<!--quoteo(post=1822269:date=Jan 7 2011, 01:20 AM:name=FocusedWolf)--><div class='quotetop'>QUOTE (FocusedWolf @ Jan 7 2011, 01:20 AM) <a href="index.php?act=findpost&pid=1822269"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->The marines are refraining from fighting the skulk and are now preferring long range attacks and stationary camping tactics where possible, because distance is the only advantage of the marine and close quarters is certain death.<!--QuoteEnd--></div><!--QuoteEEnd-->
This right here is the entire point of range vs. melee. This is what it's supposed to be when it's lmg vs. skulk...
that being said, somebody mentioned reticle position being sent from server and client, if they can stop that and just have the server be a slave to the client mouse movements, then you would have an easier time swinging your lmg around in close quarters to get the kill.
<!--quoteo(post=1822304:date=Jan 7 2011, 05:08 AM:name=Uzguz)--><div class='quotetop'>QUOTE (Uzguz @ Jan 7 2011, 05:08 AM) <a href="index.php?act=findpost&pid=1822304"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I don't think it's been mentioned in this thread yet, but part of the problem with mouse responsiveness is that your aim direction replicates in both directions - that is, <b>the server periodically tells you where you're aiming.</b> It was especially noticeable back when the "backlog on join" problem was around - if you turned around while your client was catching up, the outdated "updates" from the server would snap you back to the way you were facing when you joined - but as far as I know it's still like that now.<!--QuoteEnd--></div><!--QuoteEEnd--> I think they might have fixed this but it's true. The server should only snap back on position and not orientation. Otherwise, people with good twitch reflexes are if anything confused about why quick precision aiming is so tough and this would be exactly why.
NOTE: <u>In the HL engine when there was huge lag you could always aim around independent of it and this kept things very smooth for aiming at least</u>.
<!--quoteo(post=1822304:date=Jan 7 2011, 06:08 AM:name=Uzguz)--><div class='quotetop'>QUOTE (Uzguz @ Jan 7 2011, 06:08 AM) <a href="index.php?act=findpost&pid=1822304"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I don't think it's been mentioned in this thread yet, but part of the problem with mouse responsiveness is that your aim direction replicates in both directions - that is, <b>the server periodically tells you where you're aiming.</b> ...as far as I know it's still like that now.<!--QuoteEnd--></div><!--QuoteEEnd-->
WTFFFFFFFF Are you kidding me?!
If that's true... Case closed, problem solved. This is why skulks are owning marines. This is what's causing the input-output lag.
This game will never be playable like NS1 as long as this is in place.
That may account for some of the input lag, while also the cursor in commander mode lags way behind your mouse motions. I can't imagine the server correcting your cursor location, since that would be just absurd.
<!--quoteo(post=1822396:date=Jan 7 2011, 12:33 PM:name=juice)--><div class='quotetop'>QUOTE (juice @ Jan 7 2011, 12:33 PM) <a href="index.php?act=findpost&pid=1822396"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->WTFFFFFFFF Are you kidding me?!
If that's true... Case closed, problem solved. This is why skulks are owning marines. This is what's causing the input-output lag.
This game will never be playable like NS1 as long as this is in place.<!--QuoteEnd--></div><!--QuoteEEnd--> Seriously, given the fact that this game uses a mouse and everyone expects to be able to rotate orientation near instantly if needed, we shouldn't be allowing the server to reset this on our client, the server should just take the next orientation packet (sent from the client) and be relaying that to everyone <b>except</b> the client who sent it (they already know where they should be looking). Really the only time we should be allowing the server to adjust the client orientation is when spawning or blinking or something. So yes, the client should only be <b>sending</b> the orientation that it is facing as you move throughout the world and if you are shooting, immediately append the packet with the shoot command and I guarantee it will then feel correct... (doesn't guarantee the skulk still exists in that position on the server though :\)
Isn't this the way it should have been done to begin with?
And as a further note I have a hunch that the black screen was/is being caused by bad orientation angles being sent to the client resulting in invalid matrix calculations from NaN positional angles and rendering a black screen with non-geometry still displayed. (this is an educated guess.. I had run into this issue once on my own networked 3d game engine that I wrote in a semester by myself and of course not optimized but the player themselves never felt like they were lagging; however opponents could sometimes appear slightly jerky due to no positional prediction done but self orientation angle was always left up to the client to notify the server of)
..swear I need to take a break from working on my map and dig into the LUA code to try and diagnose some issues... wouldn't be surprised if hydras (and sentries) are target scanning every tic instead of say every .5 - 1 seconds between 'scans' and not all scanning on the same tic either. (the performance difference would be HUGE unless its already doing this)
<!--quoteo(post=1822423:date=Jan 7 2011, 02:51 PM:name=Thaldarin)--><div class='quotetop'>QUOTE (Thaldarin @ Jan 7 2011, 02:51 PM) <a href="index.php?act=findpost&pid=1822423"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->The problem is right now, props don't like skulks sticking them. My map tests have shown that so far.
