Author Topic: Lerp Guide v2.3.4c  (Read 108173 times)

ProdigySim

  • *
  • Posts: 1233
    • View Profile
Re: Lerp Guide v2.3.4c
« Reply #15 on: August 17, 2012, 11:37:53 pm »
Would you say that using 0 lerp would be the same as using 33 lerp (if everyone had to use the same lerp)?
No... and I don't see the point in everyone using the same lerp.

In terms of advantages/disadvantages of lerp it's ping+lerp = total delay. So, even if everyone uses the same lerp, they all have different ping.

jdtyhdgfhgf

  • *
  • Posts: 71
    • View Profile
Re: Lerp Guide v2.3.4c
« Reply #16 on: August 18, 2012, 03:55:56 am »
So, if the delay still depends on the ping it shouldn't matter if the minimum lerp is 0 or 33 (in terms of restrictions).

The reason I'm asking this is that as you know the witches bug much more often when nb_update_frequency is not 0.1, but if we set it at default then people that use 0 lerp complain the commons seem jerky (they shouldn't use 0 lerp in the first place I know). So we got a dilemma here.

I'm thinking of starting to limit lerps based on cl_interp_ratio 1, as that's what the minimum should be anyways, and how the netcode works (correct me if I'm wrong):
tick 30: lerps limited to 33.3ms - 100ms
tick 60: lerps limited to 16.6ms - 100ms
tick 90: lerps limited to 11.1ms - 100ms
tick 100: lerps limited to 10.0ms - 100ms

Riven

  • ***
  • Posts: 161
    • View Profile
Re: Lerp Guide v2.3.4c
« Reply #17 on: August 18, 2012, 04:51:41 am »
Is this why I get 16.3/16.4 lerp on the Pwgs when my rates are as following:

cl_interp_ratio 1
cl_updaterate 100
cl_interp 0.01

Sorry for deviating the thread.

jdtyhdgfhgf

  • *
  • Posts: 71
    • View Profile
Re: Lerp Guide v2.3.4c
« Reply #18 on: August 18, 2012, 05:06:32 am »
Is this why I get 16.3/16.4 lerp on the Pwgs when my rates are as following:

cl_interp_ratio 1
cl_updaterate 100
cl_interp 0.01

Sorry for deviating the thread.

For now, pWg accepts lerps between 0 and 100.

Riven

  • ***
  • Posts: 161
    • View Profile
Re: Lerp Guide v2.3.4c
« Reply #19 on: August 18, 2012, 05:22:44 am »
I have no clue on why I get 16.3ms lerp then. Oh well, nevermind. I will force 10ms lerp via interp_ratio 0.

Thank you for your time.
« Last Edit: August 18, 2012, 05:24:18 am by Riven »

stabs

  • ****
  • Posts: 323
    • View Profile
Re: Lerp Guide v2.3.4c
« Reply #20 on: August 18, 2012, 07:15:31 am »
So, if the delay still depends on the ping it shouldn't matter if the minimum lerp is 0 or 33 (in terms of restrictions).

Why wouldnt it matter? IMO, it's exactly because delay depends on ping that broader ranges of lerp settings should be allowed, if the very minimum legal setting was 33, that is another 33 ms of delay that players would be forced to play with, locking high-ping players into higher delay than what they already have. For low pings it's not that bad, but the higher your ping, the more that extra delay matters (counter-intuitive since the ratio is smaller with higher ping, but in my experience, high ping + high lerp is a lot more of a bitch than low ping + high lerp). Please do correct me if I'm being a moron.

jdtyhdgfhgf

  • *
  • Posts: 71
    • View Profile
Re: Lerp Guide v2.3.4c
« Reply #21 on: August 18, 2012, 09:41:18 am »
Why wouldnt it matter? IMO, it's exactly because delay depends on ping that broader ranges of lerp settings should be allowed, if the very minimum legal setting was 33, that is another 33 ms of delay that players would be forced to play with, locking high-ping players into higher delay than what they already have. For low pings it's not that bad, but the higher your ping, the more that extra delay matters (counter-intuitive since the ratio is smaller with higher ping, but in my experience, high ping + high lerp is a lot more of a bitch than low ping + high lerp). Please do correct me if I'm being a moron.

Ping always mattered on online fps games, you can't just ignore that fact by letting people use incorrect network settings. How do you explain that in CS and CS:S people have to use the same lerp? And we're talking about a game that relies much more on hitreg than L4D/2.

