95% of packets sent
And box plots:
Sent.png
Sent.png (52 KiB) Viewed 7637 times
REC.png
REC.png (42.39 KiB) Viewed 7637 times
Guess which match was labeled as "bad"?

User avatar
Slender
Posts: 1548
Joined: 25 Jan 2020, 17:55

Re: The Game Feels Sluggish, Enemies Are Too Fast, and You Die in a Millisecond

Post by Slender » 07 Jul 2025, 10:25

it seems that packets are not processed by nic in real time, as if there is some buffer, or rather a queue, like one packet is waiting for another to be processed. something prevents the internet controller from working normally. It is for this reason that disabling interrupt moderation and taskoffload work with desync for some time.

Post Reply