Tickrate or "netcode" problems are killing the game. Video inside.

Looks just fine to me. Anybody complaining about server/latency issues is an idiot.

You cannot "fix" latency, you can only mitigate the effects - A simple fact that has not changed since the first online multiplayer games.

The most important piece of information in that video is the one we don't know - the ping of the player "SiIberwanze". For all we know he could of had a ping of 800ms making the server tickrate almost irrelevant.

SIIberWanze ping - 800ms (theory crafting)
Server tickrate - 200ms (using your number)
AudieVS ping - 110ms (average)

Meaning, worst case, there could of been a delay of 655ms ( (800/2)+200+(110/2) ) from the time "SiIberWanze" got the kill to the moment AudieVS was notified he was dead.

Improving the server tick to a more realistic 25 gives us 40ms delay. That only reduces the delay to 495ms... thats still practically half a second!


As frustrating as this is for us all. The biggest problem with client-side hit detection is the ping of the player shooting us not the tickrate.

Ideally we want people to use a wired instead of wireless connection.
Ideally we want them to have a fibre internet connection.
Ideally we want them to play on a server thats in close geographic proximity.

Sadly that isn't realistic.


The only other options are; we move hit detection server side. That would dramtically increase server load and players would have horrible fire-delay after pressing fire.

Or increase the number of servers so more people have a server located closer to them. Still, that would split the PS2 population even thinner between servers.

Final option is the most drastic - kick high ping players. I hate this solution. We want more people to play Planetside not less. Yet, I recognize this is the easiest solution to implement and would probably have the desired effect. Who would want the task of telling people they can no longer play their favourite game because they have a ping of 200ms+ though?

I beg to differ. I think it's actually much easier to take a dev's word as an absolute truth simply because of their position/status instead of doing your own research and drawing your own conclusions.

But hey, I choose to do the latter, so we will always be at a disagreement. I can accept that.

People that whine about such things are rabidly complaining on an internet message board. They just need to learn how to play because the game actually works fairly well.

I think you missed my sarcasm. The video I quoted is clearly an example of the game NOT working fairly well.

Actually, I think you missed my sarcasm Check the first post of page 2. Also, read the other two responses that I quoted in that same post, then read what I said to you. You'll understand

I wish I could add something that would change the course of this hilarious propaganda campaign. Somehow, for most of the people in this thread, the OP is a credible source. If you're angry about what this guy says, please go look up "tickrate" and "Netcode" and Hz and all that other stuff. Trying to just tell you would have little to no effect than real education at this point.

The OP basically created a PS2 Da Vinci Code. Enjoyable to watch, but it seems some of you are having some real emotion about this. He's taking advantage of your lack of understanding.

This is the problem. Having high latency should never be an advantage, instad, in Planetside 2 it is. Let's take Team Fortress 2, or any Souce Engine game for that matter. These games do lag compensation the proper way. I am not going to explain it here, go read it yourself over here instead. In short, the Source Engine actually does check hits server-side, but uses a smart lag compensation system to avoid situations like those seen in older games such as Halo CE, where you had to aim miles ahead of your target to score a hit.

Planetside 2 does not check hits server-side. When you hit someone on your screen, only the damage info is sent to the server, that accepts it regardless of how it was done. Maybe you used a clientside cheat engine to aim, maybe you unplugged your router, then shot everybody in the base, then re-plugged it and all the damage info flooeded out towards the server - the server does not care. You can have 1000 ms ping and use that to shoot people behind corners - the server does not care. In the Source Engine, the server only allows lag compensation up to 1 second back in time (1000 ms latency, while PS2 will accept any damage and hit info no matter how old they are) - however, in Source there are a number of other things that make the game unplayable well before reaching that thresold (for example, input prediction which causes severe stuttering at overly high latencies).


Of course, server side hit detection + lag compensation via time rewind is not feasible in Planetside 2 due to the huge load that servers already have to endure. However, I'd like to see a system similar to Source's 1 second maximum lag compensation time. Basically, make it so that if damage info arrives and it is over, let's say, 500 ms old the server completely ignores it. How old a data packet containing damage info is can be calculated by either checking the sender's ping as soon as the packet is received, or, if you want more accuracy, by implementing a timing system (if it's not already present); the server would check when the damage was recorded (this piece of info would be present in the damage info packet) and compare that time (say 29.300 s) against its current time (say 29.470 s), and if the difference is greater than 500 ms, the damage is denieded.

This is just something that popped up into my mind - it may or may not be technically doable and it may or may not be right.

There's not much propaganda to do - everybody, me included, is suffering from networking problems. That particular video may have been wrong, but I doubt that the other 50+ videos highlighting this kind of issues are all wrong. See the video at the top of the page.

I'm just trying to see how many other people are having these kind of issues (apparently many) and trying to remind the developers that they should start with fixing networking before anything else.

