139 replies to this topic

[daw22]

he is serious btw

[unkemptfiber1]

looks insane so far reading it all. Heisa will this include better side wall arieal reads and wall reads to redirect balls to the goal, lets say for example psychos, from your own half, that it registers that it can use the under side of the car to redirect it into  the net, psychos is a bit much but just for the sake of argument. Let me know if i need to explain this better.

[Heisa]

looks insane so far reading it all. Heisa will this include better side wall arieal reads and wall reads to redirect balls to the goal, lets say for example psychos, from your own half, that it registers that it can use the under side of the car to redirect it into  the net, psychos is a bit much but just for the sake of argument. Let me know if i need to explain this better.

Better sidewalls, better backboards, better overall for scoring, yes. It increases the chances of scoring than the other aerial modes.

 

 

Better psychos? No. At least not the first version of the redirect mode, as the goal was to first get something better at scoring than the other modes. It will require that you and the ball are facing the goal. 

 

Psychos are way too complex to predict and execute, and for the sake of time not delaying the release further, I'll be pursuing that later on

[faitbellik]

Hi @Heisa

 

I'm actually GC1 on all my acc without any software

I'm really curious about the improvements you'll do on V4

 

I have a fews questions for you

 

Do you think GOS will be good enough to reach high ranks like GC3-SSL ?

Does the movement with Aerial Bot looks randoms/robotics with the new pathfinder features ur developping ?

If yes , can we risk manual bans ? Like micro adjustments that looks unlegit like if we're using some bots/aerial bot

Do you have a release date ? ( even if it's a beta version ) (ETA > 2 months ? 1 month ? )

 

Thanks in advance for your answers :3

[Heisa]

Hi @Heisa

 

I'm actually GC1 on all my acc without any software

I'm really curious about the improvements you'll do on V4

 

I have a fews questions for you

 

Do you think GOS will be good enough to reach high ranks like GC3-SSL ?

There are people who have reached SSL with it, but not all of them, so I can't say for sure. They have posted on the forums if you search for it
 

 

Does the movement with Aerial Bot looks randoms/robotics with the new pathfinder features ur developping ?

If yes , can we risk manual bans ? Like micro adjustments that looks unlegit like if we're using some bots/aerial bot

It could look robotic, as it orients rather quickly and it almost instantly "locks in" the rotation of the car. That could be made more sensitive, but will see..
And yes, you do risk a ban, both from automatic detection and a manual report. And that can happen at any time, Epic Games/Psyonix has the ability to sneak another detection system if they wished so.

And I wouldn't say the current Aerial Bot is that much obvious in the v3.9 versions. The demolition bot or the wall wave dashes can be spotted if you abuse them, of course. It all depends on you
 

 

Do you have a release date ? ( even if it's a beta version ) (ETA > 2 months ? 1 month ? )

 

Thanks in advance for your answers :3

An ETA can only be given after I have finished it and thought it's ready to be released. So, no.

In the past I did try to estimate it, but proved myself wrong that it it's not that easy to estimate it. Many things happened, many things had to be changed/improved/fixed in order to continue moving forward.

I will be posting updates in this thread though. Checkpoints of what has been done and what's left. 

So stay tuned

[MYTDarkness]

Hello, any news about V4 release Date ?

[downfvll]

Hello, any news about V4 release Date ?

Ye why read literally the text above you from the Dev when you can just ask the same question

[Heisa]

Hey guys, 

 

Here's an update on my current progress on the V4.0.0 version.

 

Ball Prediction:

I've completely rewritten the ball prediction. It now works on all maps as long as the game mode and physics are supported, which as of now those include Soccar, Snowday ( basic puck prediction ), Hoops, Dropshot, and Heatseeker ( now that mode is too easy)! 

So now there will be no need to add support for new maps everytime Epic Games adds a new one. And there's no need for other types of ball predictions, such as the old engine ones. 

 

It now also runs externally from the game's process, so there will no longer be performance issues caused by the ball prediction. So that means a performance boost!

 

Then there is also a new stability improvement mechanism that I've implemented to reduce discontinuities between predicted ball positions. So no more jumpy lines that are distracting !

