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

[CatofNineTails]

The game positions you based on when the cast time finishes. This has been tested before.

minus the fact what the game see's is not whats being shown on the players screen IE: Latency lag to the players. this is not a internet connection issue. the ping and connection quality are too good for that to be the case. also the issue is mostly in instances and not during regular play. knowing you are 3-5 seconds behind what the servers are seeing can be extremely problematic. the action bar has already completed before the game even puts up the tell to move. no amount of skill can beat impossible timing issues.

[Cienna]

Is this really a latency issue? Or just a failure to understand the mechanics of how the aoe system works in this game?

The game determines if you are "in" the area by your position when the casting bar is 100%.

Here is just one of the MANY video examples in this thread.

http://www.youtube.com/watch?v=QsHuXAl-bOY

At 6:21 landslide cast begins
At 6:22 player leaves aoe with cast bar at 75%
At 6:24 player is hit by landslide and knocked off

They were out of the red area well before the cast bar finished. They still get hit.

[TheRac25]

try casting somehting on yourself see if you can interupt it at 90% castbar by moving, thats right you cant
effect goes off at 100% cast bar my ass

[Crafter]

The game positions you based on when the cast time finishes. This has been tested before.

Have you seen Sunarie's 2nd video above? She was out of AoE one second before the cast bar went to full, but she still got hit. My lag is just like hers. It's sometimes as bad as 2 seconds for me. Maybe the test you're talking about was done by someone who doesn't have the lag like we do.

[Cienna]

The game positions you based on when the cast time finishes. This has been tested before.

Great, then you agree it is broken! Look-

http://www.youtube.com/watch?v=QsHuXAl-bOY

At 6:21 landslide cast begins
At 6:22 player leaves aoe with cast bar at 75%
At 6:24 player is hit by landslide and knocked off

They were out of the red area well before the cast bar finished. They still get hit. There are many more examples in this thread, like Sunarie's 2nd video a page or two back, and heaps more in the 100+ pages of this thread. It is broken.

[Sovjohn]

I'd like to add something to the conversation here -

The following is a traceroute to the NA / EU datacenter in Montreal by an ISP in southeastern Europe (Athens, Greece):

Non-prime time (23:50 BST - 01:50 EET)
Tracing route to 199.91.189.25 over a maximum of 30 hops

  1     1 ms     1 ms     1 ms  dsldevice.lan [192.168.1.1]
  2    25 ms    24 ms    24 ms  bbras-llu-kln-15L500.forthnet.gr [213.16.246.12]

  3    25 ms    24 ms    25 ms  213.16.247.73
  4    28 ms    28 ms    26 ms  core-kln-12Be3.forthnet.gr [213.16.247.17]
  5   25 ms   27 ms   28 ms  te-3-2.car1.Sofia2.Level3.net [212.162.46.109]
  6   65 ms    68 ms    65 ms  4.69.141.109
  7    71 ms    76 ms    73 ms  ae-10-10.ebr3.Frankfurt1.Level3.net [4.69.141.34]
  8    66 ms    66 ms    66 ms  ae-73-73.csw2.Frankfurt1.Level3.net [4.69.163.6]
  9    66 ms    66 ms    65 ms  ae-2-70.edge4.Frankfurt1.Level3.net [4.69.154.72]
 10    77 ms    78 ms    78 ms  4.68.111.146
 11   157 ms   156 ms   157 ms  if-7-2.tcore1.FNM-Frankfurt.as6453.net [195.219.50.2]
 12   156 ms   155 ms   156 ms  if-5-2.tcore1.AV2-Amsterdam.as6453.net [195.219.194.13]
 13   163 ms   160 ms   159 ms  if-2-2.tcore2.AV2-Amsterdam.as6453.net [195.219.194.6]
 14   173 ms   171 ms   172 ms  if-7-2.tcore2.L78-London.as6453.net [80.231.152.14]
 15   158 ms   159 ms   158 ms  if-20-2.tcore2.NYY-NewYork.as6453.net [216.6.99.13]
 16   156 ms   157 ms   155 ms  if-2-2.tcore2.MTT-Montreal.as6453.net [64.86.226.13]
 17   158 ms   158 ms   166 ms  if-0-2.tcore1.MTT-Montreal.as6453.net [216.6.115.89]
 18   159 ms   161 ms   157 ms  if-5-2.tcore1.W6C-Montreal.as6453.net [64.86.31.6]
 19   164 ms   164 ms   163 ms  66.198.96.50
 20   166 ms   166 ms   170 ms  192.34.76.2
 21   182 ms   183 ms   183 ms  199.91.189.234
 22   166 ms   165 ms   165 ms  199.91.189.25

Trace complete.

