Fixing position Bug

[MCocktail 183]

There are many bugs, such as some skills failing to work, graphical errors, or typos. Even the loss of ontire quest texts did happen. But those werent ever as annoying as the

 

Position Bug!

 

Anyone knows how devastating a last hit, or better, a hamstring from 7 tiles across is. Many times.

Apparantly this bug was just explained with bad internet. But its quite easy to reproduce: Walk around. The more you do the more the possibility that the server location is different from the one in your client.

 

This bug was never fixed.

 

 

This is not a regular "plz fix dis dev y u do dis gm" topic. This is an idea on how ti do it!

 

So lets do so!

 

Here is an idea:

 

Players can choose if they want high or medium or low traffc.

Unlike other options, these settings should be backed up on to account instead to the device, so that the server can handle it easier, since client setting may cause the server to check everytime what the settings are, before loading the map. This is adding more loading time, more traffic and thus would be better of of saving THAT Setting in server.

 

Low traffic users have the current system. Since the lack of sending of positions may be a reason for ws's insanely low mobile cell data usage, this option has to be kept for low end devices and low traffic internet users.

 

Medium Traffic users have the server constantly send all the char positions, not just on map load, but on every Approx 1-2 seconds. If the client coordinates are different from the server ones, the client coordinates are violantly changed, and to ensure that the user sees what is actually happening, the position is violently drawn back to the server position. Looks like one would flash away. But at least, that way what we see is (mostly) true. And lets get that "mostly" away... With

 

High traffic users. Instead of every one second, their map positions refresh whenever possible.

 

 

So yeah. This is a concept on how to fix this.

Seems like noone at aigrind HQ cared about that, at least now they get their idea served.

 

Because getting hammed from 7 tiles away. Is. No. Fun.

[benjaminbt 325]

Agree! Totally agree! It doesn't even sound impossible to do.

[Lobsel 9]

Yes that would be kinda awesome, especially in such places as lab

[Gladiator 1145]

Not only lab, duh!.. Even when you afk, sometimes you afk  showing your face to the screen but actually, your back is to the screen!!! 

 

Yes! Fix pls, if can, it'd be awesome.

[MCocktail 183]

Not only lab, duh!.. Even when you afk, sometimes you afk showing your face to the screen but actually, your back is to the screen!!!

 

Yes! Fix pls, if can, it'd be awesome.

If u go out and get back in fast enough not to let the char vanish, you still have your facing looking in the screen. Others who watch, would, since the char ist destroyed, still just seee the Direction, EG. Right. That way you could conlucde that the Direction where the char is facing doesnt even seem to existieren in server

[MCocktail 183]

Wubwubwubwub da- unzunzunza bumpabumpabump. ~

[Nosotraes 350]

Wubwubwubwub da- unzunzunza bumpabumpabump. ~

[Kaleidoscope 142]

I'm used to the "position bug", but it's still a thing that bugs a load of ppl, me included.

Agree with the suggestion.

 

But it's still easily said, hardly done.

Maybe it's not even that easy to fix it.

Edited November 12, 2015 by Raezer

[MCocktail 183]

 

Wubwubwubwub da- unzunzunza bumpabumpabump. ~

Wubwubwubwub da- unzunzunza bumpabumpabump. ~

[MCocktail 183]

I'm used to the "position bug", but it's still a thing that bugs a load of ppl, me included.

Agree with the suggestion.

 

But it's still easily said, hardly done.

Maybe it's not even that easy to fix it.

As someone who does a bit of programming, it should be possible. Quite easily. (Recieving data isnt that hard if u understood stream protocols.)

 

Especially with the money they get from us they should be able to fix at least this.

 

Maybe even easier than having your own client coordinate system

Edited November 12, 2015 by Molotow Cocktail

[Kaleidoscope 142]

As someone who does a bit of programming, it should be possible. Quite easily. (Recieving data isnt that hard if u understood stream protocols.)

 

Especially with the money they get from us they should be able to fix at least this.

 

Maybe even easier than having your own client coordinate system

sounds fair

[Ninja Owl 210]

It would solve this problem if the game client waited for the server's response before moving the character(s). This solution would cause slight delay between actions especially with slow Internet connection but there wouldn't be any position bugs.

I'm no MMO developer so I don't know if it works, but it seems logical to me.

Edited November 13, 2015 by Ninja Owl

[MCocktail 183]

It would solve this problem if the game client waited for the server's response before moving the character(s). This solution would cause slight delay between actions especially with slow Internet connection but there wouldn't be any position bugs.

I'm no MMO developer so I don't know if it works, but it seems logical to me.

That would work too:)

 

And the system u mentioned is allready applied to skills, so why not do it on map entities

[MCocktail 183]

Bump

[nabnecro 304]

I agree on this.

 

We all known each time someone complains about this, Roland comes in and say it's just because of the lag.

Well, of course you can blame lag, but at the same time, the game should handle things more elegantly in the case you don't have a perfect internet connection (with minimal delay and zero packet loss).

