Catwho wrote:
In the case of SoCal, it's probably the way that the local ISPs are routed through a particularly busy backbone.
I have very little lag where I am because I'm one hop away from Atlanta, which has a backbone that runs directly to Montreal, and the SE data center is one hop away from that. The issue is that the more big cities there are before you and your final destination, the more backbones you'll pass through. Someone out in SoCal has to hop up the coast a few times, then go through Phoenix or Salt Lake, then Minneapolis or Chicago, then New York, and only then do they get near Montreal. Depending on the ISP there might be more or fewer hops, but it boils down to lots of milliseconds adding up to half second delays each way.
This is also why satellite connections are the worst for video games, because their ping times are pretty horrible. But their packet bandwidth is huge so they're great for one way streaming. Cable internet sits in the middle for Quality of Service and bandwidth, and DSL tends to have the best QoS handling at the cost of smaller pipes, because phone lines themselves have to have good QoS or else you'll have scratchy sound.
I see. I didn't know we were so close to the servers, Internet traffic-wise. From Florida my connection has almost always been stable. The only place I ever experience this animation lag that people speak of is on turn 1 ADS laser on Monday nights when the coil servers are most likely congested. I would say about 60% of my FC plays from California (and maybe 40% from SoCal) and have heard really bad stories about lag.
But now it makes a little more sense why this would be the case. Any idea why they didn't just stick the data centers in the U.S. or why they don't have two data centers in the U.S. besides cost? I mean, I feel like it's the opposite of what Blizzard did in the old Starcraft 1/Diablo 1 and 2 days. They had US East and US West servers for almost all their early games but only one Europe option and one Asia option.
Of course, Blizzard eventually changed so that their games work well in most parts of the world but their initial basis was in the U.S. and focused mainly on NA players it would seem. I am not giving SE an excuse since they have a past MMO where they have people from all over the world...I just don't think 11 saw these problems because the combat wasn't as fast paced.
Gnu wrote:
This moves the discussion forward certainly. The unintended result will be for you take take damage when you are running INTO danger on purpose. Take for example running INTO plumes to avoid Ifrit charges. You know the damage will be dealt before your run on top of plumes, so you are timing it precisely to coincide with the explosion.
There is a lot of ground AoE, Titan bombs, and other mechanics where you are trading one danger for the next, and delaying damage will just make these frustrating from the other side.
(On a side note: I do not at all support changing PvE mechanics to help improve PvP. Just a personal opinion.)
My suggestion: We need the calculation for damage vs. position to happen client side. The server tells the client the timing of the explosion. The client decides if you got hit and reports to the server. The server registers what happened. Is that not a feasible solution?
This is very true, too, and something I take advantage of quite often yet I don't ever see complaining with this. You can easily run into a landslide that has already 'gone off' without taking any damage or being knocked back. Same with big pillars of fire like Ifrit's plumes. You can run THROUGH the plumes as they go off as long as you were on the edge when they initially went off. Not very realistic to run through pillars of fire and not take any damage but this is just another side to the lag story.
It's a bit harder the other way with punishing abilities like landslide hence the backlash from the community I suppose.
Edited, Jan 10th 2014 12:52am by HitomeOfBismarck