1 reply to this topic

[Thegoatrocket0]

I had some ideas for the future of Rocket League....

The bot's overall performance is pretty good, but I find it's missing some points.

For example, regarding prediction, would it be possible to add a feature so that the bot doesn't follow the insertion point ahead of the ball, but instead instantly heads for the ball? This would give better results for defending pledges or playing good 50-50s, and for better responsiveness in full auto boost.

Example 2: Fixed the issue with type 5 above 5.0, the bot in the air stays really high, losing its attack. However, if we set type 5 to a minimum, the bot is more stable but no longer as responsive.

Example 2 goes back to Example 1.

Example 3: For the defense logic, for a shadow defense, the bot works great when the dribble and match speed are removed. However, the dribble bot no longer works, but it defends really well. Would it be possible to improve this?

I managed to get a much better shadow defense with some Python scripts, and I think it would work really well in GoS.

Obviously, there's still some finishing touches to be done, but I think these changes can make a difference, as with v4, and I support the work you've done. I'll continue to support it.

[Heisa]

I had some ideas for the future of Rocket League....

The bot's overall performance is pretty good, but I find it's missing some points.

For example, regarding prediction, would it be possible to add a feature so that the bot doesn't follow the insertion point ahead of the ball, but instead instantly heads for the ball? This would give better results for defending pledges or playing good 50-50s, and for better responsiveness in full auto boost.

No, that's not how it works. The intersection point is the earliest ball state the car can reach in time. If you aim at the live ball - in the now, you'll always be behind in time and almost never catch up it.

 

The ball is always moving, by the time you aerial to it's current position, the ball will already be at another position, somewhere else. It's simple physics. So you don't aim where 'it is' in the now, you aim where it will be.

 

 

Example 2: Fixed the issue with type 5 above 5.0, the bot in the air stays really high, losing its attack. However, if we set type 5 to a minimum, the bot is more stable but no longer as responsive.

Example 2 goes back to Example 1.

Example 3: For the defense logic, for a shadow defense, the bot works great when the dribble and match speed are removed. However, the dribble bot no longer works, but it defends really well. Would it be possible to improve this?

I managed to get a much better shadow defense with some Python scripts, and I think it would work really well in GoS.

Obviously, there's still some finishing touches to be done, but I think these changes can make a difference, as with v4, and I support the work you've done. I'll continue to support it.

I suggest you just wait for the next version and see