How does one make good controls?

That wouldn't be hard considering mario slips around like he's permanently standing on ice in mario64.

Maybe the control stick is too fresh or something?

Not sure. With the joycons you also had the motion controls which could be used to enhance some abilities. I don't know about the feeling with any other controllers.

Controls are good, but it's sometimes hard to wrestle the camera when trying to precisely go in one direction. I'd rate them a small step down from Sunshine, but the hat gives some decent recovery options.

This, also. Vertical locations usually have few possible routes and that limits fun a bit.

Mario 64 controls were just Mario in a bare bones room with some obstacles and props.
If you can make the controls fun with just how they let you move within a room like that, then you've made great controls.

How exactly to do that, requires trial and error plus study into what sort of game you're making.

Responsive character controls are one half of fun gameplay. The other is a good camera system. Just try slogging through the japanese version of spyro.

The GameStop Switch and JoyCons just seem to handle like shit and for some reason the control sticks almost seem to lag. Probably because they're not supposed to be hooked up the way they are.

...

Well making controls "so good Modders wouldn't want to change them" generally isn't how it works. Usually what ends up happening is changing the controls ends up being too much of a hassle because of how the game's engine is designed, if you wanna change the controls you have to change other things as a knock on effect. The key lies in how objects interact with each other and even how you go about making the maps. As an example despite all the praise Republic Commando gets for instance, it's really just a bunch of interactive set pieces plopped down all over the place and your teammates relies on a finite state machine to co-ordinate between different states as determined by which orders you give them, that's why they react to your orders in real time and why they go retarded when you order concentrated fire on an enemy. They also probably have a decision tree they operate under independent of the finite state machine but the is besides the point (and topic) the controls of the game is essentially "interact with this set piece" and "change state X to state Y" for otherwise autonomous bots. But the whole game is set up so that you have to basically solve fire team themed puzzles to proceed through the map so people love it.

It felt really tanky for me, you had the turn radius of a fucking truck. The only way I could walk into gun stores was by aiming my gun at the door to make the slav fucking go the right way.
I think what makes games controls good is responsive controls and freedom of movement. .
I really liked MGSV's controls. The retarded fucking rock collision, making you slide off rocks that were 30º up, kind of fucks the freedom of movement part, but walking around felt really good.
The reason for that is that it simply makes the character go where the player tells him to go. In GTA IV it makes the character go towards the general direction the player's pointing. If you wanted to go to something behind you, in GTA IV you'll do a fucking half circle truck truck, in MGS V it'll simply spin the character around in a quick animation, then go straight towards where you want to go. That made MGSV feel really good to control IMO.
GTA SA was also pretty good. Being able to jump up a wall then climb up a roof by jumping towards the roof and climbing the ledge of a roof then getting a SMG. It allowed the player a lot of freedom, letting him climb shit. The movement itself wasn't anything special, but it did what it was supposed to be and let the player go where he wanted to go. GTA IV went a step further with the freedom of movement thing and allowed you more choices in missions, like letting you climb ACs and shit on the side of a building to get an advantage point on one mission, but it ruined it with the truck movement. GTA V went a step further down and removed that, now if you jump at a wall from 10cm away you just bang your head and fall down like a retard.

Damn. It's nothing new, but it's still nice to see a dev publicly speaking out about this sort of shit, even more so about their previous contractor.

Super Mario 64 was being developed before the Nintendo 64 hardware was even finalized.

Miyamoto and others emulated what the hardware was presumed to be like on a supercomputer with mock controllers as well. They simply had Mario and Luigi in a room and would jump around and everything else. Apparently they simply did this until the controls felt right.

So, to make good controls, simply keep tweaking things until you stop finding things wrong with it. Reducing latency is a good way to start. Once there's a minimal amount of latency, try with other things until things feel right. A human being is the one who will be using these anyway.

Aside from that, you should test on others. I know they also did this with children playing early versions of Super Mario 64 and I recall reading how they noticed the kids having fun just running and jumping around and not even doing anything much. For a video game, it's obviously important for the controls to feel good for the largest amount of people, but since people seem to have similar taste in controls like this, this shouldn't result in much compromise, if any.

Here's the interview I read: shmuplations.com/mario64/

Here's an important excerpt:

Seems like a great combinations for actions that have to be repeated over and over and are grounded in reality.
Though there's a bunch of things that I think should be considered when making a great control system, especially regarding movement options: if some option gets overused unintentionally i.e. the combat rolls in Zelda or that annoying Billy Hatcher dash, then it needs to be reworked into the main movement option and rendered less broken. One thing I notices is that in games that have charging/dashing mechanics, there should be an equal amount of risk and reward for using either the normal options or the advanced, charged movement styles, or you risk having your player rely on them too little or too much. I.E. Sonic's spindash in 2 and 3 is an example of a good charged move that offers more speed for a short burst, charges comparatively fast, trades acceleration for handling and overall gives you a hard time if you're not accustomed to the game. In comparison, shit like Sonic Adventure's charging to get through ring chains is slow and bothersome, while shit like the combat roll in Zelda are bothersome to keep doing and offer just advantages compared to moving normally since they're short and can be chained with more precise turning.
Also, for games that rely a lot on vertical mobility, there need to be more than just two ways to jump real high. It's great to see people use different tech for clearing out different obstacles.

But aren't you just limited to a joystick and a single action button? Have we gone so far back that the Atari 2600 is the new standard for portable controls?
Though I do agree that onscreen buttons are garbage, but it just seems like you're limiting yourself unnecessarily with that scheme. There's this one game, PES 08 Wii, which really opened my mind to how a game as set in stone as sportsball sim #1000 can still play the same if you stretch things a little bit. Essentially, you waggle for shooting or tackling, but ever other action is performed by pointing the remote at the screen and pressing a button once, twice, holding it or dragging a player to that area with that button selected. In a way, this could save you from using a joystick and just limiting the on screen button to a row of three or four of them for a combination of moves. Obviously there was more to that than just waggling but that's not really the point.
One other thing, I don't think that removing features can be copyrighted, but then again I'm not real good at spotting new Jewish tricks.

That is also true. Though you can get around that by either making the game "seemingly easy" by having floatier controls on jumps/bigger areas to land on/moves that don't require too much space to execute etc. or just put a fixed camera and get around technical limitations.

I spent time writing what I believed was a good post yesterday, but it was deleted. My main point was to work on the controls until they could be perceived as good, since a human will actually be the one using them. Reducing latency is an obvious way to start and then testing with a wide range of other people should help.

Here's the interview I linked to in my much better post: shmuplations.com/mario64/

Bumperinos