ETPro HitBoxes

[PiNoY]

well done Sun. seemed to me that this is a successful tests. it was fun reading those tests indeed meeting fps, nudge and ping rate playing a role on hit boxes on top of aim accuracy.

 

re-quoting that last funny statements

 

if you get beaten by a high pinger, then you suck.

 

 

lmao

[SunLight]

You know what's the funny (or maybe sad) thing about it?

 

Today and yesterday I played on an etpro public 2 times. In these 2 sessions I have been called unhittable both in the 1st and the second. 'get a head' 'get inet' etc. etc.

 

Not only I am bad, but whenever I manage to kill someone I get accused of being 'unhit', that's more than my patience can handle.

Why I am so mad?

First of all, because from those screenshots you can see there are no unhittable players (at least on etpro) .

And also because I think that insulting enemies when you lose is the worst aspect of this game, even worse than cheaters. Whoever plays in etpro servers knows how much they cry about 'unhittable players' when they lose.

 

---

Another funny anecdote:

In 2010 I happened to be in Canada for a short amount of time (maybe someone remembers me with a Canadian flag in hardcore server), so I had a nice and low ping in FA servers.

I was playing in one of FA jaymod servers, and a player (who is better than me) connects. (I won't say his name, but it starts with 'a' and ends with 'y' )

I don't know if I had a good day, or it was because of low ping (he connected from Europe), but I killed him a lot of times, and he started to insult me in all possible ways ('low jaymod player' etc. etc.)

At some point he probably looked at the scoreboard, an since I had a low ping he wrote:

'aah, you are American... all Americans are unhittable'

 

lol? So I'm unhittable when I have a high ping, I am unhittable with low, I must have some magic unhittability script in my config without knowing lol...

[Moita67]

After this work,with respect i only can say one thing!

 

I Love you man

 

Regards

 

p.s. I´m not gay and i can prove it

[PiNoY]

So, wanna test this in production environment like a VPS? Maybe for several days, we could create one or two USA-based linux VPS and see? Minimum RAM we can create is 512, but we can scale that up to 64GB RAM swiftly anytime, only a reboot is needed to do that. Just come up with your already existing plan of procedure analysis then we can be good to go anytime.

 

Anyway, this is quite nice findings you are sharing, it is not quite easy to create a simulated environment with 2 hands. Even getting SS while one player is jumping, one player fires a gun and another one captures the scenario, that just to have handful screenshots with convincing results.

 

well done Sunlight.

[SunLight]

So, wanna test this in production environment like a VPS? Maybe for several days, we could create one or two USA-based linux VPS and see?

You mean making a test server where players can connect and test hitboxes? hmmm I see several problems...

 

1. That fake player is an entity drawn in real time by the client only, so other players in the same server wouldn't see it, and it doesn't go in demos, etc. etc. It's more or less like with that etpro 'trickjump ghost' feature. Everyone would see his own 'fakeplayer' only.

 

2. the client number of the enemy (whose entity had to be copied) was hardcoded in that test, so I'd have to find a way to select it, like using a cvar.

 

3. To do the tests, I modified the client on the fly in ram. I could even make a real modified client, and put it into a pk3. But that would pose some extra problems as well, and of course the mod name shouldn't be 'etpro' or it would overwrite stuff in real etpro folder... it looks a bit complicated.

 

Anyway, I don't know what you think... I think I tested more or less everything that was possible to test. If you want to test it while you fight (and not just hitting an enemy moving left or right) your screen would be filled with debug lines everywhere.

 

Actually, an easy way to prove that hitboxes and antilag work well, and if you aim on the player you hit, would be to make a server without randomness in bullets (and maybe without bullet knockback). It has already been done some time ago.

Edited February 23, 2013 by SunLight

[SunLight]

I changed in the cgame file and not in ram anymore, now I just have to load the game from that etpro installation to test hitboxes, but on a test server it would be too messy during fights (I also raised the number of frozen players to 16, and they fade out with cg_railtrailtime like bullets)

 

Look at this screenshot (it's Bremen, but I did r_drawworld 0, to see only the entities)

 

lol way too spammy:

 

if I make a cvar for the client number I could even post the modified cgame_mp_x86.dll that draws these players, but I think it would be more of interest on a site like crossfire where most play etpro (but they are also a bunch of trolls, so I imagine already what they would say , on the lines of 'et is already dead' 'you did this years too late' or 'players from country X are still unhittable', not to mention 'who the f. are you' )

[Quovadis]

if I make a cvar for the client number I could even post the modified cgame_mp_x86.dll that draws these players, but I think it would be more of interest on a site like crossfire where most play etpro (but they are also a bunch of trolls, so I imagine already what they would say , on the lines of 'et is already dead' 'you did this years too late' or 'players from country X are still unhittable', not to mention 'who the f. are you' )

 

lol you do read crossfire

[Moita67]

Hi guys

 

PiNoY are right,this is an amazing work,and sharing these here, only from a good person who love

Wolfenstein enemy territory!

