Netcode.

  • Two Factor Authentication is now available on BeyondUnreal Forums. To configure it, visit your Profile and look for the "Two Step Verification" option on the left side. We can send codes via email (may be slower) or you can set up any TOTP Authenticator app on your phone (Authy, Google Authenticator, etc) to deliver codes. It is highly recommended that you configure this to keep your account safe.

edhe

..dadhe..
Jun 12, 2000
3,284
0
0
43
Scotland
www.clanci.net
With the advent of UTComp's 1.6 netcode that has lots of people creaming, and others worrying, what will happen with Envy?

Will they stick to the old fashioned buggy netcode that forces newstart players to relearn the game completely due to leading their or will Envy be on the receiving end of brand new, properly developed netcode that will make Online gaming seem like Instant Action (up to a point).

Oh yeah.. and think they'll finally get rid of the 'dud projectile' bug? ;)

Imho it will make the transfer of players from SP into MP smoother, and probably make the community more global and, because it's built into the game vanilla style, nobody will get to moan about it being bad or good.

No offence to the UTComp Dev, but putting a multimillion pound company onto the idea will surely bring better results.
 
T

Tournament0

Guest
I think it will be better than UT2004's Netcode, but, it might not make a difference. :hmm:
There will still probably be bugs somehow.
 
Last edited by a moderator:

T2A`

I'm dead.
Jan 10, 2004
8,752
0
36
Richmond, VA
that guy who posts a lot of useless crap and needs to be RO'd said:
One thing I'd like to see is better ping.
So you'd rather have Epic redo all the networking across the planet? Sounds good.

Making it more like IA would be great, but there needs to be a way around the high pingers having more of an advantage. I've never actually seen it in action, but Raffi, Brizz, et al, say that you have an exponentially greater chance of hitting someone with a high ping under UTComp's netcode.

I haven't seen these things that people complain about with UTComp's netcode at all. Granted, I don't have too much experience with it, but the games I've played on it went much smoother than UT2004's netcode. Hell, just yesterday I was playing some TAM on ViperBU (non-UTComp) and sent a flak shell at Raffi. He was pinging 36ish, I was 45-52ish, and my flak shell totally whiffed and went right by him, yet he gibbed as if it were a direct hit. Another time, same map, I sent two rockets straight into Skold's face and he shrugged it off like it was nothing. People complain that everything screws up under UTComp's code, but in my experience more crap goes wrong with the default code. Sure, I have seen some dead lightning shots, but I'd rather a lightning shot go bad than a precisely-placed flak shell any day.
 

spineblaZe

VFX Extraordinaire
Apr 8, 2003
2,423
0
0
44
MN
Visit site
Erm, UT2004's default netcode is rock solid. Having a high ping to servers has nothing to do with the default netcode, and I know the dud projectile bug has nothing to do with the default netcode either, because I've experienced that offline also.
 

JaFO

bugs are features too ...
Nov 5, 2000
8,408
0
0
I hope the new game simply tells players as they join in gigantic letters :
[sarcasm]
If you can't hit a bloody thing and get fragged all the time it's cause your aim sucks, you don't know what you're doing and your neighbour decided to download tons of pr0n-movies and share his music-collection to the world.

Oh ... and would you please turn off that p2p-program you're running.

signed,
Epic
[/sarcasm]
In other words :

The best netcode in the world can't compensate for the fact that the entire ff-ing internet has not been designed to handle real-time network communication.(period)


Other features of this new netcode would be :
- the game dumping you as soon as you mention the amount of 'lag' in your conversations
- doing likewise for the words : aimbot, cheater, hacker, pwned (and its variants)
 
Last edited:
T

Tournament0

Guest
edhe said:
How sensible.

But back on topic, would people rather see the 'adjusted' idea of the utcomp netcode, or good old fashioned style UT netcode?

Everyone wants it to be more like UT99.
I don't think this is a good idea. If you want it to be more like UT99, then go and play UT99.
This will be a new game.
:rolleyes:
 

JaFO

bugs are features too ...
Nov 5, 2000
8,408
0
0
No seriously. The internet just isn't suited to the kind of games we play.

