Opinions on Aim Assist? (poll)

In my FPS game I'm very focused on speed and movement and I'm starting to think that I should consider an aim assist on some weapons.

My thinking is that I will do a sphere cast and a ray cast when firing a hitscan weapon.

It then compares the the sphere cast to the raycast.

If the raycast fails to hit an enemy it then looks at the sphere cast.

If the spherecast hits an enemy it makes the target the center of mass of the enemy.

This way if the player does have a head shot, they get the head shot. But if they miss, but the sphere cast gets the enemy it will slightly cheat for the player to make sure its a hit.

I'm also thinking that I can increase the size of the sphere based on the player's velocity.

The idea being 1) rewarding the player for moving fast and 2) not forcing the player to slow down to take shots.

Again, I'm focused on speed and movement. I have zero desire to make a game about twitch skill like CoD. I want the game to be about picking targets, controlling the battlefield, rewarding aggression without some dumb mechanic like (UGH!) glory kills. (if you can't tell, I absolutely HATE Doom 2016).

There are some other benefits to moving fast such as increasing damage and enemies taking fewer shots.

Obviously I will have it a toggle in the options.

What do you all think about this and aim assist in general?

 

You may simply want to have two (the second being a capsule) colliders on characters. The second is simply scaled up by a tweakable factor and used by auto-aim. If the raycast hits the regular collider, all good. If it only hits the second: auto-aim adjust. Adjusting should work by moving the ray horizontally towards the enemy more than vertically by the enemy‘s height vs width ratio.

Be sure to cast the ray again, if the second time it still only hits the autoaim collider you should discard autoaim and treat it as a miss. That gives you a second value to tweak for cases where the humanoid is facing the player sideways, ie the target area is thinner and chance of missing therefore higher. Unless you represent this by an appropriately scaled collider. just some thoughts, ultimately it needs to feel right for your game, but details like this can make a big difference.

Why a capsule collider? You mention headshot so I assume humanoid characters, which are better represented by a capsule than a sphere. Otherwise missing sideways will give a much stronger autoaim effect than missing above the enemy.

As for design, it‘ll be tricky to tweak so that autoaim assists without it being noticable by players. If they do notice it takes away the reward of aiming well.

 

My personal preference is for weapons which don't need precision, over the game playing part of itself for me. If I have a sniper rifle and a rocket launcher and a shotgun and I choose the sniper rifle and miss, that's on me. I could have used either of the others to have a much easier time dealing damage.

There are exceptions to my don't-mess-with-my-aim rule. Returnal has pretty significant auto-aim and is still great, for two reasons. First, much like your own game, their focus is quite clearly elsewhere. There's plenty of challenge, so I don't feel like the game is going easy. Second, they don't try to hide it. Their reticule has a big box drawn around it, and if there's something in the box you'll generally hit it.

 

It would probably be faster to just increase the size of the head collider instead of spherecasting. I'd just count those as headshots, lots of games do this to some degree. What you propose could be perceived as "buggy inconsistent hitreg" when you sometimes do full headshot damage and sometimes not, while trying to hit a head.