Board index DeFRaG How to DeFRaG

More exploits!

Users helping users.

More exploits!

Postby amt-morbus » Jun 15th, '11, 8:36 pm

Hello everyone!

In this thread I want to post some techniques and demos that are not widely known.
Few people know about them but I guess you would also like to know.
There are no proper names for them - so I'll just call them as you see below.
Hope you get inspired and can use this content somehow for new routes/maps etc:

1) Sticky nades
2) Ceiling movement
3) W3sp-think17 bounce


Let's start with an easy one:

1. Sticky nades:
Sticky nades are nades that get stuck in the ceiling. All that you need is some +z speed and
a greade-launcher. Right before you hit the ceiling above you, fire a nade straight up.
The nade will get stuck with proper timing (or disappear in the volume of the brush :D ).

Demo:
AMT-Freestyle1_sticky_nade_AMT-Morbus.dm_68


2. Ceiling movement:
It seems that this one has been already discovered years back but just lately I got introduced
to it by Space. You need a ceiling and a gun with a good fire-rate (like plasma, bfg, lg) and some knockback. What you do is to press crouch right when you hit the ceiling and start firing like you would do a plasma climb on a wall. Adjust the angle to your needs (28-32° is good for plasma). Plasma-only is too weak, you need quad-damage. Bfg is strong enough - you won't need quad-damage or haste. Check the demos:

Demo:
AMT-Freestyle1_Ceiling_plasma_AMT-Morbus.dm_68

AMT-Freestyle1_Ceiling_bfg_AMT-Morbus.dm_68

Btw: You don't need to hold crouch while you are stuck in the ceiling. Ok, happy traveling!


3. W3sp-think17:

I did not see any demos on this map so far - so here you go. This is how the map is done:
Like the mapname says, there is a "layer of distractions". All this timer-stuff and bfg part is not part of the actual way to stop the timer. You need to get stuck in the maps brushwork (rebounce) and bounce to the finish-line. You will slide down to the floor and bounce - no matter what height you start.
After I downloaded the map for the first time, I tried that but failed. In order to get stuck between the two prism, you gotta do this: Adjust your yaw-angle to 87° and hold the jump key after you get stuck (otherwise it won't work). Check out the demo:

Demo:
w3sp-think17[df.vq3]00.09.872(AMT-Morbus.Austria).dm_68

Last but not least: You will not be succesfull to apply this exploit in other maps. The coordinates of the affected brushes in W3sp-think17 are not integer (the vertices have odd float numbers for their coordinates).


Oh, maybe somebody wants to explain the maths behind all this.
Greetings, Morbus ;)
You do not have the required permissions to view the files attached to this post.
Image
User avatar
amt-morbus
 
Posts: 231
Joined: Jun 4th, '10, 8:28 pm

Re: More exploits!

Postby Nt.run » Jun 15th, '11, 8:44 pm

nice, especially think17, gz \o/ :D
Nt.run
 
Posts: 38
Joined: Oct 19th, '09, 2:25 pm

Re: More exploits!

Postby Nt.run » Jun 15th, '11, 10:28 pm

By some coincidence I was just about posting another trick which I haven't seen posted before:
triplejump, the trick solving w3sp-think10
When you get an ob on a slope, lowjump happens (you jump just like you're on flat surface). The trick is to get 3 jumps in a row on the slope with second one being a lowjump. That way you'll hit the slope right after first doublejump and get another one on it. :)
w3sp-think10[df.cpm]00.01.728(Nt-Runaos.Russia).dm_68
The demo is scripted, strafing itself isn't any hard while jumping that fast is certainly impossible for hands: one frame between second and third jumps. Though, there is one record in oldtop somehow. :roll:

Also, khetti has finished this map a long ago with other original way:
w3sp-think10[df.cpm]00.06.592(khetti.us).dm_68
You do not have the required permissions to view the files attached to this post.
Nt.run
 
Posts: 38
Joined: Oct 19th, '09, 2:25 pm

Re: More exploits!

Postby bb gl » Jun 16th, '11, 6:14 am

gj!
bb gl
 
Posts: 12
Joined: Jan 18th, '10, 5:41 am

Re: More exploits!

