cl_nodelta

[ouroboro]

Or they are just misguided.

Or that. I guess I wouldn't enforce it with PB in case someone (like me) changed it for some misguided reason, but using ETPro to change it to 0 seems fair enough.

[brugal]

alot of bandwith with 'cl_nodelta 1'

on a local etmain server with only 2 clients:

cl_nodelta 0
to client: 9413 packets, 715973 bytes, time:476.867795, 19.739224 packets per second, 1501.407743 bytes per second, 76.000000 bytes per packet
to server: 39175 packets, 1375294 bytes, time:476.871784, 82.149964 packets per second, 2883.991140 bytes per second, 35.000000 bytes per packet

cl_nodelta 1
to client: 14801 packets, 9360459 bytes, time:748.698547, 19.768971 packets per second, 12502.306884 bytes per second, 632.000000 bytes per packet
to server: 61607 packets, 2116651 bytes, time:748.731479, 82.281835 packets per second, 2826.982783 bytes per second, 34.000000 bytes per packet

[LilleBror]

ReyalP:=> Im not going 2 bust ur ballz... but u can get rid og lagg on demand... its easy..

u can code it out "antilagg style" or "freeze" clients that dont deliver packets...

"antilagg... easy.. "

1)
"nonlagging client can hit all positions that are -/+ X ms from last postion..plus intermediate states... ei. relax the hitsacan vs laggers...."

2)
client lagg X% msec... do a area increase of the SIZE of the hitboxes multiplied by the X% miliseconds off lagg... so la client missing a SNAPS gets 50 msek = 1.5 mutiplier = x*1.5 by y*1.5 * z *1.5 increase.....

bottom line... i.e u are a bullit magnet if u dont diliver packets...

"freeze"..

1) if u dont diliver ....u dont move and u dont hit...

2) missing X frames SNAPS or packet dilivery u go 2 go spectator...

3) "time" punishment: "AT ur last ON time dilivered POS" the server will not procces ur action commands untill Z number of seconds have passed... u will still get packets(damage and so forth)...

Think FAIR... and Try 2 make the game work 4 the players that dont expliot the antilagg code....and diliver there packets ON time ALL the time..

soory i cant keep the peace i promised... ur statement the lagg on demand will allways be possible is FUBAR...!!!! jesus on ICE...


deej:=> As i have told u before u aim is 45+% but ur reaction speed is not pro.... live with it.. if u have a problem with me... so be IT--->
I dont mind getting my ass handed 2 me by some 12 year old that is a god on the mouse... somehow i belive u do...
I got my ass handed 2 me by playing 1v1 on the OLD boomtown 1.. shooting 60% acc...thats the real game....
u got nailed 2 the wall when u played there.... grow up!...u had reaction speed of about 200 ms.... Im around about 100.. and some are at 50 or lower,,,, the best i played was at 35... and he was evil...

all:=Z> cvars should be set optimal 4 the server... and if cl_nodelta is FUBAR.... PB the "god damb cvar" so it cant be altered during ONLINE FIRST PERSON SHOOTER... game....

off topic: i dont buy ONLINE games anymore.. if i cant "hit" whats in the crosshair im not going 2 spend the money...

if ur whant a fair game STOP buying games that allow lagg cheating!

Its my policy...


Until the Shoftware is "Cassio" Style fair.. Im not going 2 spend one dollar...

real world:=> Im going 2 get banned 4 this post... so check up! and i hope that a lots of players will read the post before its removed....

over and out
LilleBror....

p.s Bani i love ur mod please help me out u have my mail.....

[ReyalP]

u can code it out "antilagg style" or "freeze" clients that dont deliver packets...

In ET (and just about every other Q3 engine game), you already freeze if your commands aren't received. In ETPro (with antiwarp on), you are also prevented from making big warps when the commands start flowing again. If you have demos that show otherwise, please show them. Note that if your commands arrive in bursts, you will still look jerky on other peoples screens, because you are frozen some of the time, and moving some of the time.

Anyway, lag on demand is still possible. Your suggestions just punish the lagger, they don't stop the lag. From what I can tell, in the current etpro, lag on demand doesn't give you much of an advantage. Again if you can prove otherwise please do. Longwinded, hard to read posts full of creative spelling aren't proof. Demos or reproducible test cases are.

"antilagg... easy.. "

antilag (with one g) is a specific technique for dealing with hitscan weapons. It has nothing to do with player movment.

real world:=> Im going 2 get banned 4 this post... so check up! and i hope that a lots of players will read the post before its removed....

Not this one.

