So I've gotten into basic Super Mario World level design with Lunar Magic. Now I'm starting to create my own levels, and I've done some background reading on 2D level design.
I especially enjoyed the "cells and portals" theory put forth by A Framework for Analysis of 2D Platformer Levels (Smith, Cha and Whitehead, 2007) that each level is designed with one portal (exit) and several "cells" or pathways. I've also raided Gamasutra's servers for any interviews and essays on level design, though sadly they are mostly on 3D level design.
Keeping in mind I'm working with the Super Mario World engine, what advice would you have as players and/or level designer for a fledgling 2D designer such as myself (and yes, I have taken the "play more than you design" aesthetic already). I also have a few questions for anyone with 2D design experience:
1. What is it that makes Nintendo's designs so successful?
2. How prominent is pacing in a semi-linear platformer such as Super Mario World? Do individual levels need to "feed" off one another's pacing?
3. Are there any geometric rules to follow when placing blocks?
4. What's the easiest way to tell if a level's difficulty is unbalanced from the outside (I found that constant playtesting can create difficulty immunity)?
5. Should puzzles be designed around a single "screen" or backtracking throughout the level? Is there ever a "right time" to create a backtracking puzzle?
I'd appreciate any constructive advice thrown at me!
My current project: Something mysterious (oooooh!)
My username is all lower-case letters. Please get it right :(
I'm afraid I can't offer you any advice. This is something I've always found difficult, and I can never understand how people make entire sprawling 3D worlds, when I couldn't design a single puzzle room to save my life.
I've never been impressed with Nintendo's level design in SMW. Occasionally they make some interesting puzzles, but the majority of that game is run-right. I would suggest you look at the Klonoa series (the GBA games have some good puzzles, the PSX/2 games have some fantastic visuals and are still 2d-gameplay) for some design that I appreciate.
First of all, don't look back at Nintendo. What they're doing is not ideal; it's still improvable, which has been confirmed on numerous occasions. If I were to recommend a decent platformer developer, I'd rather name Konami or Treasure (also Nifflas and Pixel, if talking about indie scene).
1. Anything that can be completed by as much as holding right and pressing jump at irregular intervals is doomed. It's alright only for the first stage or two.
2. Not everything should be simple and intuitive. That being said, things that aren't shouldn't be brick walls, either. It's alright to make the player die once or twice to understand what he's doing wrong. Surprise deaths are bad, though. It's always better to challenge the player by patterns rather than something they can't read beforehand at all.
3. Difficult/tricky parts ("chokepoints") of the levels shouldn't require waiting from an experienced player, since that will just ruin the flow. You have various tools to playtest these. Backtracking is fine as long as you can make it equally fun with each pass.
4. Lack of balance comes from the sudden change in required precision. If the window of opportunity in executing a required action is small in either timing or positioning, it'll most likely be unfair.
Warp wrote:
Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.
Joined: 3/11/2004
Posts: 1058
Location: Reykjavík, Ísland
First of all the two most important rules of all:
* Do NOT use invisible platforms. Unless you know how to use them, you'll just screw it up and make the level crappy.
* Do NOT use platforms as a background if it looks solid. If you make a tile that is normally used as a solid into a background, your level will suck.
(the above rules can be broken for non-essential areas like secrets, etc, but you should still leave at least a hint so the player has a chance of finding them)
Players should always know exactly what they can stand on and what will kill them. Do not "increase" the challenge by making invisible spikes or pits, or fall-through platforms. You won't increase the challenge, only frustration, and players will stop playing your level.
Do make interesting things. Don't make just static level. Put some enemies in, moving platforms, etc. Don't overdo it though.
Make each level have a theme. It could be anything. Moving platforms, sloped hills, trampolines, pipes, or a combination of those. As with anything, don't overdo them though. Just look at Yoshi's Island. Every single level in the entire game has some new enemy, element and/or theme.
Do not put too many of the same things on one place. More than about 3 enemies in the same place should really be avoided except for special occasions. Same things with powerups.
Make each level interesting and a work of art. Make each level beautiful, so that when you look at it, you wish you could get in there and jump around. It should be like a window to another world. I guess this is pretty vague. :)
Obviously, don't make the level go for too long without a checkpoint. Especially if it's an autoscroller. Autoscroller level-layout should generally be easier than non-autoscrollers, because part of the challenge is keeping up with the screen.
Don't make your levels extremely hard. You will only make an overly frustrating level that very few people will want to play. Remember that to you, your own levels will be easy, even if they are really hard for other people, because you'll be testing it/practicing all the way through, and you know the level better than anyone else. So it's a good idea to get others to play-test your level. If they keep dying in one spot, swallow your pride and make it a little easier, even if it means they won't have to do that ultra-cool move you discovered. You want your players to play your level, not throw away the controller in disgust. (of course, if you're trying to make ultra-hard levels, please go nuts, just don't be surprised if no one else ever beats it.)
