Please fix your netcode
NS-Soldier
Join Date: 2013-01-16 Member: 179856Members
Join Date: 2013-01-16 Member: 179856Members
I am tired of seeing players with over 150+ ping ghosting/dominating at #1 of the scoreboard.
It's really hard to hit these so called "good" players that has no hitbox and they think they are good.
don't believe me? go play on a high ping server outside of your country, then you'll be dominating at #1 spot and everyone calling you hacker.
It's really hard to hit these so called "good" players that has no hitbox and they think they are good.
don't believe me? go play on a high ping server outside of your country, then you'll be dominating at #1 spot and everyone calling you hacker.
Comments
https://trello.com/c/TIyg0TRQ/456-hit-registration-de-sync-issue]
Hitreg may only be part of the issue.
https://trello.com/c/jRYpkdnm/259-bug-4910-null-move-insertion
Also, this post from ironhorse
http://forums.unknownworlds.com/discussion/comment/2240381/#Comment_2240381
"Oh, and gravity will now work for players that loose their network connection."
No more spinning wildly out into space when servers redplug hahaha
TBH this is the feature that keeps bringing me back.
Learn to aim
Yea ... it is quite interesting to run two clients and a server locally, trigger a jump on one client then watch how long time it takes for the second client to see the jump.
While most of that delay is true of all lag-compensated games, in NS2 is just matters a lot more because of the extreme speed and ranges at which you do combat compared to other games.
The "most of" qualifier is actually because NS2 has a three-stage rendering pipeline*; it might be that some other games has a two-stage pipeline which cuts 1 frametime of the delays.
This has one practical effect in that high fps matters more in NS2 than in other games... and since it has been harder to get that higher FPS in NS2, it can add to the effect.
But it has gotten a lot better lately, and 275 will see a nice little boost as well.
/Mats
* NS2: Visible frame <- graphics rendering <- game rendering <- input** handling frame = 3 frames from input to render
Other: Visible frame <- graphics rendering <- input and game rendering = 2 frames from input to render
** "input" == your input and the input from the server
I agree 10 times (and a half).
I see some guys that have a ping <40 on the scoreboard but are slow like if it was 150 (Benny Hill show effect). I fear and suspect they use VPN to induce lag. An Australian (just an example) using a VPN located in Germany and then he connects to a UK server gives the following. The use of VPN do not make the connection better. It just hides the real entry point in the network. The Traceroute will stop at the VPN hosting node. Instead of seeing Client->server we see VPN->server. The real path is Client->VPN->Server, which won't make the ping better.
It would be really really nice to have at disposal several things
-A "netgraph" like the one on NS1/HL1. Not only for client but for everybody. A second scoreboard that provide the last 5 minutes. Something like that.
-A flag in the scoreboard if a client is using a VPN. The list is easy to build / find. If i remember correctly there is an official list of all VPN in the world (it help preventing attack, etc.).
-A FPS average for every client. Sometime it's not the network that is bad but the PC. So it helps to see why this one is warping.
how can i aim someone who is ghosting aka disappearing every half a second
skills man skill
Nah, that warping is quite oddball sometimes. High ping with very fast moving aliens. that combination will kill lag compensation in any game... Still how did Quake and GoldSrc manage that stuff on modem/isdn..............................
If I remember correctly, both teams could have pings around 300 or so.
By practicing! Idk maybe it doesn't bother me cause I don't use draw damage, who knows
You can practice all you want. You can get really really good at "reading" other players, figuring out where they're going to go before they go there. That's skill. And that has almost no bearing on being able to hit a "slideshow" player. There's no way to predict when a player might suddenly just freeze in midair or when they might suddenly teleport back to where they should be.
it's more funny when I 80 pinger walks around a corner then gets killed by 300 pinger that somehow its magic bullets go around the corner as well
That's just a result of the latency between computers... not sure a solution is physically possible. Only way is to reduce the latency... somehow (better internet, closer location, or a high-ping kicker :P ) You may have been around the corner, but on the other guys screen that kill was legit, it just hadn't gotten to you yet.
Never denied that there were something wrong with it - too many videos showing that - but as noone could find a way to reliably reproduce it, finding it was just too hard.
Also, as the game has good enough hitreg to be playable, it was not seen as critical.
Just pure luck that Remi started to wonder about what happens when the client misses an update and starts to extrapolate; once we started digging in that direction, it turned out that the server happily used all the network updates it had sent to the client, even if the client had never seen them.
Which explains why it was so hard to reproduce it; only if you had happened to drop just the right network updates at just the right moment...
Fixing it just means having the client note what network updates it used when it performed the move and force the server to only use those updates when it reproduces the move.
Now, that only explains one part of the hitreg issues - there turned out to be a bunch of other smaller instances where the client used non-networked data to produce animations, which of course caused animations differences between client and server.
Most blatant for wallwalking skulks.
Animation differences are not the same issue as the network issue though - the network issue can cause a client-side bodyshot to turn into a server-side miss, while animation issues will pretty much only turn near misses into just hits or vice-versa; the major impact on the game would be faulty client-side hit feedback (blood and no damage or the other way around).
Anyhow, those fixes are still waiting to be verified and will probably be found in b276; we will want to release b275 as soon as we have verified that the upgrade to VS2013 seems to have worked.
Plenty of goodies in 275 that we don't want to delay any more than we have had to.
Glad I was correct in that trello card is more what he was experiencing.
Nordic: the pro Trello guy. Knows everything you need to know. A question about what's going on in the CDT, their future projects? Call Nordic, he's at your disposal 24/7.
A sign I have too much time on my hands and possible obsession.
i tried to play a few rounds... just end up good K/D ratio but not at top spot or being called (lag) hacker
Maybe you needed to go on an even higher ping server?
edit: Noticed the player count now. Those k/d ratios aren't even good.