For those who might be confused, my point was that if cl_nodelta can be used to induce lag, the same effect can be achieved with your favorite bittorrent client, rate limiting software or a million other ways, so there really isn't too much point in restricting the cvar. OTOH, if it makes you feel good, there isn't any harm in restricting it, since 0 is the only reasonable value for normal play.

[ouroboro]

I think half the reason LilleBror is so intolerable is the creative spelling you mentioned. I realize he's from Denmark and English is a second language and I congratulate him for that. But the superfluous arrows, numbers instead of letters, intentional misspellings and completely random capitalizations and ellipses all add up to make me dread his posts.

It's a shame because there might actually be some coherent logic in his posts, but it's too hard to sort through all the above to find it.

P.S. =FF=im2good4u gets away with it because his posts are somehow cute and lovable.

Edit: Here is a b_antiwarp test I did back on 3.2.1. I forgot about the affects of testing on LAN, so the results might be a little better than online. But they're still impressive. Even at 90% packet loss, b_antiwarp does a fine job, IMO. Read the chat to follow what's going on. And disregard the horrible hill jumping, I considered the normal running back up the hill to be the main focus of movement.

[LilleBror]

ReyalP:=> You are right, u cant "force" clients 2 dilver and recive packets in a manner that is server optimal, with regards 2 freqency, timedelta and ping. But it is possible 2 write code that gives the advantage 2 the players that have near perfect server-client information exchange.

Im not restarting the topic.. but i asked 4 a server cvar option... that would "give" the advantage 2 the players that have good connections and systems.

It dosent matter 2 me how the "problem" is resloved, but one thing is 100% sure, if the game is unplayable 4 clients that "lagggggg" they will stop playing or fix the whatever issues that prevent them from having a smooth transfer of data between their client and the server that they desire 2 play on....

ouroboro:=>sooory 4 the spelling.. its "brainstorm style"... so i punch ... when i think..
> 4 (its 2 u ..)... CAPITAL LETTERS 2 "describe" raising my voice.. in a verbal conversation --- <...line or tread... meaning can u follow my pattern of thinking.... but ur right is looks like spam style...

[ReyalP]

ReyalP:=> You are right, u cant "force" clients 2 dilver and recive packets in a manner that is server optimal, with regards 2 freqency, timedelta and ping. But it is possible 2 write code that gives the advantage 2 the players that have near perfect server-client information exchange.

ETPro already gives the advantage to people with good connections. There may be specific situations where a lagger gets some advantage, but on the whole, a good connection is better. If you can show specific, reproducible cases where lagging or a bad connection gives the lagger an advantage, please do so.

It dosent matter 2 me how the "problem" is resloved

As long as you don't demonstrate the existence of the claimed problem it certainly isn't going to be resolved. What we are NOT going to do (I don't speak for the other etpro coders, but I'm pretty sure they agree with me on this) is arbitrarily punish people with bad connections.

ouroboro:=>sooory 4 the spelling.. its "brainstorm style"... so i punch ... when i think..
> 4 (its 2 u ..)... CAPITAL LETTERS 2 "describe" raising my voice.. in a verbal conversation --- <...line or tread... meaning can u follow my pattern of thinking.... but ur right is looks like spam style...

It is also very hard to read, and doesn't encourage people to look seriously at what you are saying. It may make your line of thought clearer to you, but it tends to make many of us just skip to the next post. For me at least, the same goes for abbreviations like 2 4 ur etc.

[LilleBror]

ReyalP:=>
Setting up experiments to prove specific and reproducible cases is giong to be very hard.

Ill give it go. If you are willing to help me with input and information on the experimental setup, you are more than welcome.

Existence of the claimed problem, i think thats its self evident, clients have far to low accuracy compaired to the spead off the weapons in the game.

The demo of this years nations cup battle UK vs Germany (Semifinals) UK marvel, makes a run with the OB as Allies on goldrush with 3 german players targeting him, can u please justify how he can make that run if there are no critical problems of lag expliotation possible?

[fretn]

The demo of this years nations cup battle UK vs Germany (Semifinals) UK marvel, makes a run with the OB as Allies on goldrush with 3 german players targeting him, can u please justify how he can make that run if there are no critical problems of lag expliotation possible?

spread ?

[ReyalP]

ReyalP:=>
Existence of the claimed problem, i think thats its self evident, clients have far to low accuracy compaired to the spead off the weapons in the game.

