Limit lag compensation to 300 ms
biz
Join Date: 2012-11-05 Member: 167386Members
Join Date: 2012-11-05 Member: 167386Members
looking for a way to cap lag compensation.
ideally it would be defaulted to something (eg. 100 ms) and people who want to adjust that number can run custom settings...
i've seen games that are much more responsive put it even lower than 100 (eg. cpma / quake live / warsow)
it just seems impossible to play NS2 as it was designed to be played atm
NS2 is not about deathmatch. it's supposed to be about reacting to enemies.
1 example:
put a skulk in a room and a marine in an adjacent room.
the marine doesn't know the exact location of the skulk.
the skulk doesn't know the exact location of the marine.
the skulk and the marine both know the exact location of the doorway between the rooms (and are staring at it)
this happens on every map in basically every round multiple times during the first 5 (and most influential) minutes.
when this scenario plays out as the game was designed / balanced (low-ping / LAN enviornment):
1) if the marine moves forward, the skulk will have an advantage because he sees the marine but the marine still has to find where the skulk is and aim there
2) if the skulk moves forward, the marine will have an advantage becuase he will see the skulk before the skulk will find him or reach him
when this scenario plays out with a high-ping marine:
1) if the marine moves forward, the skulk can be discovered and put at a disadvantage before the skulk's client even registers that the marine entered the room
2) if the skulk moves forward, the skulk will have an advantage because the skulk will be biting the marine before the marine's client even registers that the skulk entered the room
when this scenario plays out with a high-ping alien:
1) if the marine moves forward, the skulk can be discovered and put at a disadvantage before the skulk's client even registers that the marine entered the room
2) if the skulk moves forward, the skulk will have an advantage because the skulk will be biting the marine before the marine's client even registers that the skulk entered the room
note that it doesn't matter which side actually has high ping. the high-ping game simply favors whichever player moves forward.
excessive lag compensation turns the game upside down. it's not just a laggier version of the same basic game - it alters the decisions and tactics in a bad way
i'd rather play a game that plays out as it was designed & balanced
if I need to give myself more time to react when I have high-ping, then that's perfectly acceptable (and commonplace in every other FPS game...)
maybe i'll play more of a support role and less of a twitch combat specialist...
instead, the game wants us to throw strategy out the window and just try to do damage asap & be constantly moving
it doesn't help that the game is like twice as laggy as the pings would indicate. there's definitely some extra latency in the netcode making things worse
ideally it would be defaulted to something (eg. 100 ms) and people who want to adjust that number can run custom settings...
i've seen games that are much more responsive put it even lower than 100 (eg. cpma / quake live / warsow)
it just seems impossible to play NS2 as it was designed to be played atm
NS2 is not about deathmatch. it's supposed to be about reacting to enemies.
1 example:
put a skulk in a room and a marine in an adjacent room.
the marine doesn't know the exact location of the skulk.
the skulk doesn't know the exact location of the marine.
the skulk and the marine both know the exact location of the doorway between the rooms (and are staring at it)
this happens on every map in basically every round multiple times during the first 5 (and most influential) minutes.
when this scenario plays out as the game was designed / balanced (low-ping / LAN enviornment):
1) if the marine moves forward, the skulk will have an advantage because he sees the marine but the marine still has to find where the skulk is and aim there
2) if the skulk moves forward, the marine will have an advantage becuase he will see the skulk before the skulk will find him or reach him
when this scenario plays out with a high-ping marine:
1) if the marine moves forward, the skulk can be discovered and put at a disadvantage before the skulk's client even registers that the marine entered the room
2) if the skulk moves forward, the skulk will have an advantage because the skulk will be biting the marine before the marine's client even registers that the skulk entered the room
when this scenario plays out with a high-ping alien:
1) if the marine moves forward, the skulk can be discovered and put at a disadvantage before the skulk's client even registers that the marine entered the room
2) if the skulk moves forward, the skulk will have an advantage because the skulk will be biting the marine before the marine's client even registers that the skulk entered the room
note that it doesn't matter which side actually has high ping. the high-ping game simply favors whichever player moves forward.
excessive lag compensation turns the game upside down. it's not just a laggier version of the same basic game - it alters the decisions and tactics in a bad way
i'd rather play a game that plays out as it was designed & balanced
if I need to give myself more time to react when I have high-ping, then that's perfectly acceptable (and commonplace in every other FPS game...)
maybe i'll play more of a support role and less of a twitch combat specialist...
instead, the game wants us to throw strategy out the window and just try to do damage asap & be constantly moving
it doesn't help that the game is like twice as laggy as the pings would indicate. there's definitely some extra latency in the netcode making things worse
Comments
The game runs on Lua, and I'm not sure how much more performance can be squeezed out without major engine changes (major meaning "tear out Lua and put something faster in there"... so pretty major)
the game just feels way laggier than it should be, and it's impossible to quantify with just 1 client
Happens when I skulk as well, land a bite, blood spatters but no damage.
It can already be customized.
Was wrong. Refer to ghouls post below.
Skulk moves forward, gets 4 gl nades to the face before the marine's client even registers that the skulk entered the room
When i have to join a Russian server and play because in Asia , there is all empty server Japan
Those russian kick me because i have 350 ping >> Still owned those Vodka drinker . Lucky things is my team Marine vote F2 - say No to as-*hole on internet
Nope the warning will come up as soon as you change the net vars in any way. You can hide that warning so maybe you already did
Tweaking the net vars can improve the game-play depending on the given server but can also make the the hole situation a lot worse than normally when you set them not correctly.
..that ain't no russian server.
And i'm pretty sure most of them don't drink no vodka either.
VB *spit* maybe, not vodka.
All pings are around 50.
My millage may vary. If you can compare packets / update to wagons. It would be like put more wagon on a slow train. It won't go slower in this case, but not faster. In fact it can make things worse.
There are 3 key elements.
-The clients (PCs). Some things can be done to make things faster but basically, sends update to the server and receive updates.
-The network (Internet). Transports all packets to the destinations. Client - Servers relations. Only ISP can improve their network and if they want to (it require maintenance...). Like make some game ports (ex: 27015) a priority.
-The server response time. If all packets get there fast the server takes some time for calculations. It takes time to process one incoming update and sending the updates. Depending on this delay, even if you have a good ping (from your location to the server: only network) it may be "laggy".
This last item seems to take huge loads of time on NS2.
A test can be implemented when you make that kind of programs. If none was done, I'm afraid it's not a CDT priority to make such test. It would only reveal with precision the real delay we all know. Not sure it's gonna change anything.
The test would be to create an update (client to server vector) and see when the update results comes back. An action like setting up a mine on a wall. The entity doesn't exist in the "world" as an entity as long as you carry it. This little change can be measured. You set up mine, the sever receive update; make some changes and send back the update for your client to create the entity in your local world.
The server calculation time is almost a zero. In fact, one iteration of the game logic fits in the time between updates (by design). If it does not fit then the server admin has completely overloaded his server machine with some other stuff.
The very nature of increasing update rates means you are updating more frequently, and thus information is transmitted at an increased rate : In other words, information is transmitted faster.
We don't need to setup a test to determine such a thing.
We've known exactly what they are for years, the inherent default delays are all very measurable and are now configurable.