The best thing Epic (or anyone else) can hope to achieve is somehow minimize the amount of data (and perhaps use some kind of compression) combined with redundant packages for error-correction.

In fact the best thing for a high-ping connection would be a game that runs as slow as possible with weapons that have an extremely low rate of fire.

There's nothing that can compensate for the fact that your connection has to cross larger distances across compared to other players.
Introducing client-side hit-prediction will only serve to shift the problem to the low-ping connections + it'll make cheating even easier.

Anything that claims to magically fix this stuff is make-believe.
 

edhe

..dadhe..
Jun 12, 2000
3,284
0
0
43
Scotland
www.clanci.net
Tournament0 said:
Everyone wants it to be more like UT99.
I don't think this is a good idea. If you want it to be more like UT99, then go and play UT99.
This will be a new game.
:rolleyes:

Did you actually read the post, lardass?

Do you actually think your comment is in any way relevant to the thread? Can you even justify your opinions?
 

Discord

surveying the wreckage...
Nov 6, 2002
639
0
0
Somewhere on Route 666
JaFO said:
Introducing client-side hit-prediction will only serve to shift the problem to the low-ping connections + it'll make cheating even easier.

It's not clientside, JaFO. Here's something I managed to drag from the flames over on INA in a gut- wrenching feat of derring- do :eek: :

The UTComp ReadMe said:
First of all, technically the word netcode is a tad misleading: UTComp's Enhanced Netcode does not have any impact on the actual network data being sent between clients and servers. The better word would be predictive weapon firing. However, due to the fact that gaming communities typically refer to this as netcode, we decided to use a term that players would easily be able to identify. Please do not write us angry emails claiming Enhanced Netcode isn't really netcode: we know.

As opposed to past "netcode" modifying mutators for the Unreal Tournament series, our netcode does not decide anything clientside that would normally be decided server side, or change the position players are displayed at. Our netcode works similar to Half-Life's, by having the server remain the ultimate authority.

The entire point of this is to remove the aim differences between various pings: while many players can deal with "lagged" fire, (having to lead targets depending on their pings) several players cannot, or would rather not if they had the option. Players using Enhanced Netcode can simply fire directly at targets, regardless of their ping.

Technical facts about Enhanced Netcode:

Hitscan Weapons:
How it works: The client draws its own client effects. The client sends info to the server. The server decides if the player would have hit if he were not lagged.

The immediate downside is 'through the wall' hits. These are not nearly as noticable in UT as you might imagine. Most experience with it comes from counter-strike, where the movement caused by damage is also reverted to 0 ping conditions. What this means is, that unlike CS, you won't get stuck behind a wall being hit over and over. You will be hit only if you would have been hit versus a lower pinger. This is currently capped at 350 milliseconds.

Another downside is that the client may register hits, or misses that will not actually register on the server. These are rather infrequent, but will happen nonetheless. The majority of these come from already existing innaccuracies in UT's netcode. In ut2k4, when you fire a shot, the shot is not only not guaranteed to be fired when you shot, but also is not guaranteed to be even aiming in the same direction, or from the same place as you shot. For example, if you move your mouse *QUICKLY* from left to right or up to down. There is basically a random cone of fire, even on weapons like the shock rifle and lightning gun, and even on lan. This is why snap-shots are so innacurate in ut2k4. Basically, the errors are made more visibly noticable.

Projectile Weapons:
Projectiles are reverted much like the hitscan weapns, but are currently limited to 75ms. You cannot be hit behind a wall by the projectiles.

Downsides: The immediate downside is that the projectiles are actually slightly harder to dodge vs a higher pinger. However, the magnitude of this effect is equal to the effect that the high pinger feels from the lower pingers projectiles. For example in the case of a 10 pinger vs a 75 pinger, they will both have equal difficulty in dodging projectiles, the approximate difficulty of a 85 ping player. Above 75 ping, the lower ping player has an advantage in dodging.

Shock Combos:
Your own shock balls are moved forward approximately as far as your ping, so that you do not have to lead them.

