next »

[TwilightVulpine]

Well, and then they mix those "realistic" jumps with with frustrating jumping puzzles that make you want to scream "I HAVE HANDS! WHY CAN I NOT JUST CLIMB ONTO THIS 2 FOOT HIGH SHELF!?" Personally I thought jumping was not terribly well done in HL2, though I've certainly played much worse too.

Mirror's Edge was a 3D first-person game that used realistic jump in an interesting way. It had it's deal of frustration, since first-person games are inherently not very good for jumping around, but it was a game in which you could climb stuff and actually have (kind of) realistic movement not limited to jumps. It didn't use any type of movement unseen in games, but it brought them to a new perspective, literally.

[baconman]

All FPS games need to fix that, is a MegaMan X style wall jump, and an appropriate (depthwise) margin of error. Maybe a slight change in perspective when doing that, so you don't accidentally RocketWallJump yourself to death (and not to mention, the side of the hole you just came from).

Well, then; platfomer games can measure jumps in tiles. And while 3D games can do something similar, arguably; most people do NOT build 3D games out of consistently-sized (uhh..."cubetiles?"). And when they do, it's not always obvious how big they are; unlike with most standardized 2D games. That, I believe, makes all the difference in judging jumps, however realistic/surreal the physics may be.

[gimymblert]

All FPS games need to fix that, is a MegaMan X style wall jump, and an appropriate (depthwise) margin of error.

I curious to see how it would work using a First person perspective with a wall blocking the view? (by the way: warsow)

[Derakon]

The way I'd see it working is, while you're in midair, the game would figure out that there's a ledge you can grab onto (basic jump prediction; this is fairly tricky, but should be doable) and give you the option of hitting a button to grab on. If you do, then your view becomes locked to the ledge, so you look up as your body slams into the wall, and then haul yourself up and over.

[gimymblert]

but it's ledge jump not megaman X style wall jump

[Cevo70]

'scuse me while I jump back on topic for a sec

Someone else metioned this already, but one thing that bothers me is the overuse of "realistic" physics engines used in platformers.   How brutal.   We actually plugged a physics engine in to our platformer in the early stages of development and quickly hated it.   It never feels right/clean, nor truly realistic, and certainly not fun.  Just sort of buggy and restrictive.

Part of of the appeal of a platformer is to simplify and exagerrate something that is actually pretty mechanically complicated and not very exiting (for humans anyhow).

There is a subgenre that executes it well but those tend to choose that path to deliver a sense of adventure.  Those games are more about exploration and probem solving - case and point the sexy new Limbo

. 

[Kegluneq]

Interesting topic we've got here!

My proposal for, what did we call it? varied jump heights which only require input at the start of the jump (a la real life):

Unfortunately this would require the use of a gamepad, but somethings better than nothing right?

In you're typical platformer, the only buttons you'll be using are you're movement (left control stick or d-pad), jump (most easily accessible button, usually A), and some secondary button. The input method that is rarely ever used however, is the right control stick.

My idea for jumping involves "flicking" the right control stick (which from now on will be referred to as the jump stick). The distance you pull the jump stick will determine the power of the jump, which will be executed upon release. A little similar to walking and running with our movement stick.

Now with some extra calculations involving the movement stick, I feel like this has the potential to let players execute a large variety of different kinds of jumps.

While standing still, you could do you're basic jump with varying powers in 180 degrees of direction: straight up, forwards or backwards.
The same could be applied to jumping while moving, only it includes your initial velocity from running.

I feel like this could open up a new kind of platforming that has at least semi-realistic physics (no mid-air direction or height changing) but still be challenging and entertaining.


What do ya'll think?

[SirNiko]

I've seen some flash games that use a similar method by making you click a certain distance and angle from your 'hero' to do different length jumps.

My concern is that this will be really inaccurate, and that the variety of possible jumps will be ruined by the fact that you'll need to practice a great deal in order to perform the jump you want precisely and accurately. The player will just keep repeating the level and trying until they coincidentally get the right jump, which is how a lot of the flash games seem to work to me.

The ability to perform a standard length jump in any direction you press the right stick could be implemented better, as the player need only aim and not adjust power, which increases the precision by a great margin.

But maybe other people are more positive about the flick-jump idea.

-SirNiko

[baconman]

That has good potental for a jump-dash stick (since theoretically, a "dash" can simply be a "horizontal jump"). But again, ideas vs. execution.