Postby Nt.run » Jun 16th, '11, 6:59 pm

This way requires some luck =)
You do not have the required permissions to view the files attached to this post.
Nt.run
 
Posts: 38
Joined: Oct 19th, '09, 2:25 pm

Re: More exploits!

Postby amt-morbus » Jun 16th, '11, 7:15 pm

Nice one run!

Btw: Using crouch works better than holding jump.
Image
User avatar
amt-morbus
 
Posts: 231
Joined: Jun 4th, '10, 8:28 pm

Re: More exploits!

Postby <hk> » Jun 16th, '11, 11:16 pm

I have a rough think map like layout for the ceiling bug for something like a "mdd-know01.bsp". I never finished it, like most of my maps.
All 'think' maps should actually be named 'know'. There is no thinking involved in solving them.



Strange I tried to solve think17 exactly like this. But I never bounced high enough. I gained massive amount of negative Z-Speed, though. In the end, I did get it to work, but had to make the triangle brushes a few units higher. You have to release the jump button in time to get stuck, I noticed back then. A one frame jump script did the trick. Still no luck.
I never tried holding jump though. :oops:
So close, man that's annoying. :(

I expected it had something to do with the falling code snipped i stumbled upon while browsing through the code.

Here is what I think goes wrong in the code to have this bug. I didn't work it out because I should have spend the time on my diploma thesis instead on code reading. Interesting how the coolest things turn up when you don't have any time for them :) Long story short. It's not much I came up with, but it's a start:

Usually the player's Z-speed will be set to 0 if the player is stuck within a wall for some reason. That's why you don't fall when you are stuck in the wall on SJT2.
At this brush you aren't stuck in the wall somehow and therefore gain massive amount of negative Z speed without moving down that much.



Here is the code that actually should prevent this bug form happening:
Code: Select all
if (trace.allsolid) {
      // entity is completely trapped in another solid
      pm->ps->velocity[2] = 0;  // don't build up falling damage, but allow sideways acceleration
      return qtrue;
    }


Since you slide down the wall you eventually come close to the floor. So close the well known bounce bug takes effect. This bug is actually a bounce. I don't have any data to support this theory. Maybe setup can help out.
Why don't we name it something like "wall stuck bounce" or "stuck falling bounce"?

I don't know why this trace doesn't detect the player being stuck in the wall. Probably some floating point stuff.
It's funny how the 'cylinders', that should actually have this bug, don't have it. Apart from all the other objects one which aren't really distractions, one sticks out.

On the original map shio made there were cylinders causing the bug. The cylinders here are too short for this on purpose, though. You have to put 1 and 1 together. Still this map doesn't have to do much with thinking. You have to grind code for this.
User avatar
<hk>
Menstruating 24/7
 
Posts: 1040
Joined: May 2nd, '09, 12:06 am
Location: Ingolstadt Germany

Re: More exploits!

Postby amt-morbus » Jun 17th, '11, 2:31 pm

With the right angle, you'll also get stuck between the other 2 brushes:

w3sp-think17bounce_AMT-Morbus.dm_68
You do not have the required permissions to view the files attached to this post.
Image
User avatar
amt-morbus
 
Posts: 231
Joined: Jun 4th, '10, 8:28 pm

Re: More exploits!

Postby Nt.run » Jun 18th, '11, 7:18 am

<hk> wrote:Usually the player's Z-speed will be set to 0 if the player is stuck within a wall for some reason. That's why you don't fall when you are stuck in the wall on SJT2.
At this brush you aren't stuck in the wall somehow and therefore gain massive amount of negative Z speed without moving down that much.

On Sjt2 you also gain massive amount of negative Z speed, without moving at all. :)
I don't know yet how this trick works inside, but its effect slightly differs between engines, some compilation stuff probably.
You do not have the required permissions to view the files attached to this post.
Nt.run
 
Posts: 38
Joined: Oct 19th, '09, 2:25 pm

Re: More exploits!

Postby amt-morbus » Jul 28th, '11, 8:16 pm

Image
User avatar
amt-morbus
 
Posts: 231
Joined: Jun 4th, '10, 8:28 pm

Next

cron