Apologies mate.

The gameplay experience improved, no doubt. There were a few costs however, most notably perhaps the resources used on the server(possible server fee increases) and the amount of resources used by the client (both upstream and downstream). BF4 with the 30Hz tick rate needs 1Mbits/s for BOTH upstream and downstream. That is pretty damn high when comparing it to other multiplayer games. Now consider Planetside 2's servers, you can't possible imagine the amount of data being sent back and forth from a client with over 200 players in one area at a tick rate 6 times faster than what it is now.

Increasing the tick rate comes at a cost of resources and essentially higher server rental fees for SOE and also the client's bandwidth and usage. Is it worth it? Given the details unveiled by the OP it seems SOE is saying no. Oh well

I pretty much stopped paying attention to what you said here, because Source has terrible net code. I die behind walls far more often in TF2 than I do in PS2.

Today this is getting ridiculous in Miller, turn behind a corner, get shot 1 second later, and it's not lag and happening in even small fights

Problem is, to generate that revenue they need to provide a product people are willing to pay to support. A lot of people have cancelled membership, or refuse to spend SC on this product, all because of its current performance. In other words... to make money, you need to spend money.

"I suppose I should also note that I think the EM6 is the worst LMG in the NC arsenal. I'd seriously rather have the EM1, mostly because I fully suspect that server tick rates were lowered at some point, rewarding higher rof weapons and punishing lower rof weapons"

I posted that in this thread last week. I don't follow reddit, but I am glad that I am not the only one that thinks this.

I was wondering if it is possble to compare accuracies from pre-pu02 to accuracies now on groups of weapons. I suspect that higher ROF weapons will have >/= the accuracy they had then and now and lower ROF weapons will have < accuracy now than they did then.

[youtube]

keep an eye on the hitmarkers while i'm fighting maxes....

I don't get how people don't see how important the tick rate is. Assuming the OP is correct on the current tick of 5 Hz, that's 0.2s before each batch of information is processed by the game.

A 600 RoF weapon, a fairly average rate, can fire two bullets reliably within that time frame. 0s tick -> 0.05s shot -> 0.15s shot -> 0.2s tick.

Taking the Gauss Rifle since it fits the RoF, we have 2 reliable shots at 167 damage with a 2x headshot multiplier. Assuming they don't miss, this is 334 to the body, 267 with max NWA, or 668 damage to the head.

The tick rate of most CS:S servers is 66 or 100, leaning to the latter in my experience. This translates to 0.015s or 0.01s between each process by the game. Many of these servers are 32v32, even if the "official" play is 5v5. It's not perfect in the real world nor is it the same game and completely comparable, but the 95% difference in the tick rate can definitely be felt, especially in close to mid range (0m-50m) fights (Read: Most of PS2's infantry gameplay.)

@Firejack

How can you not acknowledge a 25% decrease in the delay is important? Even if you want to just focus on how a higher tick impacts people who are laggy, that's silly.

Let's take my situation with your numbers.

My ping is pretty steady at an average of 60 ms (0.06s). On a 0.2s tick, my inputs are delayed up to ~400%. Maximums below are set to be close enough for government work as I don't feel like doing limits for a forum post.

-0.05s I shoot on my end -> 0s tick -> 0.01s shot is received by server -> 0.2s tick, shot is processed. 0.205s delay. Difference between my delay and theoretical 800ms delay you figured is 0.45s (.655s-.205s)

On a 0.4s tick, I get a maximum of: -0.05s I shoot -> 0s tick -> 0.01s shot reaches server -> 0.4s tick, shot is processed. Total of 0.45s delay. Difference between my delay and the second theoretical 800ms delay is 0.455s (.495s-0.45s).

The lagging player gets a better game, the people with good ping get a better game, and the difference between the two is not significant enough to outweigh the benefits.

But like the OP said, an official word on the tick rate would be nice.

Maybe for NC. Not so much for TR with RoF hitting over 900.

Comment thread where muldoonx9 said tick rate had not changed
Comment thread where muldoonx9 said hardware has not changed

If anything it likely just the fun bits of clientside hit detection. I highly doubt that this is misinformation and maybe the servers just need to restart (or an unlikely DDoS attack).

EDIT:
muldoonx9 says that they did not downgrade hardware and that he may have an idea of what the problem could possibly be.

That's technically possible... but why?

You need to do that math overhead for every tick, but 99% of your players are going to have <500ms ping, so you're doing it for nothing. If the player is cheating they can spoof the timecode as well, so you don't stop that either.


And beyond all that, the video you're using as proof of how bad the netcode is follows your suggestion anyway. Remember, that player isn't being killed by the people across the bridge but by an LA hovering above the door.

So again, what is gained by adding all that overhead to each server tick (and then, as the more vocal are suggesting, multiplying the tickrate)?