I don't see it. For example, if you watch the quacecon demos, those players continue to shoot about the same as they did in online matches. That represents some of the best players in the world, on high end machines and a LAN. Unless you claim that they were all abusing some network setting, that should pretty much be better than any internet server will ever be.

If I have understood correctly, your essential claim is that laggers are hard to hit, to the extentent someone who got say 30% acc on your average pub could get 50% if everyone had high end machines and perfect connections. As I already said, quakecon and other LANs seem to put a limit on how big that effect could be, well below what you have claimed. Furthermore, an aimbot can get something like 70+ acc with an SMG and 90ish with a sten, on a normal internet connection. So obviously the human factor is the main limitation that keeps normal human players in the 20%-50% range.

It is difficult to set up conclusive experiments to test hit detection. Perhaps the clearest would be to use aimbots (same on both computers) and /give health a very high number. Obviously, this would be done on a private, PB off server. Decide on a set of tests, like tracking an opponent going in a straight line, one is strafing back and forth, face to face with both players strafing etc. Do it enough times find out the normal variation in results. Now vary network settings, or introduce packetloss, latency etc (there are fakelag settings in ET, or better use something like a linux based router that). and see how it affects the results.

[RoadKillPuppy]

> No sir, I am not using a bot, I am just debugging this àç!* game

But let's stress this again: do NOT test like this on a pb on server, and better use seperate pc's/configs/installs/... You don't want to end up on the MLB or pb hardware ban list. ( + you can forget about comp. play)

[LilleBror]

Lag on demand.

Spread of the weapons in ET is particaly zero.

Aimbots that dont reach 99-100% acc is a indicator that something is wrong.

Without considering external programs thats can infuence client server communication.

Variable range = r

Static condition cvars = a

s1 = rate s1 = [s1]x[r1]
s2 = cl_timenudge s2 = [s2]x[r2]
s3 = pmove_msec s3 = [s3]x[r3]

Float conditions cvars = f

f1 = cl_nodelta f1 = [f1]x[r4]
f2 = cl_avidemo f2 = [f2]x[r5]
f3 = snaps f3 = [f3]x[r6]
f4 = cl_packetdup f4 = [f4]x[r7]
f5 = com_maxfps f5 = [f5]x[r8]
f6 = pmove_fixed f6 = [f6]x[r9]
f7 = cl_maxpackets f7 = [f7]x[r10]
f8 = r_swapinterval f8 = [f8]x[r11]
f9 = r_displayrefesh f9 = [f9]x[r12]

Commands = c

net_restart c1
in_restart c2
ui_restart c3
shrug c4

step 1) [startpoint.script] [r1,...,r12]
step 2) Run a [lag.script.start] = at [ r1,r2,r3] while [r4,...r12,c1,..,c4] are changed/run.
step 3) Repeat [x] times, or run a differnt combination ie [lag.script.next]
step 4) Repeat step 3 [y] times
step 5) Run the [endpoint.script] setting [r1,..,r12] = [startpoint] [r1,...,r12] without running c
step 6) Bind the keys

My thesis is that there is a combination, that can induce lag on demand.

Most likely its a simple combination.

Bell curve Acc.

Consider a skillset of players.
Newbie 40 N
Skilled 45 S
Elite 50 E
Pro 55 P

Variation of +/-5 % bell curve distribution.
Movement, evasion, forced crosshair movement taken into consideration.

1 vs 1 meanAcc% sumtotalAcc% meanAcc%range sumtotalAcc%range

N vs N 40 40 80 35-45 35-45 70-90
N vs S 35 50 85 30-40 45-55 75-95
N vs E 30 60 90 25-35 55-65 80-100
N vs P 25 70 95 20-30 65-75 85-105

S vs S 45 45 90 40-50 40-50 80-100
S vs E 40 55 95 35-45 50-60 85-105
S vs P 35 65 100 30-40 60-70 90-110

E vs E 50 50 100 45-55 45-55 90-110
E vs P 45 60 105 40-50 55-60 95-115

P vs P 55 55 110 50-60 50-60 100-120

Weapon influence
Distance shift 0.9 - 1.3 mutiplied factor from extream range to toe2toe Sten
Distance shift 0.8 - 1.3 mutiplied factor from extream range to toe2toe Mp40/Thomson
Distance shift 0.6 - 1.3 mutiplied factor from extream range to toe2toe Colt(s)/Luger(s)
Distance shift 0.4 - 1.2 mutiplied factor from extream range to toe2toe FG/Garand/K43
-weapon spread related
-target size related
-unscoped FG/Garand/K43

Tactics
(Use of Cover)�fire displace cover� �popup� tactics. 0.6 – 1.0 multiplied factor.
-reaction speed related. N 0.6-07 S 0.7-0.8 E 0.8-0.9 P 0.9-1.0

(Placement/Position)�flank/back� attacking 1.1 – maxout 100% total acc X number of rounds before fire is returned, and yes everyone can shoot well if they have time 2 zero in and open up.
-X depends on the situation awearness of the target and reaction speed.

