Forum
Projectile Aimbot
Created 1st August 2015 @ 16:44
Add A Reply Pages: « Previous 1 2 3 Next »
Quoted from Phnx
[…]
yes I have seen proof of him using projectile aim assist ;)
it is quite an obvious one tbh.
and it’s very useful, even when spamming from distance it’s better than how most pockets do it in prem now a days.
you were/are part of AC team? i don’t really think, so you haven’t possibly seen shit, there’s no possibility to create such good aim assist thing
Quoted from Phnx
[…]
when spamming from distance it’s better than how most pockets do it in prem now a days.
do you even know what you are talking about? pretty much every player does a-d taps every now and then which means that the very moment a player that’s far away starts moving it would shoot the rocket into nowhere
The accuracy of projectile aimbots is determined by its ability to predict the future.
Computers are good at extrapolating simple movements (this makes airshots ‘easier’ because there’s less room for movement). Humans are way better at predicting complex scenarios where stuff like “where is the enemy most likely to flee too” or “that bombing soldier may try to use that spire to jump towards” and adjust accordingly.
A middle ground would be a bot that simply shows you where the bot thinks you should aim and then use your brain to adjust based on your own prediction of the enemy’s movement.
Neat stuff though.
For those interesting in the maths behind this, it’s pretty simple:
In theory you can try to match equations for both the opponents movement as well as the trajectory of your projectile and try to ‘solve’ them. This is generally not doable because the equations spiral out of control quickly.
A better way is plain guessing the ‘time’ when the collision between target and projectile will happen.
Say you start at t=0.1s, predict the target that amount of time in the future. For that position solve how long your projectile will travel to hit that spot.
If you get a t larger than 0.1s that means the target will have moved away before the projectile can hit, so a potential hit is in the future, try again by increasing the t.
If you get a t smaller than 0.1s that means your projectile will hit the spot before your target even arrives at the predicted position, try again by decreasing the t.
Once the predicted target is predicted t seconds and the projectile fired at it will arrive at that same time, you will get a collision and thus hit the target.
As a consequence of this method, solving when the projectile will hit a spot is reduced to hitting a stationary target in 2D (the 2D plane defined by the vector of gravity and vector to the predicted spot). Solving in 2D is ez pz.
This method is pretty damn efficient, you can get accurate within 0.0001s within like 15 iterations. All that remains is accurately predicting the future, which is more tricky.
Last edited by Casual,
Quoted from Casual
The accuracy of projectile aimbots is determined by its ability to predict the future.
Computers are good at extrapolating simple movements (this makes airshots ‘easier’ because there’s less room for movement). Humans are way better at predicting complex scenarios where stuff like “where is the enemy most likely to flee too” or “that bombing soldier may try to use that spire to jump towards” and adjust accordingly.
A middle ground would be a bot that simply shows you where the bot thinks you should aim and then use your brain to adjust based on your own prediction of the enemy’s movement.
Neat stuff though.
For those interesting in the maths behind this, it’s pretty simple:
In theory you can try to match equations for both the opponents movement as well as the trajectory of your projectile and try to ‘solve’ them. This is generally not doable because the equations spiral out of control quickly.
A better way is plain guessing the ‘time’ when the collision between target and projectile will happen.
Say you start at t=0.1s, predict the target that amount of time in the future. For that position solve how long your projectile will travel to hit that spot.
If you get a t larger than 0.1s that means the target will have moved away before the projectile can hit, so a potential hit is in the future, try again by increasing the t.
If you get a t smaller than 0.1s that means your projectile will hit the spot before your target even arrives at the predicted position, try again by decreasing the t.Once the predicted target is predicted t seconds and the projectile fired at it will arrive at that same time, you will get a collision and thus hit the target.
As a consequence of this method, solving when the projectile will hit a spot is reduced to hitting a stationary target in 2D (the 2D plane defined by the vector of gravity and vector to the predicted spot). Solving in 2D is ez pz.
This method is pretty damn efficient, you can get accurate within 0.0001s within like 15 iterations. All that remains is accurately predicting the future, which is more tricky.
that’s what i wanted to say, but much much better, it’s really unlikely to find something that good, very close to impossible, and even the possible one, can’t be that perfect
because we should all try to crack ideas 4 better aimbots right? it’s abaud time we got this thread
also i’m not assuming baud was/wasn’t cheating i just can’t say coz i never seen any proof, apart from the funny heavy one.
Quoted from rockie
[…]
you were/are part of AC team? i don’t really think, so you haven’t possibly seen shit.
why do I have to be in the AC team to see his demos?
can stream them in 360p and 750 kbit since my net can’t handle higher, you can ask sideshow ;)
Last edited by Phnx,
Quoted from Phnx
[…]
why do I have to be in the AC team to see his demos?
can stream them in 360p and 750 kbit since my net can’t handle higher, you can ask sideshow ;)
gimme those demos, i wanna view them myself
Quoted from Phnx
[…]
yes I have seen proof of him using projectile aim assist ;)
it is quite an obvious one tbh.
and it’s very useful, even when spamming from distance it’s better than how most pockets do it in prem now a days.
haha ok
Quoted from Phnx
[…]
lol who wouldn’t
xdxd
Quoted from Spriggan
Does anyone remember this: http://i.imgur.com/ZgzyUVO.png ?
The Dreams Come True :D
Windows XP, why would anyone even want to remember that?
Quoted from kKaltUu
This is old stuff, I’m missing the purpose of this thread?
Download link… pls )))
Add A Reply Pages: « Previous 1 2 3 Next »
content rss