If I were executing an idea like that, I'd define it to 4 jumping levels:

1 = gentle tap in one direction, partially
2 = sustained hold in one direction, partially
3 = quick tap in one direction, fully
4 = sustained hold in one direction, fully

You might be able to pull off a 5th and 6th variable as well, defined by how far the stick is moved (more or less than half-radius), but doing so would take some considerable experimenting and feedback to get an actually good range-difference between the "close partials" and "far partials."

ALSO: What would your character do if you pressed the stick downwards?


(On that note, different down-up combos could result in absolutely CRAZY geometric jumping potental, too! Then the trick is making THAT actually USEFUL, and not in a frustrating kind of way. But I like the idea!)


EDIT: Is it just me becoming dyslexic, or does it seem like the forum trims words from the posts, sometimes?

[Kegluneq]

Yea it's not quite the most perfect idea. But there's some potential in it, and I feel that if anyone were to give it a go, it may actually turn out decently. I'd try it myself but I don't have a gamepad

I dunno what down would be, maybe crouch so you could make even higher jumps?

[TwilightVulpine]

It could be a body slam or other kind of fall attack, something that could break or go through passable ground under the character.

[sjakaus]

why not just define jump height by the y-axis of the movement stick. Could do direction too this way. That is, not holding the stick at all gives you a tiny, pathetic, hop, and pushing the stick full upwards gives you a full jump.

[AndroidScholar1]

I love games with semi-realistic physics, rather than mario-esque physics - you know, the kind of physics that actually gives a feeling of weight.

Two examples I can think of this are the games Oddworld: Abe's Odyssey and Heart of Darkness for the PlayStation - two fantastic platformer games.

[Seth]

Well in River City Ransom if you ran into a wall you would get knocked down.  It wouldn't hurt you but you could still die.

[ansel]

why not just define jump height by the y-axis of the movement stick. Could do direction too this way. That is, not holding the stick at all gives you a tiny, pathetic, hop, and pushing the stick full upwards gives you a full jump.

seems to me like that'd be hard to manage in situations where quick reflexes are needed. jumping too high or too low again and again before finally nailing the perfect height could be extremely frustrating

[gimymblert]

That's the too low which pose problem, dead zone create ambiguous reaction. But you make it more intuitive if you care about velocity at the same time that inclination of the stick while analyzing. A lot of tweak is needed

[jwk5]

I don't mean to raise this thread from the dead but I recently came across this page detailing how Sonic the Hedgehog's platforming engine works (it could have even been from this very topic that I found it on, I don't remember exactly where I came across it) and I've been trying it out in Game Maker using collision_line() as the "sensor lines" mentioned on the page and I am getting pretty satisfying results. I think it is definitely worth looking into for anyone looking to make a platformer. More than likely I will post a .gmk file of what I came up when I got it working smoothly.

[iggie]

I'm working out how to make a 2D platformer in a freeform (non-regular blocks) world. I'm aiming at less realistic physics, but with realistic characters.

Games like Prince of Persia and Lara Croft have developed quite strict level design hints and auto-locking jump systems to allow crisp platforming in a more loose world - also clearly marking good spots to grab, climb with different models and textures.

On the other hand I felt Mirrors Edge had too little in the way of hints and auto-correction so you had to rely too strongly on guesswork to jump correctly.

My favorite games make the actual jump work predictable and easy to land precisely where you wanted - but complicate things with enemies and timing. It is so frustrating to simply fail a jump when there are no distractions, just because you can't control the jumping cleanly. It is interesting to hear Metroid Prime's solution - consistent distances and big platforms.

If things get too messy I'll probably just constrain the world to a grid in all areas with critical jumps.

[J-Snake]

That is actually a very interesting thread and the issue is very important to me. I have given that a lot of thought but I have a certain philosophy regarding interaction-depth that is helping me out.
I am interested in creating a character-behavior-model that provides possibilities for the player to reflect himself in his actions, to grow and expand his skill-set in the performed actions and action-combos.
It can often require things like having no momentum, stopping motion directly to enhance precision but it is not true for everything. For example the player wants the character to start running directly the frame the input is given. But on the other hand there can be games where you want to position yourself pixel-perfect to pass small holes and perform breath-taking actions. So the compromise is to find fast but smooth acceleration so you can make nearly pixel-perfect stops after starting running.
It sounds that I am very nitpicky at this but it provides more interaction-depth and allows the player to reflect himself more in his actions. And least but not last appropriate controls-design allows to perform awesome things in real-time.

here is some of my philosophy for a certain type of games:



and here are some awesome actions:

[eva_]

let talk about mario babby