Prime time (19:00 BST - 21:00 EET)
 1     1 ms     2 ms     1 ms  192.168.1.1
  2    23 ms    22 ms    22 ms  bbras-llu-kln-15L500.forthnet.gr [213.16.246.12]

  3    23 ms    22 ms    22 ms  213.16.247.73
  4    25 ms    23 ms    23 ms  core-kln-12Be3.forthnet.gr [213.16.247.17]
  5    33 ms    35 ms    33 ms  te-3-2.car1.Sofia2.Level3.net [212.162.46.109]
  6    63 ms    63 ms    62 ms   4.69.141.109
  7    63 ms    67 ms    74 ms  ae-10-10.ebr3.Frankfurt1.Level3.net [4.69.141.34]
  8    64 ms    63 ms    63 ms  ae-93-93.csw4.Frankfurt1.Level3.net [4.69.163.14]
  9    64 ms    63 ms    63 ms  ae-4-90.edge4.Frankfurt1.Level3.net [4.69.154.200]
 10    62 ms    61 ms    62 ms  4.68.111.146
 11   158 ms   158 ms   159 ms  if-4-2.tcore1.PVU-Paris.as6453.net [80.231.153.10]
 12   298 ms   306 ms   319 ms  if-2-2.tcore1.PYE-Paris.as6453.net [80.231.154.18]
 13   332 ms   328 ms   325 ms  if-5-2.tcore1.L78-London.as6453.net [80.231.130.1]
 14   333 ms   316 ms   338 ms  if-1-2.tcore2.L78-London.as6453.net [80.231.130.122]
 15   314 ms   305 ms   303 ms  if-20-2.tcore2.NYY-NewYork.as6453.net [216.6.99.13]
 16   325 ms   328 ms   345 ms  if-5-2.tcore2.MTT-Montreal.as6453.net [216.6.99.30]
 17   332 ms   315 ms   308 ms  if-0-2.tcore1.MTT-Montreal.as6453.net [216.6.115.89]
 18   315 ms   329 ms   342 ms  if-5-2.tcore1.W6C-Montreal.as6453.net [64.86.31.6]
 19   319 ms   324 ms   317 ms  66.198.96.50
 20   336 ms   327 ms   316 ms  192.34.76.2
 21   316 ms   312 ms   305 ms  199.91.189.234
 22   322 ms   318 ms   321 ms  199.91.189.25

Okay. I have submitted these traceroutes to them, via support (and for the first one, which was in the initial ticket, they mercifully said they'd "forward it to the datacenter", something I honestly did not expect, despite being very rational to do so).

I have also told them the following, in an e-mail response:

You will notice that the 300+ ms in Paris or London (in bold above) are over what they should be (Europe-to-Europe connectivity should be at 100 ms or less, as you can see Athens-Frankfurt is "only" 63).

Since after Paris, the rest of the route up to Montreal is entirely within TATA (London, New York, and finally Montreal), I presume that TATA must be the primary / only provider of bandwidth for your Montreal datacenter.

So, it would be wise for Square Enix to take this up with them (TATA). Perhaps they can perform network upgrades or otherwise direct traffic from / to your Montreal datacenter through links which are less congested.


TL;DR: Square Enix / EIDOS Montreal should take up this issue with TATA, to alleviate problems with EU players in general, since the problem is not isolated or restricted to few cases only. It appears that a significant discrepancy exists in "prime time" hours, when links are severely congested and produce double the pings + packet loss, hence normal gameplay under these conditions is totally unfeasible.

Alternatively, Square Enix could decide to deploy some servers in UK, where Square Enix Europe is, and automatically stop worrying about TATA links

Conclusion

I hope you understood the issue from a networking POV, at least as adequately as simple testing can allow this to be done. There are more elaborate ways to test, however, the link congestion will not go away on its own, unless Eidos Montreal decides to significantly change their Internet bandwidth providers and / or take this up with TATA which seems to be the only ones taking EU players to them as things are.

[DenzelTaru]

I am done with Titan til this is fixed.. I am tired of the embarrassment of being hit off the edge when i am so far out of the aoe.. i am tired of getting hit by weights or making sure i run a certain way to make sure only one hits me.. and i am tired of the argument "you gotta move straight away you can't hesitate" look the big glowing thing on the floor is your indication, when that big glowing thing goes away you are too late, not half a second before it goes away, its not part of the fight mechanic is the games silly updating.

A good tip I received and trialed (with 90% success) was cast as soon as your feet leave the aoe, that forces the server to update your position, the only time this didnt work is if i was literally 1 foot out and thats probably just cos of casting not happening instantly and i guess that is also pushing it.

cast anything, it doesnt have to hit it just needs to cast so you update your location

[Goreth]

A good tip I received and trialed (with 90% success) was cast as soon as your feet leave the aoe, that forces the server to update your position, the only time this didnt work is if i was literally 1 foot out and thats probably just cos of casting not happening instantly and i guess that is also pushing it.

The detractors of this thread would call this good play style, but yea it's extremely annoying to have to do little tricks that are superfluous to the intended gameplay just to try to coax the servers to respond properly.

[Suellen]

*snip*

My husband has been examining our trace route from Los Angeles California and we are having the same issues with TATA. We are trying to see if we can use a program/service like smoothping to alleviate it until the data centers fix their shit (hopefully that's a when and not an if). This is not just a EU problem this is a problem for lots of people all over the place.

The irony is that our ping to the JP servers is steady, strong, and only about 5ms higher latency than the "NA" one. We probably should have played on the JP servers. who knew. :P

[Darkillumina]

The best part about all this was that the biggest issues all of us had in 1.0's last year were not game-play or mechanics related, but rather the massive server and positional lag and animation lock lag. They promised us that the new server infrastructure would alleviate these issues.

Guess which one wasn't fixed in 2.0? The server and positional lag. They ditched the superior battle system, got rid of animation lock by making a stale GCD system and failed to get the lag issue sorted out. At this point it seems all they did was add a red circle to tell players when they should theoretically move instead of visual cue's from the mobs and notes in the chat log.

GG SE.