Seems to me this problems comes from L4D. The competitive scene had no idea what lerp was, so everyone got used to use it at 0 like they're adjusting their crosshair color and now just the suggestion that it should be limited to ratio 1 (like every other valve game) is almost an insult. But I guess this is just one of the problems that will always remain, since people are too biased on their own interests to change to what's right.

CanadaRox

  • *
  • Posts: 589
    • View Profile
    • Confogl home page
Re: Lerp Guide v2.3.4c
« Reply #22 on: August 18, 2012, 02:04:09 pm »
Ping always mattered on online fps games, you can't just ignore that fact by letting people use incorrect network settings. How do you explain that in CS and CS:S people have to use the same lerp? And we're talking about a game that relies much more on hitreg than L4D/2.

Seems to me this problems comes from L4D. The competitive scene had no idea what lerp was, so everyone got used to use it at 0 like they're adjusting their crosshair color and now just the suggestion that it should be limited to ratio 1 (like every other valve game) is almost an insult. But I guess this is just one of the problems that will always remain, since people are too biased on their own interests to change to what's right.

CS1.6 uses very different lag compensation than the Source engine so anything in CS1.6 doesn't really mean anything here.  And I did a quick search in the ESEA CSS rules and they didn't have anything about forcing everyone to use the same lerp.

So here is how I explain it:
CS1.6 doesn't uses client side extrapolation do deal with lag compensation so changing the value could change how people move and your ability to shoot them so a single value is forced to keep the playing field even.
CSS doesn't enforce a specific lerp because changing the value doesn't actually change the way hits register and it doesn't directly improve your ability to get shots that would have otherwise been impossible, it just changes the time offset for your game state.

I don't get how changing the minimum ratio to 1 is "what's right".  If nb_update_whatever actually causes problems it should be removed.  People played with ratio 0 and choppy CI for a long time before anyone used nb_update_penis and if they want the advantage of easier deadstops they also have to deal with the disadvantage of jumpy CI.
"Now I've seen people running kite scripts making perfect and I do mean geometry class perfect circles around a moving tank." - The average pub

jdtyhdgfhgf

  • *
  • Posts: 71
    • View Profile
Re: Lerp Guide v2.3.4c
« Reply #23 on: August 18, 2012, 09:07:46 pm »
CS1.6 uses very different lag compensation than the Source engine so anything in CS1.6 doesn't really mean anything here.  And I did a quick search in the ESEA CSS rules and they didn't have anything about forcing everyone to use the same lerp.

So here is how I explain it:
CS1.6 doesn't uses client side extrapolation do deal with lag compensation so changing the value could change how people move and your ability to shoot them so a single value is forced to keep the playing field even.
CSS doesn't enforce a specific lerp because changing the value doesn't actually change the way hits register and it doesn't directly improve your ability to get shots that would have otherwise been impossible, it just changes the time offset for your game state.

I don't get how changing the minimum ratio to 1 is "what's right".  If nb_update_whatever actually causes problems it should be removed.  People played with ratio 0 and choppy CI for a long time before anyone used nb_update_penis and if they want the advantage of easier deadstops they also have to deal with the disadvantage of jumpy CI.

In CS:S you can't use 0 lerp, the minimum is based on cl_interp_ratio 1.
1 / 66 = 0.015151 ~ 15.1ms

And it's right, because no matter the tickrate, if your interp is following that formula above, you won't see commons stutter, while if it's lower than ratio 1 you will. (of course you have to test this with default nb_update_frequency)
« Last Edit: August 18, 2012, 09:12:42 pm by step »

CanadaRox

  • *
  • Posts: 589
    • View Profile
    • Confogl home page
Re: Lerp Guide v2.3.4c
« Reply #24 on: August 19, 2012, 12:33:39 pm »
In CS:S you can't use 0 lerp, the minimum is based on cl_interp_ratio 1.
1 / 66 = 0.015151 ~ 15.1ms

And it's right, because no matter the tickrate, if your interp is following that formula above, you won't see commons stutter, while if it's lower than ratio 1 you will. (of course you have to test this with default nb_update_frequency)

Well until PSim wrote the lerp tracker plugin using ratio <1 would cause issues.  In CSS you are shooting hitscan weapons at lag compensated targets (moving players).  In L4D2 you have to react based on non-lagcompensated events (player collisions) so being able to see a closer representation to live is much more important.  In CSS there is just no desire for a lerp that low because it doesn't really help the game in any significant way.  In L4D2 having a lower lerp can be a significant advantage so obviously people are interested in lower lerps.