Downsides: The shock ball will look to be in a slightly different position when the actual combo explosion occurs. This is only visible to the firer, and has no real impact on play, but it looks wrong.

Other facts about Enhanced Netcode:

* If the server has Enhanced Netcode enabled, every player can decide whether they want to use Enhanced Netcode or not. Players who prefer the regular UT2004 netcode can simply check an option in their F5, Miscellaneous menu and revert back to the normal netcode. This will have no impact on other players. Server admins who prefer not to use Enhanced Netcode can disable it entirely for all players by either changing the setting in their server.ini, or by voting the option off via the Voting--> Gametype --> Settings menu.
* Enhanced Netcode is in most cases as accurate than the normal netcode is. However, many effects now take place on the client instead of the server. This has the direct result of causing some innacuracies that would normally be transparent to the player, visible.
* Enhanced Netcode does not make it any "easier" to hit targets. It only makes it so that you don't have to fire ahead of opponents. For players who are unable to effectively adjust to their pings, it will make a difference, but it's not going to give them any better accuracies than they would get under lan conditions.
* Enhanced Netcode does not favor higher ping players. However, if those players had difficulties aiming in front of their targets to compensate for their ping times, it will obviously minimize their usual ping disadvantage. As much as we'd like to make things 100% even regardless of pings, this just isn't possible. For now, minimizing higher ping disadvantages is as good as we can do.
* Enhanced Netcode will not currently work with custom weapon mutators. So, for the time being, you would be best served to only use the enhanced Netcode with the standard weapon loadout.


Short version: it moves your shot back in time (serverside) by the amount of your ping. The upshot, apparently, is that it makes it easier to hit and harder to dodge.

Seems like six of one and a half- dozen of the other to me, though I haven't played it yet. I don't think I'd want to be forced to use that. Dealing with latency is kind of part of the game... and you move a helluva lot farther in 60ms in UT2k4 than you do in CS or CoD. It's an interesting idea and I'm glad somebody was motivated to investigate, but I'd put it in the "still- too- new- to- UT- to- get- excited" category.


And please, guys. Nowhere is it written that just because you're posting about UTComp you HAVE to flame people. I thought this was the well- behaved UT forum...
 

Discord

surveying the wreckage...
Nov 6, 2002
639
0
0
Somewhere on Route 666
Srry for self- reply, but I find the subject interesting and that previous post was already waaay long.

Anyhow, here's some further Q&A from the same unbearable thread... Q courtesy of yours truly and A courtesy of Boksha, who is not a UTComp dev but who makes it his business to know these things nevertheless. Probably the best secondhand source on a subject like this you could want.

Discord said:
1. How weird does that look, exactly?
2. Do you guys like it?
3. Is it known yet whether this will be used in competition?
4. I'm still fuzzy on this -- did ZeroPing use clientside hit detection?
5. Is there an option to force all clients to use the same "netcode?" If not, is such an option under discussion?

Boksha said:
1. Not that weird, at least for instant-hit weapons. There is always some weirdness involved of course, like with the shockrifle, the shot itself looks instant-hit, but your opponent won't be bounced backwards until 60ms later when the hit gets replicated. The hitsound also gets delayed until the hit gets confirmed by the server.
I've seen some odd happenings with projectiles though, where the projectile jerked back in a Z-shaped path. I'm not sure if that's still possible in the latest versions of the netcode.
2. I don't. I prefer lag over ping correction, mostly because of afformentioned weirdness.
3. In Europe, it might be if both sides of a match agree to.
4. Some versions did.
5. You can only force enhanced netcode off by disabling it serverside. There's currently no way to force it on.
 

Sir_Brizz

Administrator
Staff member
Feb 3, 2000
26,020
83
48
[Apoc]Discord said:
It's not clientside, JaFO. Here's something I managed to drag from the flames over on INA in a gut- wrenching feat of derring- do :eek: :




Short version: it moves your shot back in time (serverside) by the amount of your ping. The upshot, apparently, is that it makes it easier to hit and harder to dodge.