Dynamic lighting so far, doesn't show up in first person only third person mode. Which is a massive disadvantage for the ambusher.
Until those are fixed, don't ask for that too much ;)<!--QuoteEnd--></div><!--QuoteEEnd-->
you misunderstood, i simply meant that in comparing ns1 maps to ns2 maps, ns2 maps have much more varied lights and are much more detailed. it was way easier to pick off a skulk in ns1 because the maps were so simple and boxy with very flat and even lighting. this also made for easier skulking cuz wall running was better.
and we're supposed to be using collision geometry to make skulk wall running easier over props, no?
I tentatively agree that the main reason skulks seem more powerful in NS2 is that the maps and lighting provide far more hiding spots for them. And I don't just mean ambushing - there are so many ceiling spaces, floor grates, wall crevices etc that a skulk can literally disappear mid-fight and reappear out of another place a second later.
The other thing that is messing with skulk v marine at the moment is the big-leap exploit. I try to avoid exploits during play, but I've found I habitually use this exploit to gain a huge advantage over marines in the early game. Shame on me!
<!--quoteo(post=1822496:date=Jan 8 2011, 01:16 AM:name=NS2HD)--><div class='quotetop'>QUOTE (NS2HD @ Jan 8 2011, 01:16 AM) <a href="index.php?act=findpost&pid=1822496"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->The other thing that is messing with skulk v marine at the moment is the big-leap exploit. I try to avoid exploits during play, but I've found I habitually use this exploit to gain a huge advantage over marines in the early game. Shame on me!<!--QuoteEnd--></div><!--QuoteEEnd--> that exploit doesn't work for me..or I just do it wrong. (pressing jump and leap at the same time if I recall correctly)
<!--quoteo(post=1822518:date=Jan 7 2011, 08:30 PM:name=Thaldarin)--><div class='quotetop'>QUOTE (Thaldarin @ Jan 7 2011, 08:30 PM) <a href="index.php?act=findpost&pid=1822518"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Tig go load Tram or Rockdown up in Spark. A lot of props aren't covered in the collision meshing.<!--QuoteEnd--></div><!--QuoteEEnd-->
i have noticed, but i've also noticed that they put that stuff on the floors and in weird places as if its an afterthought.
i only loaded tram to get a good idea of how to organize my layers and how UWE went about building as far as props vs geometry and to get with the style, i didn't spend too much time in there because i didn't want to end up ripping it off, i want my level to be original.
im trying to put logical paths to the ceiling for skulks because i know how frustrating it is to have to plan out how to get from the ground to the ceiling for an ambush.
Hmm... I was going to post the issue I mentioned earlier on GetSatisfaction, but a quick test on a 400 ping server didn't reproduce it; my position would be corrected, but my camera angle would not, just as you'd hope. Perhaps it has been fixed after all. Could I get a second opinion?
ScardyBobScardyBobJoin Date: 2009-11-25Member: 69528Forum Admins, Forum Moderators, NS2 Playtester, Squad Five Blue, Reinforced - Shadow, WC 2013 - Shadow
<!--quoteo(post=1822164:date=Jan 6 2011, 06:55 PM:name=Tig)--><div class='quotetop'>QUOTE (Tig @ Jan 6 2011, 06:55 PM) <a href="index.php?act=findpost&pid=1822164"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->i really don't think its that imbalanced. the lag just heavily favors skulks.<!--QuoteEnd--></div><!--QuoteEEnd-->
Lag amplifies, rather than causes, the imbalance. I predict that skulks will still dominate marines (though to a lesser extent than now) when the lag is reduced.
<!--quoteo(post=1822611:date=Jan 8 2011, 08:22 PM:name=luns)--><div class='quotetop'>QUOTE (luns @ Jan 8 2011, 08:22 PM) <a href="index.php?act=findpost&pid=1822611"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->the problem is mostly in stability issues, and hit detection. Too much lag on both sides, and attacks not being registered correctly at all.<!--QuoteEnd--></div><!--QuoteEEnd--> Indeed. Contrary to the prediction above, I predict that once the hitreg improves, Marines will be ten-shotting straightliners en-masse. Better make the most of that starting Leap, Skulks.