 

Ball Intercept Type 6:

Also had a change, it now models and predicts the jump, the turn time, and finally the boosting in the direction of the ball to find the earliest intercept. But it is a subject of change at the time I'm writing this post.

 

Path Finder:

This was and still is the most time consuming aspect of the new GoS:RL because I had to re-iterate many times by going back to change things in order to move forward.

Mainly the decision on the best path possible algorithm towards the target. In RocketLeague there's just too many variables to account for, to have the car drive on the best path. 

For example, by only going on the path that has the shortest time towards the target.. that alone does not always guarantee feasibility. It doesn't ensure the car will actually reach the intercept by going on this path.

So this alone, took a lot of predicting to actually ensure that the path is somewhat feasible. But I'm really close to an actual solution!

 

Then there you have the actual car input solver that drives on the given path by the AStar algorithm. This also took me many, many hours to get right, and it's still not 100% perfect, but at least it's close to that. As I discussed previously I had no clue where to start.

Each driving mechanic had to be re-done, steering and turning, throttling, braking, matching speed, handbraking and boosting.

 

The GoTo segments now include an aerial speed estimate solver, which once the car is in that range where it can successfully reach the target by aerialing now, it will match the speed of the aerial target. And if it's past that range, it will simply follow and stop on the target. 

This method also works while the car is on the ceiling.

I think the current approach is good starting point, because it doesn't interfere a lot by doing sudden changes to the speed of the car or overshooting/undershooting. So I decided to make it fall into a range. 

 

I'm also working on a path finder logic for when a map is not supported by the AStar algorithm, so at least we have working functionality on such maps. The car input solver will work on without a problem on unsupported maps, it's just that it will not be aware of the best possible path.

 

Another huge time consuming process - Dribbling segment. As previously discussed, it simulates possible dribbling positions of the car and picks the best one, influenced by weights.

 

It will have weights for you to tweak. Each weight will influence the picked dribbling position.

 

Currently there are: 

 

Bouncing - What bouncing outcome would a position have if the car drove, met and had contact with the ball there. Scoring for this would pick lowest bounce possible in the direction of the car. This is basically the bread and butter of the new dribbling logic of the [Path Finder]

It tries to minimize the bounce and it gives you great control of the ball's bouncing behavior. By killing the bounce of the ball, the car can start controlling it earlier and prepare it much faster for flicking. 

 

For example if you want to kill the bounce of the ball, you will have to select a [Bounce] weight value bigger than 0.0, for a particular dribbling state. e.g., [Catching].

But that alone won't be enough to have a successful catch, because having only that as a decision metric, it will simply kill the ball by any means, and that doesn't necesserally mean that the bounce will go in favorable position towards the car, so this is where it gets tricky.

 

Relative Velocity - The second most important metric here is the ability to match and slow down the ball's speed. Paired with the [Bounce] weight, you already have a pretty good base for catching/carrying the ball.

 

Distance to Ball and Distance to Car weights - These are useful for centering the car under the ball or picking positions closer to the actual car. Super useful too!

 

Alignment With Goal - Then we have the next important metric, but not that much since by default the dribbling will already offset the ball towards the enemy goal. This metric basically tells us how aligned will be the ball post-hit towards the enemy goal.

So including this will pick positions that would push the ball more towards the goal.

 

You have to be combining multiple weights together. And this is the most time consuming thing. Once you start fiddling with this, you will understand my pain ( lol )

 

The dribbling segment once perfected will be the grounds for future implementation of an auto flick feature.

 

Wave Dash:

Completely rewritten. As well as rewrote old jumping code, which improves the wavedash, the kickoff bot, and the flip cancels. 

 

It now uses new orienting logic to align the car with the surfaces. It has a more aggressive ability to orient towards the ground for faster recoveries and it ensures orienting just in time to land on the wheels.

The wall wavedash was also completely revamped. It now actually works together with the engine to ensure it actually does a meaningful wall wavedash, not just spam it. 

The older version spammed jumps which created a lot of unnecessary ones that didn't even increase the speed of the car. And caused unwanted jumps away from the wall. The new one has advanced logic to keep it on the surface, even when releasing the hotkey mid-sequence.