Edit: Also, avoid putting too many things in one level. By that, I mean do not put every kind of enemy into a single level. It breaks the immersion and makes the level feel like it's not a real place. Be creative, but consistent with your themes.
Etc, etc.
Go play Kaizo Mario World and don't do anything that it did. ;)
Try to vary the level settings. You're a bit limited here because you're stuck with what SMW can do, but each level should be different from the one previous. Look at the second island: outside level with flying guys, inside level, ghost house, outside level, outside level with no floor, castle. Having lots of variety like this helps keep the player engaged.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
This is a level I made before getting this advice (where Yoshi's Island 1 is normally; the level is called Spring Garden). You can play it here http://www.megaupload.com/?d=2QJK4HRG
It centres around two themes: a) having more than one way to solve any one puzzle and b) multi-tasking platforming and puzzle-solving. I won't spoil it, because a big part of the level is figuring out just HOW to beat it.
Rather than trying to justify the level's design, I'll put this level on the operating table for it to be judged, analyzed, and have tomatoes thrown at it.
*cringes*
My current project: Something mysterious (oooooh!)
My username is all lower-case letters. Please get it right :(
Don't make your levels extremely hard. You will only make an overly frustrating level that very few people will want to play. Remember that to you, your own levels will be easy, even if they are really hard for other people, because you'll be testing it/practicing all the way through, and you know the level better than anyone else. So it's a good idea to get others to play-test your level. If they keep dying in one spot, swallow your pride and make it a little easier, even if it means they won't have to do that ultra-cool move you discovered. You want your players to play your level, not throw away the controller in disgust. (of course, if you're trying to make ultra-hard levels, please go nuts, just don't be surprised if no one else ever beats it.)
Etc, etc.
Ultra hard levels don't always involve having to do death defying stunts, it's more like Acryte's levels which require cryptic soloutions. SMW does have dreadful level design or at least lacking, but its helped by the fact that the game has a great engine. One level you could make is a maze like level you have to navigate by using the cloud.
Joined: 2/28/2006
Posts: 2275
Location: Milky Way -> Earth -> Brazil
theenglishman, links to roms are not allowed, I guess
Make an ips and share it anyway, because it's faster.
Also, megaupload sux...create an account at some file hosting place, such as Fileden.
"Genuine self-esteem, however, consists not of causeless feelings, but of certain knowledge about yourself.
It rests on the conviction that you — by your choices, effort and actions — have made yourself into the
kind of person able to deal with reality. It is the conviction — based on the evidence of your own volitional
functioning — that you are fundamentally able to succeed in life and, therefore, are deserving of that success."
- Onkar Ghate
Bisqwit wrote:
My thoughts:
Use different styles in different places. Give the players as many dimensions as possible for identifying where exactly they are in game that they can use in their thoughts and in their speech; name tags are not enough. These dimensions include:
– type of scenery: caves, water, sky, etc.
– graphics: is the scene constructed of wood, question marks, tiles; what kind of backgrounds you use
– music
– certain shapes of platforms/enemy groups
Mix things that go well together. Create beautiful style.
– Avoid creating stark contrasts in colours without a corresponding change in the terrain / pace.
– Don't mix randomly: Use the same components together. Cave stages use cave music, etc. Certain enemies only live in certain inhabitats, so don't put a flowerhead worm in a pipe puzzle, for example.
Use a pace that goes together with the expected progress speed. Don't cram screens full of things. Let the player get an impression of progressing. Variations in pace are acceptable, of course.
Don't repeat the same thing many times. Players get bored quickly if they need to do the same thing again and again.
Don't create fake difficulty. Fake difficulty includes:
– invisible platforms which you must find in order to progress
– invisible platforms that you hit accidentally and die
– long spans of deadly floor / wide pits that you must cross by utilizing very unreliable platforms, often relying on luck such as the specific characteristics of enemy AI (so you can jump from enemy to the next)
– giving the player too little time to reasonably complete the level
– making the level impossible to complete unless the player chose earlier an action that seemed completely irrelevant
– splitting the route to two separate sections where you must have some kind of premonition in order to know which route leads to progress and which one is a dead end
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
You have to question yourself if it's meant to be a hardcore gamer game only or if you want as many people as possible to play it. What I mean is that you can make sure that almost everyone can go through the level, but the hardcore player gets rewards for more difficult tasks. I believe this is nintendo's approach on level design, keeping it casual fun, but with challenging optional tasks.
But if you choose for hardcore gamer, try not to make it way too hard too, because it will be boring for a homebrew game.
