.3 Second Positional Update Delay (Requesting Dev Response)

**CatofNineTails** · 10-28-2013 05:11 AM

being in Canada which is where the servers are and having the same issue pretty much shows this as not a distance from server problem. my ping is great most of the time. i always have Excellent connection listed to my world yet the moment i enter an instance with less than max players from my server I have additional lag if I do have all my server players the lag is about half of random groups add this to the .3 delay and tells can show to me at the point the servers have already counted down most of the cast bar yet on my screen the cast bar is full and the ground tell just appeared so I evade with lots of time left before the cast bar ends and wham get hit with the Damage.

**Asheilin** · 10-28-2013 05:24 AM

As someone who fits in the middle of both latency groups, I have good days and bad days where AoE is horrible for me or not. Doing Titan the first attempt was a bad day, as every time he landslides anywhere near me I would fly off; even when I was out of it before the cast bar was at 10% complete. Other days I can run out wait til 98% and run back in and not get hit.

I played wow for a long time. Upgrading my computer right before mop came out to a better graphics card (I easily exceed all of FFXIVs min recs) but I never had any issues with AOE there. Nor in RIFT, DC Universe, D&D Online, XI, or a slew of others that I attempted to get into at one point or the other. I'm looking forward to when addons make it, because there is always a "latency fix" I used wow's before the computer change; still did after but it didnt have a noticeable effect.

Wow I ran on high(custom) settings and got anywhere from 35-70fps in average play; 25mans I could tank to 2fps. FFXIV I run at 23-36 fps and yet the aoe and npc/mob load is the worst I've ever seen in any of the games I've played.

Fates are horrible for me, Instances run at 36fps when I remember to check it. the ping to check servers needs to be fixed or they need to accept that the memory has to be larger so that the local client can handle some of the issues (and therefore the ping can also focus on less and succeed more)

**ByakkoSeven** · 10-28-2013 05:34 AM

Errrrm, don't think this is still a problem, wireshark says it's 0.1sec positional updates, with positional updates also given with skill commands submitted. Tested quite a bit to make sure it was accurate. Seems this happened ~week or so ago, I know it used to be .3sec updates.

**Zdenka** · 10-28-2013 06:02 AM

Originally Posted by Sunarie

Are there work arounds?

The only real workaround is the make everything client side based instead of server side based. This would make everything you see on the screen actually happen.

An example of a client side based game is WoW, and you can compare how much smoother that game plays.

**Troop86** · 10-28-2013 06:20 AM

One of the first thing I learned in this game is that you need to move out of dmg around ~0.5 seconds before it lands. I thought it was how the game was built.

**Sybreed** · 10-28-2013 06:23 AM

I've been playing mmos since EQ1 and have never seen these sorts of issues to this extent. I'm running the game on its highest settings and still get between 40-60 FPS, with most boss fights being in the 50s, so it isn't an FPS issue. The character select screen always says I have an "excellent" connection, and I haven't checked on this game, but in WoW I get between 80 and 100ms latency; I'm on the east coast in America.

Despite all that though, I CONSTANTLY get hit by aoes even when I'm 2-3 character lengths out of the red zone when the cast finishes; so far Titan and the Chimera in Cutter's Cry have been the worst offenders.

On top of this I also get "hitching" issues, which is something I haven't seen since Vanguard: Saga of Heroes. For those that don't know, hitching is when there is a slight pause in the screen and your movement, that then rapidly corrects itself, giving your character the appearance of being a cart hitched to something that is stopping and going.

**Sybreed** · 10-28-2013 06:24 AM

Furthermore, I've been a high end raider in EQ, I played WoW for 7 years and was a raider in that game as a main and off tank nearly the entire time, on MUCH crappier computers than I have now, and I have never before seen bosses and encounters in a game that are ONLY hard because you're getting hit by boss moves when you clearly shouldn't be. I can't think of any encounter in any dungeon or raid in any mmo I've played before that this was so true of. It's really unacceptable from a company like Square.

**Sunarie** · 10-28-2013 07:51 AM

Originally Posted by ByakkoSeven

Errrrm, don't think this is still a problem, wireshark says it's 0.1sec positional updates, with positional updates also given with skill commands submitted. Tested quite a bit to make sure it was accurate. Seems this happened ~week or so ago, I know it used to be .3sec updates.

I don't know that I would consider Wireshark an official source.. as the .3 second update thing came from Yoshi himself, along with a statement that those with a great connection would still see a .1 second delay on the cast bar (perhaps Wireshark misinterpreted this?). Given the responses, and what I see in the forums, and the video examples, something is still an issue. Using skills after you get out of an aoe doesn't work for everyone, though it does seem to work for some. Please don't assume it's not a problem for others, when it clearly is, just because you don't have issues with it.

Originally Posted by Sybreed

I've been a high end raider in EQ, I played WoW for 7 years and was a raider in that game as a main and off tank nearly the entire time, on MUCH crappier computers than I have now, and I have never before seen bosses and encounters in a game that are ONLY hard because you're getting hit by boss moves when you clearly shouldn't be. I can't think of any encounter in any dungeon or raid in any mmo I've played before that this was so true of.

This is what really gets me. I'm not new to raids or dungeons. I've been playing MMOs since I was 18, got a job, and could afford them myself (mom and dad wouldn't pay a sub). That's a decade of MMOs for me.. from Shadowbane to WoW, CoH/CoV and Rift.. I've played almost every MMO on the market at some point (and I'm including a slew of F2P MMOs I've messed with), and done end game stuff for a good 5-6 of them. This is the only game I've ever been hit when I was out of a ground effect before the cast bar was done.

**Astarica** · 10-28-2013 07:59 AM

The lag is deliberate which is why this issue is depressing. It's not like SE is just using cheap hardware that randomly lags, or that their servers are in Antarctica so everyone has poor ping to it. The game chooses to do most thing on server side and the price you pay is lag, and in this case crippling lag. The problem is that on an ability like WotL, if you move the moment you saw the ability, you'll get out of it with at most 0.5s to spare. Assuming WotL is 2.0s cast, that's 25% on cast bar left, which is actually a pretty reasonable time to escape an AE. However, now add in 0.3s of lag (possibly more), and your margin of error is way smaller than what you think you have. Basically the last 20% of the cast bar might not really be there depending on lag, and that's what frustrates people because you can't rely on your screen to be accurate.

**ilitirit** · 10-28-2013 08:19 AM

Originally Posted by Ceol

Wouldn't that just let people make programs that tell the server a player avoided an AoE no matter where they are? There has to be some server-side validation.

That's why 99% of big-budget multi-player games use predictive netcode with rollback.
http://www.gabrielgambetta.com/fpm1.html

How a game like this exists in this day and age is beyond me...

To see rollback netcode in action, take a look at any FPS or a fighting game with GGPO-style netcode (this is P2P, but the concept is the same).

eg. http://www.youtube.com/watch?v=q0TZdY9_XHo

One client "thinks" the opponent opponent died so the "KO" sound plays. However, the opponent did actually manage to block in time so the match "rolls back" to a previous state and continues.

Thread: .3 Second Positional Update Delay (Requesting Dev Response)

Thread Tools

Search Thread

Display

Tags for this Thread