[tribes education volume 2] netset

unless he is lying it looks to me that lower pingers should be allowed to use netset because 1.40 provides little assistance to them, and interpolate 0 will give them a more accurate drawing (or simply using netset to now supply themselves with prediction which would be a superior choice)

that also means that higher pingers r now at a distinct disadvantage again since they deal with more warping and trigger lag

ban net set please :) hehe
 
A proper control demo would compare server player data to client player data. Not what Marv3lo posted. Just saying.

That would prove what's more 'accurate'.
 
i actually thought of that myself as well since he is just taking one set client data and applying different network settings to the same demo repeatedly

that said i don't think it is possible for interpolate = 0 to draw the image too far forward, so if setting interpolate to 0 on with 1.40's stock pft value draws the image forward, then that image must be more accurate unless i am mistaken?
 
KnownPosition->PredictedPosition
KnownPosition->InterpolatedPosition->PredictedPosition

It's your call. I'd rather have the latter, with no jitter.
 
(and this is only good news and not bad, i am glad to be wrong [and it looks like i am] - i still don't think it's a cheat and i have always been for a level playing field)

(i still most likely wont use it as i ping very high and i find that the prediction is at a very good setting for myself)

i agree nofix in that i'd rather have the model interpolated and then predicted

but it appears that in stock 1.40 andrew provides little to no prediction for people with low pings, and interpolate 0 on top of stock 1.40 is superior to stock 1.40 by itself. that said, if they really understand whats going on and they use netset to improve or tweak their settings people should really just stop toying with interpolate entirely and start to understand the utility of pft (unless you ping really high, like 180+)
 
Last edited:
thx 4 discussion guys i tried 2 follow along as best as i could and i think i get most of it

i am gonna continue to test 1.40 with netset

polate at 0
pft at 16

thank u i learn so much!
 
opsayo said:
i am assuming when he typed 1.30 he either used a 1.30 bypass or meant 1.11 as in the other cases

@nofix: i believe he is setting them equal because he found the following to be true (not because he assumed they are inverses of each other):
Click to expand...

Correct on both counts. I used '1.11' and '1.3' a bit too interchangably.
 
wat i dont get is y andrew made 1.40 to "simulate" 32 ping and not 0 ping 4 every1 y? (does he hate ppl with 16 ping????)

also can we get wdeath back
 
dont ask nofix 4 netset, dont ask me 4 netset, it is around the community already many people have it here are the insturctions if u can't understand this google "command line"

after i config'd i put my shit in a .cs file


1. Extract to Tribes
2. Tribes.exe +p netset

NetSet Variables:
$net::interpolateTime,
$net::predictForwardTime, and $net::predictForwardTimeMethod.

Both Interpolate and PFT use their default values unless set otherwise.

Example:
$net::interpolateTime = 64; // Default 1.11, 1.30 and 1.40
$net::predictForwardTimeMethod = 0; // Default 1.40 PFT
implementation. Using this method, the PFT setting does nothing.
$net::predictForwardTimeMethod = 1; // PFT setting is used now, rather
than the 1.40 implementation.
$net::predictForwardTimeMethod = 2; // PFT is set similar to 1.40's
algo, but in increments of 32. An experimental setting meant to reduce
jitter.
Click to expand...
 
marv3lo said:
As above, nothing critical.

And here I was thinking it was about the accuracy of player positioning. How stupid of me.
Click to expand...

Accuracy would be determined with server/client data. Not demo, and client variable altering. You can't distinguish accuracy using client data alone.

Preference, as in do you enjoy jitter, with more accurate positions? Or do you enjoy smoothness, and slightly less, or maybe slightly more accurate positions?
 
opsayo said:
marvelo i am curious

do u think it is possible for an image to be drawn too far forward to the point of inaccuracy
Click to expand...

Well yes I do, certainly.

Nofix: alright, I guess we're at an impasse of sorts. Perhaps you could let me/others know how to test this in layman's terms? Would it be as simple as hosting a non-dec server and using that recording?
 
Last edited:
Back
Top