I'll believe it when i see it.

"We believe it's a clientside extrapolation error, not an issue with the unlagged code. "

It's not us, it's you.

:/

What part of quake do you own?

?

My comment was a quote from SyncError and my opinion of what it seemed he was saying, nothing more.

If you must know, I own a nice piece of quake down by the river, where I park my van.

except he wasn't saying that at all...

Except I wasn't serious. It was more a play on words that was formatted poorly.
D'oh!

You did not understand what he was saying. The client means Quake Live on the player's machine and its representation of the world. In simpler terms:

Quake Live was displaying the enemy position incorrectly on player's machines.

When he says there's no problem with the unlagged code I guess he means that the code on the server that looks back in time to check for hits that should have happened is working correctly.

I don't doubt that, but see my post above.
I wasn't being serious and didn't think my response would've been taken as such. :P

learn to write or people will not take you seriously.

I know how to write fine, but I post from work and don't exactly have time to proofread everything for cogency. :)

when they gonna fix the problem with people going through and hanging inside each other?

Its fun just hanging inside someone else..

That's called sex. Sex is awesome. How dare you petition to remove sex.

You'll be able to play plenty of quake down by the river, when you're living in a van down by the river!

erok foley, motivational speaker

:D :D :D

"His name's Matt Foley and he's been in the basement drinking coffee for the last four hours."

Don't call it a hit box fix. It's not so complicated that you cannot understand. The issue is with where the client draws the enemy.

title must be short, so everyone understand what we're talking about, even better than "client misplacing hitbox due incorrect extrapolated position" :>

It's not the hitbox. It's the image of the opponent. They aren't the same thing, in Q3 if you change time nudge you can move the opponent's displayed position around, nothing happens to their hitbox. It's not a helpful way of communicating, it's just wrong so please fix it.

'Enemy position fix on its way' would be more accurate or 'netcode fix' or 'extrapolation fix'.

ok i made it clear :D

\o/

I wonder, when id will stop talking and start doing instead. :/

I wonder when will id start and defeat the laws of physics, making 50 ping acting like 10

Have you ever played saurbraten? Shit gameplay aside, that game is totally playable with 200 ping and even more. I think pretty much every recent game i can think of has been more high ping friendly then ql is...

That is probably the case because quake is about accuracy.

Haha n1:D

I even played chess, once
once in '66
no plus for you, though

dont xerp me bro

Why would the client be extrapolating if the server is using backwards calculation to determine hits? If a combination of the two are being used, this whole perceived hitbox delay issue is just an adjustment of figures,no ?Or am I misunderstanding something here?

"Why would the client be extrapolating if the server is using backwards calculation to determine hits?"

To try to show you where the opponent is rather than where they were.

i was thinking the same when i discovered cpma uses backward reconciliation.

I mean, on one hand, the client tries to draw the opponent where he is rather than where he was, on the other hand the server determines hits based on where the opponent was when you shot him.
Even though it obviously works well in cpma (xerp whiners notwithstanding), it doesn't quite make sense to me.
Also, why doesn't cpma have round-the-corner hitscan bs?

It does, it just uses a smaller amount of backwards reconciliation so it's less noticeable.

what he means i think is that the client is over extrapolating. as in drawing the enemy further forward along the direction from a previous frame than he ought to be, which is resulting in all the "i hit behind you and yet it registered" errors.

How does this explain the feeling that you can constantly be hit even when you just stepped behind the corner?

That's backwards reconciliation and ping. The shot was fired at you when you weren't behind the corner, the server checks for the hit and awards it, then once you have moved around the corner your client's informed by the server you got hit so you think the shot happened after you went around the corner.

i think atleast part of it is also placebo. because people (and by that i mean mostly me :) ) assume if u cant see the enemy he cant see you, which isnt necessarily true. dunno for sure but im guessing ur pov is taken from a single point within ur model/hitbox whatever? which would mean that even tho the line of sight from that point has gone beyond the corner there are still parts of you visible.

not saying the netcode is perfect but just that the problem is exaggerated. and as far as i can tell this doesnt happen all that much more in ql compared to cpma.

love this comment: short, understandable and contains all the necessary info. The should quote it in their next news, so maybe some more people get it.

its still annoying as fuck as it makes timing your own moves close to impossible sometimes.