Comments
This right here is the entire point of range vs. melee. This is what it's supposed to be when it's lmg vs. skulk...
that being said, somebody mentioned reticle position being sent from server and client, if they can stop that and just have the server be a slave to the client mouse movements, then you would have an easier time swinging your lmg around in close quarters to get the kill.
I think they might have fixed this but it's true. The server should only snap back on position and not orientation. Otherwise, people with good twitch reflexes are if anything confused about why quick precision aiming is so tough and this would be exactly why.
NOTE: <u>In the HL engine when there was huge lag you could always aim around independent of it and this kept things very smooth for aiming at least</u>.
...as far as I know it's still like that now.<!--QuoteEnd--></div><!--QuoteEEnd-->
WTFFFFFFFF Are you kidding me?!
If that's true...
Case closed, problem solved. This is why skulks are owning marines. This is what's causing the input-output lag.
This game will never be playable like NS1 as long as this is in place.
If that's true...
Case closed, problem solved. This is why skulks are owning marines. This is what's causing the input-output lag.
This game will never be playable like NS1 as long as this is in place.<!--QuoteEnd--></div><!--QuoteEEnd-->
Seriously, given the fact that this game uses a mouse and everyone expects to be able to rotate orientation near instantly if needed, we shouldn't be allowing the server to reset this on our client, the server should just take the next orientation packet (sent from the client) and be relaying that to everyone <b>except</b> the client who sent it (they already know where they should be looking). Really the only time we should be allowing the server to adjust the client orientation is when spawning or blinking or something. So yes, the client should only be <b>sending</b> the orientation that it is facing as you move throughout the world and if you are shooting, immediately append the packet with the shoot command and I guarantee it will then feel correct... (doesn't guarantee the skulk still exists in that position on the server though :\)
Isn't this the way it should have been done to begin with?
And as a further note I have a hunch that the black screen was/is being caused by bad orientation angles being sent to the client resulting in invalid matrix calculations from NaN positional angles and rendering a black screen with non-geometry still displayed. (this is an educated guess.. I had run into this issue once on my own networked 3d game engine that I wrote in a semester by myself and of course not optimized but the player themselves never felt like they were lagging; however opponents could sometimes appear slightly jerky due to no positional prediction done but self orientation angle was always left up to the client to notify the server of)
..swear I need to take a break from working on my map and dig into the LUA code to try and diagnose some issues... wouldn't be surprised if hydras (and sentries) are target scanning every tic instead of say every .5 - 1 seconds between 'scans' and not all scanning on the same tic either. (the performance difference would be HUGE unless its already doing this)
The problem is right now, props don't like skulks sticking them. My map tests have shown that so far.
Dynamic lighting so far, doesn't show up in first person only third person mode. Which is a massive disadvantage for the ambusher.
Until those are fixed, don't ask for that too much ;)
Dynamic lighting so far, doesn't show up in first person only third person mode. Which is a massive disadvantage for the ambusher.
Until those are fixed, don't ask for that too much ;)<!--QuoteEnd--></div><!--QuoteEEnd-->
you misunderstood, i simply meant that in comparing ns1 maps to ns2 maps, ns2 maps have much more varied lights and are much more detailed. it was way easier to pick off a skulk in ns1 because the maps were so simple and boxy with very flat and even lighting. this also made for easier skulking cuz wall running was better.
and we're supposed to be using collision geometry to make skulk wall running easier over props, no?
The other thing that is messing with skulk v marine at the moment is the big-leap exploit. I try to avoid exploits during play, but I've found I habitually use this exploit to gain a huge advantage over marines in the early game. Shame on me!
that exploit doesn't work for me..or I just do it wrong. (pressing jump and leap at the same time if I recall correctly)
or I just can't press them simultaniously.
i have noticed, but i've also noticed that they put that stuff on the floors and in weird places as if its an afterthought.
i only loaded tram to get a good idea of how to organize my layers and how UWE went about building as far as props vs geometry and to get with the style, i didn't spend too much time in there because i didn't want to end up ripping it off, i want my level to be original.
im trying to put logical paths to the ceiling for skulks because i know how frustrating it is to have to plan out how to get from the ground to the ceiling for an ambush.
Lag amplifies, rather than causes, the imbalance. I predict that skulks will still dominate marines (though to a lesser extent than now) when the lag is reduced.
Indeed. Contrary to the prediction above, I predict that once the hitreg improves, Marines will be ten-shotting straightliners en-masse. Better make the most of that starting Leap, Skulks.