Rallaz wrote: ↑12 May 2020, 15:38
Yes it happends offline somedays in freeplays that the game feels "heavy" but after connection to a match or two its usually gets better as well in freeplay.
Strange. It could be a combination of training effects (Common terms are the "training effect", "muscle memory", etc)
Sudden lag changes creates weird effects. Lag increases versus lag decreases.
A person get used to a specific latency and automatically compensates by aiming slightly early or slightly late, in order to aim correctly (e.g. shooting at moving targets).
Like Archery where you have to shoot your arrow a bit early -- in order to hit a sideways-moving archery target. I call that the "invisible target" that you have to aim ahead of the real target arriving to its real position.
The same problem happens in online / FPS games when latency changes. It's like an invisible target location that changed -- one moment, the invisible correct target to aim at is ahead of the target, the next moment the invisible correct target to aim at is behind the target. Ping jitter will gyrate the location of that "invisible target" making it so frustrating. In fact, you prefer "higher but consistent lag", because at least that invisible target is predictable location (whether it's a lagbehind target or a lagahead target).
Even a sudden decrease in lag can cause scoring misses, because you're trained/used to a latency.
If you play for days offline, you shouldn't feel this effect. But if you alternate between online and offline, you will definitely feel this effect.
If you want to converge online feel to the offline feel, you need as low latency as possible AND low latency jitter.
Yes, gigabit FTTH is overkill for a game that only uses under 10 megabits per second. But latency and latency jitter is superlatively low
on a lightly-loaded gigabit FTTH, so you're paying overkill for an overkill safety margin, even if you only use 1% of the Internet speed.
Loaded latency of most Internet connection suddenly gyrate up and down. You see this when doing
www.fast.com and clicking "Show more info". If your unloaded latency and loaded latency is dramatically different, that's very terrible for your game. It can kill your network game if your loaded/unloaded latency is 10x different. Now, fast.com is a terrible analyzer, you need better software... But my point is unloaded latency can be perfect but totally crap during loaded latency creating a hairpulling 100ms of latency jitter (especially as somebody suddenly starts watching Netflix/Disney in the middle of your game, or suddenly downloading a Steam game, or other latency-changing events).
Now a lightly-loaded connection (e.g. only 10Mbps of 1000Mbps) will have very low ping jitter on the important 10Mbps, you're paying for the speed overkill to purchase less network variance. Bandwidth is not everything, latency stability is also important. And because residential connections share each other, when your neighbour on another floor or another separate house suddenly downloads something off Steam, it may slow your connection down. Or someone watching stuff on streaming and suddenly stopping. So you might need to simultaneously use overkill margins (business-it + FTTH-it + dedicate-it to gaming PC only) to brute-force Internet gaming stability. Yes, you're overpaying. But a lot of the paid professionals have been doing it. If you have to score in those leagues, you just have to pony up.
20 milliseconds of ping jitter is equal to 5 different 240Hz refresh cycles, you know...for a crappy 75 megabit overloaded Cable Internet connection shared by your neighbours.
Higher refresh rates and higher frame rates amplify visibility of network jitter. 60Hz had refresh cycles that lasted ~16.7ms but today's 240Hz have refresh cycles that last only ~4ms. So you _really_ feel the network lag much more at 240Hz than at 60Hz.
Heck, even 4ms of network jitter = 1 refresh cycle of jitter on that moving target. In some games, it's noticeable enough, e.g. 4000 pixels/second high speed target (airplane flying over you in Fortnite) -- 4ms lag jitter creates a 16-pixel offset in that invisible target (Archery metaphor). (4000pps / 1000p x 4ms = 16p offset).
Almost everybody's Internet connection changes more than 4ms between unloaded and loaded in your favourite network analyzer (or even just merely
www.fast.com) -- now you understand why Internet connections are just often crap, a major bottleneck in the refresh rate race to retina refresh rates, in the
Vicious Cycle Effect -- where bigger resolutions, bigger FOV, higher resolutions, larger screens, higher refresh rates, better motion clarity, amplify each other -- amplifying those milliseconds to human-visibility or human-feelability of "why am I scoring crap?" effects. It was easier to be a LPB in the 320x200 60Hz Quake days if you were lucky to have the first DSL connections, today, the 1920x1080 240Hz, really amplify visibility of network latency issues.
Anyway, the feel of FTTH is closer to LAN feel than with most other connections (cable/DSL) when you're playing to local servers or playing through a good low-jitter network (including potentially a gaming VPN that bypasses the low quality latency-jittery ISP backbone).
You also might be witnessing multiple simultaneous effects at the same time
(A) Lagfeel changes from offline vs online Internet (training effect)
AND
(B) Lagfeel changes from peak to offpeak Internet (network effect)
AND
- Interactions between the two of the above
You try to frag a fast sideways-moving enemy, you've got that invisible aim target problem (like aiming arrow at a sideways moving Archery target). It's affected by both (A) and (B) above. The mere act of switching between online/offline can suddenly create a "why is my aiming bad?" effect even for sudden lag decreases. But network jitter can also jitter that invisible target around too.
Sometimes you can't feel the milliseconds directly, but you just simply have a "Why am I suddenly missing my frags now?". Exactomundo, exactomundo! That invisible lagbehind/lagahead target is moving all over the place.
That's why well-trained people who mainly play over Internet (instead of LAN) sometimes aim at consistent latency (but higher latency) can often score better than some people who try lower (but jittery) latency -- at least they're resolving the lag training effect through various techniques such as via low-jitter Internet, via framerate capping, or via other techniques that adds latency consistency over absolute lag. There are situations where one can be trained to prefer a 20ms reliably consitent lag, over very volatile 3ms-15ms lag (combined lag, frametime lag, engine lag, netcode lag, display lag, in the entire latency chain, combined)
So you might have a fog of red herrings during a wild goose chase, it may not be a single cause. You sometimes just have to bruteforce your way (lowest lag + lowest lag jitter)
Another big problem is that many other players won't have as good Internet connection, so your game may be bottlenecked by the average Internet connection of the whole online game. That can turn things a bit shitty for the best-Internet players on that particular game. The best thing you can do is do your best, and be aware of (A) and (B)
This may not be your latency problem but it helps to understand the (A) lag training effect and (B) network latency effects, and both (A) and (B) can be happening at the same time (confusingly so).