It’s okay for me most of the time. It tends to be individual high pingers that I have hit reg issues with. Just played 1 game with I’d say only one instance of not getting the hit and the other player was 100 ping.
It is kind of funny: if it is doubly bad EVERY time on patchday, it might be a good idea to double capacity, use different locations/providers/do whatever is needed to avoid that?
It is bad we almost think that is normal and acceptable …
Every patch day there’s a player count spike. This happens with every game.
Doubling the number of servers won’t change anything because you can still only fit 64 in a server. All will still be a mix of low and highs. The only thing adding more servers will do is reduce wait times on queueing for MM.
There’s no such thing as lag compensation. Packets are processed on a first come first serve basis. Having a low ping means your packets make it to the game server faster.
Ping: Packet Header Timestamps are used to determine how old the packet is.
On creation a timestamp is placed in the header. On receival the timestamp value (integer) is subtracted from the current time timestamp (integer). This results in a ms value for one way travel time. This value is doubled which results in a Round Trip Time … PING.
Input TimeStamps: Packet payload timestamps per input are used to determine how old each specific input is by the time the packet reaches the server and vice versa.
Thus at 0:59:987 game time the player pressed W to move forward, then at 1:00:000 game time the player pressed the jump key.
e.g. I send a packet to the server with my input actions. One of the actions is “jump”.
The server reads the input timestamp for the jump command. Rewinds its simulation and updates its simulation for that moment in time (1:00:000) with me starting a jump.
Lag comp simply determines:
“WHERE SHOULD THE PLAYER BE AT CURRENT GAME TIME BASED ON THE LAST INPUT RECEIVED AND HOW LONG IT TOOK TO GET THAT INFO”.
So if my input time for my jump was 1:00:000 (1m, 0s, 0ms) and it took my packet 20ms to reach the server and 1ms for the server to process the packet, then at the current time in game (1:00:021) the server will have me at the 21ms position of the jump animation.
^^^^^^^^^^^^^^^
That’s Lag Compensation on the server.
On the client-end we do the same thing. Server said dude jumped 21ms ago. Have dude jump, but we need to interpolate him into the jump 21ms. So start the jump and then increase the playrate of the animation to accommodate for the delay.
This results in the jump being executed a bit faster than normal on our end by just a few frames. Say for example the entirety of the jump is 30 frames normally. By increasing the playrate of it we can shrink its duration to 28.5 frames. Thus speeding it up.
If the delay is small then the client will interpolate across multiple frames smoothing out the transition. Otherwise the player is warped. Just like client-side prediction and server correction of your movements.
The server CANNOT send packets more frequently (sooner/faster) than what its tickrate permits.
60Hz Tick only allows for an update rate of 1 packet every 16.67ms. (1000ms / Hz)
The tickrate controls how often the server updates its simulation via “TICK EVENTS”. Every Tick the server executes a sequence of a events and calculations. Once those calculations are done it then compiles packets to be sent out.
Can someone please add a videoclip/picture of how the new (2 or 3 patches ago changed) network performance graph and stats?
I seem to be missing the network stats like and I want to know if it is my game that is broken or everyone is missing that info.
I do have the graph and kill info, but don’t have the latency, tick rate, srv tick, packet loss, etc numbers.