Seems like six of one and a half- dozen of the other to me, though I haven't played it yet. I don't think I'd want to be forced to use that. Dealing with latency is kind of part of the game... and you move a helluva lot farther in 60ms in UT2k4 than you do in CS or CoD. It's an interesting idea and I'm glad somebody was motivated to investigate, but I'd put it in the "still- too- new- to- UT- to- get- excited" category.


And please, guys. Nowhere is it written that just because you're posting about UTComp you HAVE to flame people. I thought this was the well- behaved UT forum...
This is actually worse in alot of cases than client side hit detection, then.
 

Kantham

Fool.
Sep 17, 2004
18,034
2
38
Ban UNR3AL again.

Everything that you post must contain :hmm: or :con:

Never satisfied and you post like unr3al , you used numbers in the ''tournament>0<<< account'' that make me about UNR3AL TOURNAMENT0 that remind me that nerd that posted 100 post in a single day.

It is not the first time ppl blame him for his own post. And ppl here start to get Really anoyed.


Think twice before posting. Ping is a no-problem for epic so don't blame them for that.

Sorry for posting that in your thread Edhe.
 
Last edited:

JaFO

bugs are features too ...
Nov 5, 2000
8,408
0
0
I don't know about you, but I'd rather have to lead my targets then watch bullets/projectiles travel along weird z-shaped paths and similar weird/unrealistic effects.

I think what Epic (and any other company) should do is take the other route :
make instant-action appear more like 'real' internet/lan-games ...

So instead of trying to negate lag in on-line games we should be thinking of ways to add lag-like effects to off-line games.
That would make it easier for players to practice with the effects of a 'real' on-line game.
 

edhe

..dadhe..
Jun 12, 2000
3,284
0
0
43
Scotland
www.clanci.net
That's pretty arse over tit. You'd want top notch performance offline, not a bogged down game that's trying to be a pain just to try to train you to go online.

They're moving to less ping based weaponry (i hope) so maybe that'll help against those who can't fathom some latency differential. I still think if they can make a workable 'prediction client side drawing of effects' work well then that'd be a break through, on the otherhand i'd rather have raw and comprehendable good old-fashioned latency, than psuedo non-latency.
 

JaFO

bugs are features too ...
Nov 5, 2000
8,408
0
0
I wasn't thinking of using that 'simulated' lag to all off-line games.
I think an on-line simulator-mode does have some merit in a game that's going to have problems with lag (and similar effects), because I see no reason in trying to compensate that which can't be compensated.

Especially for a fast-paced game like UT trying to 'predict' stuff is almost useless.
Perhaps Epic can add minor prediction to weapons (say +/- 10-20 ms), but anything more and the risc of visual weirdness is just too great.

There's so much happening/changing in such a short period of time that any attempt to compensate is has to take care of an enormous amount of variables in order to reduce the amount of weird effects.

I think the only way such 'prediction' has a chance of succes is if there's a minimum amount of players, but then the data needed to transmit/receive changes becomes so little that lag becomes less of a problem as transmitting additional redundant data becomes easier ...
 

T2A`

I'm dead.
Jan 10, 2004
8,752
0
36
Richmond, VA
I actually wouldn't mind the option of being able to simulate lag and packetloss offline. I tried a couple days ago to start a dedicated server with ucc on my computer with lag and packetloss but I couldn't get it to work. The ping was always ~16 rather than the 150 I tried to force on it.

I suck online compared to offline, mainly because I've played offline for so long, that even with my low pings, my aim goes completely out the window in most cases. Learning to deal with ping without having to try to find a server that's playing the game you want with the right amount of people you want via a mutator or something would be great.
 

Discord

surveying the wreckage...
Nov 6, 2002
639
0
0
Somewhere on Route 666
edhe said:
'prediction client side drawing of effects'

That would be great if it would work, but AFAIK that would just cause warping.

Warping in online play is caused when the server predicts a high- pinger's location and sends that info on down the pipe to other clients. Then, the HPB turns out to be somewhere else entirely and the server updates instantly upon receipt of that info... et violá, the warp. Moving that prediction downstream to clients would probably just make the effect more pronounced? I dunno, but I think so...