Sorry SE, but endgame won't work if you don't fix something.

[Gravehill]

Here's another test for you -

Run in a straight line (do not stop) and cast Esuna.

What point in the cast bar does it finally interrupt for you?

For me it will sometimes interrupt at around 00.25, other times the bar will completely fill, animation will go off, but in the chatlog it will say "You cancel Esuna".

I think Esuna is a really easy way to test latency on casts because of this, as it's about a 1 second cast time, which makes it very easy to see how long it actually takes for the server to tell your client that you interrupted the spell... For me it is basically between 700-900ms... Pretty crazy huh?

Ive done this with summoning a mount - I can mount up while running and the cast bar gets pretty much 100% full before it even interrupts. I get 20 ping to the NA server.

[Altena]

Ive done this with summoning a mount - I can mount up while running and the cast bar gets pretty much 100% full before it even interrupts. I get 20 ping to the NA server.

Didn't think of mounts, but yep that's basically the same as Esuna/Leeches. Cast time seems the same / similar.

So it seems you are experiencing the same problem as others here. 50/50 on spell complete is 50% too many completes. 300ms position check is at least 250ms too long if they are going to server authenticate everything. Always been 70 and 30 interrupt means you are playing at 700ms - 300ms latency for that, which is solidly in the "terrible" category.

These are the issues we are referring to and talking about, issues that every single player is experiencing no matter what their connection or PC.

Well yes I don't think Namko was really disputing the fact that there is some degree of latency there.

The one thing I will agree with some of the people in this thread is that internet / computer does have SOME effect on the latency, but overall it's a compounding issue.

Poor ping (due to even geographical distance) as an example means that once your client sends the data, it still takes 250ms (well ping dependent) to send to the server, when paired with a poor check interval (currently seems to be about 300ms), this is compounded to 500ms or so, then remembering it takes another 250ms (ping dependent) for the server to send it back to the client and say "that action is invalid, you need to cancel that". So this is probably the sort of place we are seeing this 700-900ms latency... The rubber banding effect of client > server > client..

When someone lives right next to the servers in montreal and gets 5ms, they only have to deal with the check interval, which is why the issue is still present but not as obvious.

Well that's my understanding of the issue..