 

Guys,if you need a "guinea pig",you can count on me,ok?

 

I´m not a "big brother" but i´m watching!

 

Regards

 

p.s. In last post,i forgot put at the end of p.s. !

[S3ti]

Googled some silent mod stuff and somehow landed here Nice read Sunlight.

 

Although I know that in theory that all works very well and netsettings make no huge difference, I have to say that hits SOMETIMES don't register when they should with crosshair spot on, no spread, not getting hit yourself and that stuff...for whatever reason. For me it happens the most when someone is crouching and moving on some bumpy terrain. Also the faster someone is going you have to aim rather behind the model you see on your screen to get decent hits, maybe the hitbox trailing the model or your client predicting the model too far ahead.

Edited June 3, 2013 by S3ti

[SunLight]

Googled some silent mod stuff and somehow landed here Nice read Sunlight.

 

Although I know that in theory that all works very well and netsettings make no huge difference, I have to say that hits SOMETIMES don't register when they should with crosshair spot on, no spread, not getting hit yourself and that stuff...for whatever reason. For me it happens the most when someone is crouching and moving on some bumpy terrain. Also the faster someone is going you have to aim rather behind the model you see on your screen to get decent hits, maybe the hitbox trailing the model or your client predicting the model too far ahead.

If you read the pdf, you can see that not only hitboxes, but also your bullets don't match perfectly with your viewline. And not only the endpoint, which is something everyone knows (spread) but also the origin.

 

And spread is *never* 0 in ET. There is a minimum value forced in the code. There is no way for you to shoot a series of bullets all in the same place. Even if you crouch.

 

So even if your aim (viewline) was perfect, it's not necessarily so for your bullet line (which you can't see, and if you could it would be too late anyway)

 

 

Hitboxes don't 'trail' behind, but the server has to do different calculations than the client interpolating between frames, so that's probably the reason why hitboxes aren't spot on, only close.

Then you have monitor input lag an other stuff, bullet lines 'trailing'...

Every player with enough experience tries to figure out how to compensate for that kind of things, maybe subconsciously.

 

Unluckily demos don't record everything, so it's hard for a player to know the reason why a bullet didn't hit the target. You miss the bullet origin, the position of the enemy when you saw him, etc.

It's possible to test only in special ways, for example with a modified client and a test server, as I did here.

 

Anyway, the most important thing I was interested in (when I made these tests) is whether a player can have an unfair advantage over others or not, and as you can see nobody can move his hitboxes (after all it's obvious, since hit detection is server side, and the server is the one who sends you the snaps with players positions as well)

[S3ti]

If you read the pdf, you can see that not only hitboxes, but also your bullets don't match perfectly with your viewline. And not only the endpoint, which is something everyone knows (spread) but also the origin.

 

And spread is *never* 0 in ET. There is a minimum value forced in the code. There is no way for you to shoot a series of bullets all in the same place. Even if you crouch.

 

So even if your aim (viewline) was perfect, it's not necessarily so for your bullet line (which you can't see, and if you could it would be too late anyway)

 

 

Hitboxes don't 'trail' behind, but the server has to do different calculations than the client interpolating between frames, so that's probably the reason why hitboxes aren't spot on, only close.

Then you have monitor input lag an other stuff, bullet lines 'trailing'...

Every player with enough experience tries to figure out how to compensate for that kind of things, maybe subconsciously.

 

Unluckily demos don't record everything, so it's hard for a player to know the reason why a bullet didn't hit the target. You miss the bullet origin, the position of the enemy when you saw him, etc.

It's possible to test only in special ways, for example with a modified client and a test server, as I did here.

 

Anyway, the most important thing I was interested in (when I made these tests) is whether a player can have an unfair advantage over others or not, and as you can see nobody can move his hitboxes (after all it's obvious, since hit detection is server side, and the server is the one who sends you the snaps with players positions as well)

 

Sure spread is never 0, I meant the lowest spread you can have in those situations. I have often been called unhitable myself on etpro, and even my mates (which I trust to tell the truth) said it when we played each other. Agree with you that one can't influence the own hitbox by settings, but those real 'unhitable' situations do exist after all, whatever the reason might be.