 

I've also added the ability to have a cooldown between wall wavedashes. It now also greatly minimizes random jumps away from the wall which are bad to have and is in general much more stable on the wall.

So you can basically hold it and drive on the wall, and it will wall wavedash for you!

 

And lo and behold, the rewriting of the wave dash allowed me to integrate more seamlessly, a feature that was requested many times - Zap Dash! A mechanic popularized by one of the fastest players - Dark!. 

 

This feature lets the user quickly gain speed even without boost or when no dodge is available.

From my current tests, it works on 30-240 FPS, and low-high ping ( including if you have high ping, which I've tested at 300ms ), works on ground, when jumping off from a wall or from an aerial, it can recover even if you don't have boost or a flip! 

 

The new Wave Dash bot also will come with new continuous mode functionality. You would be able to select between a classic normal wave dash and the newly added Zap Dash. 

For that I have also added a new feature which lets you control the continuous wave dash modes by holding the jump button and holding the wave dash hotkey. Giving us more control over it and decreasing unwanted jumps from it.

 

If you are interested in it, you can read more about it how I have implemented it, and user feedback, here: https://gamingonster...p-dash-mechanic

At the end of this post I will link a video I've made about it, to show it in-game, how I have *currently* integrated it and it's use cases. It is super OP! You'll be able to play just like Dark!

 

Flip Cancels

Also rewritten to a better manageable system. Removed old code and completely changed automatic flip cancellation to better detect the direction of the dodge so it can cancel it correctly. That includes flip stalls, now it should no longer try to correct the car on stalls.

This is also a potential fix for the users experiencing orientation problems when aerialing and double jumping. ( Was unable to produce on my end.)

 

With that change, I've also included the new Zap Dash mechanic. Currently it is only added after a speed flip using the hotkey. 

So once the user presses the speed flip hotkey, it will automatically zap dash after the speedflip. And holding the hotkey will chain speed-flip -> zap dash -> speed flip -> zapdash until hotkey was released.

 

I've also rewritten the control setting so when the user dodges to use their actual inputs, instead of the Aerial Bot's. Although this issue rarely happened on some users, this rewrite is for them in order to fix the issue of backflipping while holding the Aerial Hotkey.

 

 

For the next addition!

This one is big ! I'm sure a lot of people would be pleased, GoS:RL V4 officially has controller support! There is no longer a need for controller mapping softwares, which even sometimes caused problems such as missing frames and lost inputs! ( Yeah, I know it took a while, lol.) 

So let's all give thanks to @TheJani for letting me know about a game method that makes adding controller support super easy. This method also saved me a lot of time dealing with controller libraries, compatibilities and whatnot. So props to him.

 

I've also added the ability to map hotkeys really fast - a Capture Hotkey button which picks up both keyboard and controller buttons, as long as Rocket League actually supports that type.

It will also display possible conflicting hotkeys, really useful to see if any gamebinds and other bots use the same keys! Also if you have any in-game conflicts which might be messing up with the bots you use.

 

I've also added visual assist of the new aerial controller's path towards the target for new users, or for people who are not that familiar with the physics of the game.

Just with a glance you would be able to see if going for an aerial now is a good idea! It shows the stages of an aerial, or what it would take to get there, aiding even the new comers to the game who have no knowledge whatsoever about aerials.

 

That being said, after making the new Zap Dash, Flip Cancels and Wave Dash, mechanic, currently I'll have to go back in the process of tuning the behavior and the feel of the path finder, the new aerial controller logic as a whole (including the new feather boosting controller) together with the new aim at goal simulator, and the aerial modes associated with it (especially the redirect mode). 

Once the core functionality is done, updates with new features, forum suggestions, etc., will be much more frequent.

- I have plans of completely revamping the [Demolition Bot] too, the ideas are

- It will have the new aerial logic.

- Completely change the interception of cars

- A long awaited automatic jumping feature, not just simple - *if car has jumped*, therefore jump, but a full-on prediction to ensure the jump will actually hit the enemy car.

 

As well as the [Kickoff Bot] completely. (Have great ideas for that also, for example a kickoff recorder/replayer).

 

However, to avoid delaying the V4 release further, I will address those improvements in future updates. Although, I was watching some Dark gameplay clips and was inspired to add a simple Pogo aerial mode just for fun, lol. But we'll see as it is pretty exciting just even thinking about it.. 