And back to what I said before, if nb_update_frequency causes issues at non-default values, restore it to the default.  Then if players still want to use low interps that cause jumpy CI it is their own choice.  In Quake 3 you could have your FOV set to almost any value.  Setting FOV 160 was a setting you were free to set which would increase the amount you can see hugely, but you would have to deal with very high image distortion and middle that felt zoomed out.  Similarly if people want to use low interps that cause jumping CI, they get the advantage of easier deadstops but have to deal with jumping CI.  Their choice of low lerp has no affect on you and doesn't cause any issues with the lag compensation or hit registry on your end, so they should be free to set it as low as they want.  Similarly, their choice of high lerp DOES have an effect on you and causes unintuitive hit registration when you are the target so they are not allowed to use lerps that have been deemed "too high".

Your argument that it is right because common stutter is weak.  People should be free to deal with stuttering CI if they want since it doesn't have any impact on you.  We don't limit any other settings because setting it too high or low makes their game play subjectively worse so why should we prevent a low lerp if all it does is make their game subjectively worse?
"Now I've seen people running kite scripts making perfect and I do mean geometry class perfect circles around a moving tank." - The average pub

THyroXIN´-

  • *****
  • Posts: 519
    • View Profile
    • InnerSphere - we are not a clan we are a syndicate
Re: Lerp Guide v2.3.4c
« Reply #25 on: January 14, 2013, 07:15:32 pm »

The reason I'm asking this is that as you know the witches bug much more often when nb_update_frequency is not 0.1, but if we set it at default then people that use 0 lerp complain the commons seem jerky (they shouldn't use 0 lerp in the first place I know). So we got a dilemma here.


what is "the" witch bug?

ProdigySim

  • *
  • Posts: 1233
    • View Profile
Re: Lerp Guide v2.3.4c
« Reply #26 on: January 14, 2013, 10:59:34 pm »
what is "the" witch bug?
It's not "the witch bug" it's "the witches bug", with bug used as a verb.

He's simply referring to pathing failures--and since witches and commons share their pathing logic the same can happen to both. If the witch fails to path to a player it freaks out.

THyroXIN´-

  • *****
  • Posts: 519
    • View Profile
    • InnerSphere - we are not a clan we are a syndicate
Re: Lerp Guide v2.3.4c
« Reply #27 on: January 15, 2013, 03:43:06 am »
It's not "the witch bug" it's "the witches bug", with bug used as a verb.

He's simply referring to pathing failures--and since witches and commons share their pathing logic the same can happen to both. If the witch fails to path to a player it freaks out.

ah okay,  then I know, I also had it. So NB_update_frequenzy is not an option?!

and, how come that I read fps_max 150.
Is my setting wrong haveing fpx_max 1000 and 1000 fps?
« Last Edit: January 15, 2013, 04:36:47 am by yromem »

Visor

  • *
  • Posts: 1149
    • View Profile
    • Steam Profile
Re: Lerp Guide v2.3.4c
« Reply #28 on: January 15, 2013, 06:53:05 am »
I've made some time ago a plugin that sets the default value of nb_update_frequency whenether a Witch draws in a Survivor's line of sight. Then it restores the initial value once she's killed/juked. Shared this stuff with step, not sure if he tested it on pwg or not.

Anyway, players really don't like this feature. They'd much rather have a smooth and retarded witch instead of a laggy and smart one. Who's to blame but not Valve? By now, any source dev in the world should've noticed that ANYTHING non-default in Valve creations causes weird shit to happen. Surely their "no bosses, do what you will" policy has its benefits, but it also has a crapload of weaknesses. Modding or modifying their games is a pain in the ass and it never goes well. That's the result of convoluted code written in a random fashion by several programmers and the lack of playtesting.
$10 says you aren't anywhere near a controlling position of the L4D3 scene in any continent when it happens.

Thing is he does what he wants, cause his able to and we are letting him do that. He abuses hes place and power in this community and people like me get banned for no reason. Only thing visor wants is more and more control so he can do what ever he wants.

THyroXIN´-

  • *****
  • Posts: 519
    • View Profile
    • InnerSphere - we are not a clan we are a syndicate
Re: Lerp Guide v2.3.4c
« Reply #29 on: January 15, 2013, 04:54:52 pm »
I just checked, when i execute promod nb_update_frequency is set to 0.
I tried to avoid this by insert nb_update_frequency 0.1 in promod.cfg. but its still 0 ?
is it forced by a plugin? / how can i change it back to default. before the promod is executed its 0.1

 

A dedicated community website to competitive L4D and L4D2, ran by the community, for the community. L4DNation supports all continents of play and focuses on bringing together the community as a whole to a central hub of information.