 

Also, lag is not the only thing triggering position desync. If you play on PC, click on the crossing of the map you are, then drag the game window for 1 or 2 seconds and see what happens.

 

 

Edited November 17, 2015 by nabnecro

[Ikafati 148]

Every single online game has this "position bug". Warspear simply solves it poorly

[Lobsel 9]

But it's probably the oldest and the more reccurent bug of this game. It's not like a specific item appearance...

Fixing it would probably brings tons of anothers bugs dépendant of your device.

[maraschino 9]

Every single online game has this "position bug". Warspear simply solves it poorly

 

I agree. The WS developers can't eliminate latency, but they can probably improve software and server hardware latency and they can account for it in how skills work and players interact. There have been other obvious bugs that lingered for years (e.g. casters running to where they cast dark circle or quake) before being quietly fixed, so it would not be a surprise if there are lag-related programming issues waiting to be resolved.

 

I made two accounts on two devices and moved the players around the screen to test lag/latency. One device was a mobile phone with LTE connection & the other is a desktop with wired network. I'm on west coast USA using EU servers, so lag/latency should be on the high end. A character on one device could move 1.5 to 2 tiles before it appeared to move on the other device, about 0.6 to 0.8 seconds.

 

I'm not sure how this amount of lag can generate hamstrings from 6-7 squares away. If my BD and my druid are standing two squares apart, and both start to move in the same direction at the exact same instant, immediately after the server gets location information, my druid could move ~ 2 squares before seeing any movement from the BD and so could see a separation of 4 squares. Meanwhile, if the BD position is updated at the server an instant before the druid, the server could locate the BD next to the druid and able to hamstring. This would require everything to go wrong almost perfectly, a very unlikely case, yet it happens all the time. So, there must be more to it, maybe extrapolation of velocities by the server (e.g. start with BD and druid moving towards each other) or some other programming issues that make this problem worse, e.g. with how the server handles hits when a melee class is "locked on" to an opponent.

 

Mostly speculation but fun to try to figure out the puzzle. :-P

 

Good article about this issue here, though mostly from the perspective of first-person shooters:

https://en.wikipedia.org/wiki/Lag#Solutions_and_lag_compensation

[Ikafati 148]

Try demanding arena and get that map with a huge square in the middle to run around. Any class run and a ranged class try to catch up to him. Start like 8-10 squares away. Eventually, for some reason, you'll catch up (you'll shoot an auto attack)

[Kaleidoscope 142]

I agree. The WS developers can't eliminate latency, but they can probably improve software and server hardware latency and they can account for it in how skills work and players interact. There have been other obvious bugs that lingered for years (e.g. casters running to where they cast dark circle or quake) before being quietly fixed, so it would not be a surprise if there are lag-related programming issues waiting to be resolved.

 

I made two accounts on two devices and moved the players around the screen to test lag/latency. One device was a mobile phone with LTE connection & the other is a desktop with wired network. I'm on west coast USA using EU servers, so lag/latency should be on the high end. A character on one device could move 1.5 to 2 tiles before it appeared to move on the other device, about 0.6 to 0.8 seconds.

 

I'm not sure how this amount of lag can generate hamstrings from 6-7 squares away. If my BD and my druid are standing two squares apart, and both start to move in the same direction at the exact same instant, immediately after the server gets location information, my druid could move ~ 2 squares before seeing any movement from the BD and so could see a separation of 4 squares. Meanwhile, if the BD position is updated at the server an instant before the druid, the server could locate the BD next to the druid and able to hamstring. This would require everything to go wrong almost perfectly, a very unlikely case, yet it happens all the time. So, there must be more to it, maybe extrapolation of velocities by the server (e.g. start with BD and druid moving towards each other) or some other programming issues that make this problem worse, e.g. with how the server handles hits when a melee class is "locked on" to an opponent.

 

Mostly speculation but fun to try to figure out the puzzle. :-P

 

Good article about this issue here, though mostly from the perspective of first-person shooters:

https://en.wikipedia.org/wiki/Lag#Solutions_and_lag_compensation

For latency testing I'd rather take two devices with the same connetion.

[Nosotraes 350]

For latency testing I'd rather take two devices with the same connetion.

There is no such thing as the same connection. Yes, you might be connected to the same network, ethernet, but there are always little lags for either one of the devices. Especially when using LTE/HSPA/WiFi networks.

[maraschino 9]

Try demanding arena and get that map with a huge square in the middle to run around. Any class run and a ranged class try to catch up to him. Start like 8-10 squares away. Eventually, for some reason, you'll catch up (you'll shoot an auto attack)

 

Yeah, weird. I don't remember seeing that before the expert skills updates.

 

For latency testing I'd rather take two devices with the same connetion.

 

I just tested it with both devices on the same router. Same results. I assume most of the lag comes from the distance to the server (west coast USA to EU) and the processing on the server. Would be interested to hear if there's a lot less lag for people in EU.