 

 

There are still things that need to be implemented/adjusted/fixed. and who knows what would pop up again and what would need fixing of what I already have done.

As you can see, I've dived deep into the really complicated stuff and I'm genuinely excited to see what will come out from this long journey. It's a leap of faith, one built purely on trust and patience.

 

There are also external factors such as real life, wowz. So again, I cannot provide a release date yet, but will keep you guys posted about the futures, the changes and the work overall.

 

Here's the video I promised in the shoutbox about the new wave dash, wall wave dash & zap dash mechanics:

 

I really enjoy putting these together.

 

Thank you for the continued support through this time!

[Rglmca]

You're awesome!

[taylorsmx1]

yea this looks dope, awesome to be able to put eyes on some of the progress. love the tool and cant wait for v4!

[numsebacon]

Pogo aerial mode just for fun is crazy❤️

[Heisa]

Pogo aerial mode just for fun is crazy❤️

Yeahh, and it wouldn't take long to implement 

[Septemberhvh]

Amazing ! i've been using it for four+ years and i thank you for working so hard on the project ! V4 is a long waited update and i'm sure that you won't dissapoint us, take your time 

[Bully]

Nice work Heisa, I can tell by the tone of the post you have a bit of relief / weight off your shoulders compared to the original post. Seeing a bit of light at the end of the tunnel. I’ve supported you the whole way through including during the ban phase and I’m excited for V4 and what’s to come!

@TheJani also, I know you’ve been helping in the background because you’re just that guy! So special little shoutout to you also homie.

(Yes I know I’m not subbed right now, I’m currently spending school holidays with the kid - but I’ll be back, like always 8) )

Edited by Bully, 28 June 2025 - 10:08 .

[s2lent]

I've been inactive for some time, but the latest changes look very promising.
Has there already been progress with a ground dribble or flick bot in the past, or are there already plans for this?
Greetings and highly appreciated your work <3

Edited by s2lent, 29 June 2025 - 04:51 .

[numsebacon]

Yeahh, and it wouldn't take long to implement

So like predefined segments it has to go to with the aerial controller and then recovery?

Also anything regarding defense? 😳

[grizzy5150]

Thank you Heisa for the update on the progress! So excited to see what v4 can do!

[Heisa]

You're awesome!

 

Amazing ! i've been using it for four+ years and i thank you for working so hard on the project ! V4 is a long waited update and i'm sure that you won't dissapoint us, take your time 

 

Thank you Heisa for the update on the progress! So excited to see what v4 can do!

Thank you, guys.

 

 

Nice work Heisa, I can tell by the tone of the post you have a bit of relief / weight off your shoulders compared to the original post. Seeing a bit of light at the end of the tunnel. I’ve supported you the whole way through including during the ban phase and I’m excited for V4 and what’s to come!

@TheJani also, I know you’ve been helping in the background because you’re just that guy! So special little shoutout to you also homie.

(Yes I know I’m not subbed right now, I’m currently spending school holidays with the kid - but I’ll be back, like always 8) )

For sure there is some relief now (Seeing a bit of light at the end of the tunnel), but I will be completely relaxed once I have finished the complex features.

 

That being said, I wasn't sure if I was going to make something useful in all that time. All I had was blind trust without even knowing how I was going to do it. Just had a vision. So I kept trying and sticking to it. Trying different approaches and angles, and slowly throughout this journey, it's developing into something usable and useful

 

Appreciate the support and for being a part of the community for so long, really means a lot !

 

[anthooz]

Could you add the musty in the future? it would be interesting to combine the reset button and the musty button

[BadassCoco692002]

Hey Heisa,

a BETA version would be really cool – just for fun and to mess around a bit.

To be honest, I’m not very good at Rocket League, and the regular game modes just aren’t much fun for me at the moment.

That’s why I’d love to try out a BETA, so we can experiment a little, see what’s coming, and get a feel for the current state 

I read that V4 is incredible in Heatseeker, which is what got me interested in the first place.

Plus, I think it’s easier when the community can help spot bugs and issues just by playing – makes the whole process smoother for you