Situation
Stationary target �flank/back�. Maxout--- all shoot hit if the are no weapon influence on the range part.

Basic unless a 1 vs 1 firefight is at extream range and both are using cover, its going to be fast and at high acc.

Gibbing... yes that can lower acc...

But when we are talking 20acc... its allmost ridculas.. since u think 20-50 is humane factor..
That would mean that ur �getting� owened at 70%+ acc if ur at 20%

There is a gap of missed shots that not true to the weapons in the game.

Alien cvars.

There are non native cvars that work in ET and Etpro.

cl_125hz
cg_physics
b_flickerghost
nudge
...
ect..

plus a ton of cvars that may work..

nudge <= should be clamped with a timedelay like rate and cl_timenudge is in Etpro.


Documentation.

But the real problem is that there is no documentation on all cvars in Et/ETpro that can be set.

Even if documentation is to much work, at least post all the cvars that can be set in ET/ETpro.

I forgot about it in the lag on demand section of this post but there should be some serious implications if its dynamicaly altered.

Plus i would like if somebody in the Etpro team would write a optimal128+DSL.cfg

And also tips on netsetup, mousepolling rate, PCI latency..ect

IMHO if u download and install the game ur dead meat...

Experiments.

Using Aimbots would not be my cup of tea, first i dont have one, and im never going 2 download one either.

I think the best way is to setup a server, and let it run diffent server.cfg's, and collect all the statistics of the palyers. If my thesis is right average acc is going to go up, with more restictive server.cfg's with regards to gamemehanics/connection cvars.

The demo.

The demo of this years nations cup battle UK vs Germany (Semifinals) UK marvel, makes a run with the OB as Allies on goldrush with 3 german players targeting him, can u please justify how he can make that run if there are no critical problems of lag expliotation possible?

Can u please comment on the specific part of the demo i have mentioned.

[ReyalP]

Spread of the weapons in ET is particaly zero.

http://www.sonic.net/~rfm/et/etss/et-has-no-spread.jpg
SMG, single shots on the left, full clip on the right. Or
http://www.sonic.net/~rfm/et/etss/garand-vs-mp40.jpg
garand slow fire vs mp40 single shots. Or
http://www.sonic.net/~rfm/et/etss/garand-vs-pistol.jpg

Note that none of those shots involved turning or jumping, which both add to your spread.

Aimbots that dont reach 99-100% acc is a indicator that something is wrong.

Incorrect, even disregarding spread. There are many things that could cause an aimbot to be imperfect. If I would going to do this test, i'd code the bot directly in the et client, but even that would have some imperfection. In any case, the fact that the difference between the best players and aimbots is 40+% proves that the human factor is the dominant factor. If you dispute this, please present a clear, rational arguement.

Without considering external programs thats can infuence client server communication.
... incomprehensible verbiage snipped...

None of that adds anything factual to this discussion.

Alien cvars.

There are non native cvars that work in ET and Etpro.

cl_125hz
cg_physics
b_flickerghost
nudge
...
ect..
plus a ton of cvars that may work..

Incorrect. There is no such thing as 'non-native cvars'. A cvar either has code associated with it, or not. None of the ones you mention here are used by the code in any way.

nudge <= should be clamped with a timedelay like rate and cl_timenudge is in Etpro.

See above.

Experiments.
Using Aimbots would not be my cup of tea, first i dont have one, and im never going 2 download one either.

I mere suggested that as one way of removing the human factor, which, as I have already proven, is a major source of uncertainty.

I think the best way is to setup a server, and let it run diffent server.cfg's, and collect all the statistics of the palyers. If my thesis is right average acc is going to go up, with more restictive server.cfg's with regards to gamemehanics/connection cvars.

Unless the effects are very strong, this will be swamped by random variations. Also, the test would need to be blind, or players knowing the settings will influence the results.

Can u please comment on the specific part of the demo i have mentioned.

Not if you don't post a link to the demo, and the specific round and time (on the match clock) you are talking about.

If you want to continue this discussion, please make an effort not to confuse facts with speculation.

[ouroboro]

Oh.

My.

God.

[mortis]

Lille Bror = ownt by reyalp