It feels like there is something other to it than just ping. Since even playing with 14 ping, the shots hit where I was 500 ms ago (give or take).

It should also be your opponents ping.

I don't recall this happening in Q3. A common tactic that I use in a rail fight is to go behind an obstacle the very moment the opponents rail is recharged. However, now this tactics never works since he can just fire behind me and hit anyhow.

Not quite, that would only explain it if your own ping was sufficiently high to cause the movement prediction of your own position to be subject to latent corrections, but the 'getting hit after taking cover' phenomenon happens to me in a big way when I have a 20ms ping, which means the hit calculations must be done clientside of the HPB that shot me..

You didn't understand what I wrote.

Erno. If you have a 20ms ping and you incur lengthy retrospectively corrected hits from some HPB's railgun that is clientside hit calculations and that is LIES, 'nough said.

You have no idea what you're talking about. QL uses backward reconciliation on the server, not clientside hit detection.

This was actually posted last month, and on this very site.

Extrapolation error
arQon fixing it

you are correct sir, it's been a long time since they found a possible fix for the issue..... considering they haven't updated anything in Quake Live, i'd say they're on vacation.

remove unlagged = win

YEAAAA LETS PARTY LIKE ITS 1996 AGAIN!!!

it's not real "unlagged" code anyway... which would be 100% clientside like xbm in q3 and some other small mods.

i m sure that they ll make the game unplayable for >30 ping players... :D

thats life. high pinger will always get a disadvantage. deal with it

actually i could hit much better before they removed xerp long time ago
atm it's pretty unplayable for russians, dunno what changes

THE FIX, LIVE: Coming to a Theater Near You @ 12 December 2012!

History tell us the fix will be done and out by the time QL playerbase matches Q4 playerbase after it became decently playable.

so true

Or it'll be even smaller.

WHERE'S YOUR GOD NOW, WHINY OSP FAGGOTS

***Warning IX patrol on this thread***
***Refrain from using term HITBOX***

Oh yeah, don't try to ask for any info about this issue on the forums, this is what i got last night:




You have been banned for the following reason:
Useless Post, Bumping after Warning

Date the ban will be lifted: 06-29-2009, 04:00 AM

Brought to you by SyncError.

the ql forum consist of a bunch of noobs. Unfortunaley ID give them more attention in the forums then us. Thats why I start ridiculos troll threads to get attention.

btw the game is nearly dead but the ql forum is quite active ^^

btw the game is nearly dead but the forum is quite active ^^

That's how all Quake games work. =P

I believe syncerror is often lurking esr

To be honest, seeing how ESR people expect netcode to work , I would be tempted to call them noobs as well.

We have been over this a trillion times really.

200 ms is 200 ms is 200 ms.

They can lie visually about model positions in order to smoothen things out, in the end it there is still going to be a 200 ms latency. Obviously this can and will create cases where the engine seems to be lying to you....but that still a lot better than having the pure 200 ms represented 100% of the time.

It probably requires more tweaking on their part for sure, but I still think most people who constantly complain about the "netcode" are just living in the wrong decade or something.....you are requesting features that won't see the day before 2050.

Im surprised SyncError stepped forward with this however. Opening Pandora's Box :)

200 ms is 200 ms is 200 ms.

Except no one here is talking about 200 ms. I personally have around 7-8 ping to the QL server I play on and the problem is even more apparent there than if I've had 200 ping. As the lower your ping is, the more you have to aim behind the model to hit (which makes no sense).

I totally agree the values are out of range at the moment. The remark
was more aimed towards people who will STILL hate after the fix.

Oh yes.....they will be there, trust me :)
( they probably also "feel" the lag with 0 ping ).

That being said, your case is probably considered an exception in the code anyways, since you are in one of the rare places where Wan/Lan produce similar metrics pingwise.
( talking way out of scope here, and also out of my hat, but that ESR fair )

you are requesting features that won't see the day before 2050.

Client-based prediction, anyone?

I don't have an assload of experience with programming, but to say that, with this beauty of a feature, netcode is incapable of becoming that good is quite pessimistic if you ask me.

won't happen!

any chance we could get an update about the error, or an estimated fix date? thanks in advance.

Read #59

http://www.youtube.com/watch?v=fj5d_-01nBg&am...annel_page

the tail, man, it's the tail