[SunLight]

Sure spread is never 0, I meant the lowest spread you can have in those situations. I have often been called unhitable myself on etpro, and even my mates (which I trust to tell the truth) said it when we played each other. Agree with you that one can't influence the own hitbox by settings, but those real 'unhitable' situations do exist after all, whatever the reason might be.

 

I think one of the reasons why some players seem to be less 'hittable', is that everyone moves in a different way. If someone has a movement that doesn't match well with yours, you  will have a harder time hitting him.

And of course players with a high accuracy are harder to hit, because they give you more knockback.

 

Also, we shouldn't make confusion between warping and being unhittable. Some players warp, especially in some mods, so it's harder to *aim* at them (so they should be called 'untrackable' rather than 'unhittable').

But if you aim at a player and you don't *hit*, then you wouldn't have hit anyone who was in the same position.

 

 

Since it's impossible to repeat the same fight twice, and it's impossible to watch again what happened, players can only trust their judgment (which is often wrong ).

Knockback and bullet randomness are responsible for most misses, I don't know if you remember on Crossfire when they talked about that Etpro 'promode' without bullet randomness and knockback, and how everyone said every player was hittable there...

[rajaah thunder]

nice findings sunlight ! I am amazed  Well , now i understand why even i frag sometimes with my ping on my lucky days . So , fps does influence bullet spread doesn't it ? Then , what is the lowest possible fps for good shooting ? And more tips could help a noob like me  

[S3ti]

I think one of the reasons why some players seem to be less 'hittable', is that everyone moves in a different way. If someone has a movement that doesn't match well with yours, you  will have a harder time hitting him.

And of course players with a high accuracy are harder to hit, because they give you more knockback.

 

Also, we shouldn't make confusion between warping and being unhittable. Some players warp, especially in some mods, so it's harder to *aim* at them (so they should be called 'untrackable' rather than 'unhittable').

But if you aim at a player and you don't *hit*, then you wouldn't have hit anyone who was in the same position.

 

Well, I don't say that specific players are more unhitable than others, but there are situations where hits clearly don't register when they obviously should based on what is on your screen. I'm aware of the variables like knockback, dodge patterns, higher spread and all those stuff in some places or situtations. Or did you never experience that? Easy shots, standing on even ground, not getting hit yourself...someone is crouching and strafing and you just don't hit. Can upload a demo when it happens and when it's obvious

[PiNoY]

I'm one of those players that almost everyday experience that kind of "delayed" or "less hittable" shots to the enemy.

 

Aimed approx on heads, but I will hear the hit maybe on 4th or 5th shots even though I know I hit him on first bullet. There's always delay. I know this as most times enemy find it hard to have steady hit on where I assumed that I hit him first and hard, so his movement and his aim were affected.

 

At times, if I killed the guy, screen will show me I killed player X .After a few seconds, another screen will show me I killed same player X again (cl_dup packets probably). This is not technically possible killing an already dead guy. But this is what is happening. The 2nd screen can appear while he is still waiting to be served by medics or he's probably spawning. Strange but this is normal on my experience since.  What confuses me is that I can hear footsteps on almost on due time, no noticeable long delays not like what was described above.

 

The same goes when I was killed in front of door entrance, but I can still move and I die inside the room already. My death is delayed. But  I know I was shot outside even though there was no line of sight inside, still that same player killed me. So, at my ping of 150, to be honest , I wonder how J3 regular players are being owned  by players with 200 or 300 ping . At 200, it is really unplayable for me unless I played naab,  use panza and flame thrower.

 

The above is when I have nudge enabled.

-------------

 

Now, having a disabled nudge, I hit him at 90 X and Y axis, I can see that I am not hitting him well. I assumed enemy was on 91 X/Y axis already. So, adapting to situation, I quickly scrambled my mouse or flicked it a bit in  advanced location where enemy is going to be, or supposed to be (assumingly). This is not easy for me but at times it works, at times it doesnt. If I adjust my mouse sensitivity, my assumption goes to enemies I am not able to hit due to wide spread caused by my high sensitivity. This is my theory on that. Quite a big spread, enemy ping, my ping, and where I am truly aiming at.

 

---

 

Having the above, tendency is to be more paranoid when I'm easily dead at 2 sec close fight with a player having a higher ping than mine. I can easily accept without doubts that a low pinger player can easily frag me, but a higher pinger player that I am not able to frag even once in 3 maps is really a  suspicious for point of "view", and my laggy link.

 

---

 

Whatever the case, I tend to have fun rather than rage quit. Sry long post.