– graphics: is the scene constructed of wood, question marks, tiles; what kind of backgrounds you use
– music
This hack is based entirely on design, mainly becuase I have no skill in uploading graphics, and secondly because I just want to focus solely on level design.
FODA wrote:
You have to question yourself if it's meant to be a hardcore gamer game only or if you want as many people as possible to play it.
With the level I demoed above, I was taking a middle ground with that concept. (And for those who don't know what to do, you have to carry the springboard across the level. Come on, it's called Spring Valley for a reason!)
For example, look at the puzzle just after the midway point. There are two ways to get across, one for the casual (bad) player and one for the hardcore (good) player:
1. Activate the invisible blocks, get up to the Lakitu cloud and hold on to the springboard while flying across to the other side (casual)
2. Hold on to the springboard and rebound of the Koopas to get to the other side (hardcore)
In my tutorial level that I created (which I haven't shown, for good reason) I always tried to make a riskier but shorter path that those looking for a challenge would want to check out.
My design philosophy is that difficulty should come not from control limitations, but challenging the player to give it his/her all and make the obstacles themselves the challenge. This is in part why I chose the Super Mario World engine (the other being that Lunar Magic is AMAZING).
So...what do you actually think of the level above? What mistakes from that level should I avoid in the future? What actually worked?
My current project: Something mysterious (oooooh!)
My username is all lower-case letters. Please get it right :(
Joined: 2/28/2006
Posts: 2275
Location: Milky Way -> Earth -> Brazil
I'll tell you what I think as soon as you upload it somewhere else.
Megaupload is good for big files... storing small things there is a waste of time.
At least use Stashbox...
or
MiHD...
whatever, just put somewhere I can use.
"Genuine self-esteem, however, consists not of causeless feelings, but of certain knowledge about yourself.
It rests on the conviction that you — by your choices, effort and actions — have made yourself into the
kind of person able to deal with reality. It is the conviction — based on the evidence of your own volitional
functioning — that you are fundamentally able to succeed in life and, therefore, are deserving of that success."
- Onkar Ghate
Bisqwit wrote:
– Don't mix randomly: Use the same components together. Cave stages use cave music, etc. Certain enemies only live in certain inhabitats, so don't put a flowerhead worm in a pipe puzzle, for example.
Actually, you can't even do that without a degree of effort. In order to drop certain non-universal enemies into a habitat they are not typically found in, you would have to rewrite the enemy database for that environment to include the enemy. A bit too much effort involved for someone disinterested in altering the graphics, even.
Fly with the lakitu cloud up to the end. It's way easier. You can also jump from the help block, onto a bullet bill, to the end.
Hehe...oops. I only realized the Lakitu trick when my roommate went to test it, so...darn. Anyway, the bullet bill trick was intended but I'm altering now to make it (slightly) easier.
I've also been trying to do some colour matching. Not going very well I'm afraid :(
My current project: Something mysterious (oooooh!)
My username is all lower-case letters. Please get it right :(
Joined: 4/30/2006
Posts: 480
Location: the secret cow level
Bisqwit wrote:
– making the level impossible to complete unless the player chose earlier an action that seemed completely irrelevant
I played a Portal map where someone did that. You needed a box up until a certain point, and then you could progress down a hallway to the next area without needing the box. Once in the next area, there was a large room with two turrets behind forcefields you couldn't reach, you couldn't kill, and they could see the whole room and would kill you if you tried to reach any of the switches, platforms, etc. You also couldn't go backwards into the previous area. I ended up getting through the room by using noclip, but it always bothered me that I couldn't figure it out.
Finally one day I was replaying it, and I went looking for a walkthrough on Youtube. The player got to the aforementioned hallway, but he brought the box with him, and near the end, pushed the box into a wall, which opened up and revealed a gun ( from Half-Life 2. ) You had to bring the gun into the next area with you, and use it to shoot the turrets through the forcefields. There was absolutely no indication of the hidden alcove, and if you didn't catch it before entering the next area, you were screwed.
– making the level impossible to complete unless the player chose earlier an action that seemed completely irrelevant
In the changes I'm making to the level now, using the springboard is the easiest, rather than the only, option. For example, the player could potentially bounce off two Bullet Bills and land on the pipe (which has replaced the goal) or, if they bring a cape with, they can simply fly up.
My current project: Something mysterious (oooooh!)
My username is all lower-case letters. Please get it right :(
This doesn't really apply as much if you can backtrack to get the springboard, mind.
It's almost possible to backtrack, but I just need to have an extra flying Koopa spawn so players can jump over the large gap near the end from the other side.
I hope to post another update in the next few days with at least one new level as well as all the major changes to the first.
My current project: Something mysterious (oooooh!)
My username is all lower-case letters. Please get it right :(
I'm only a TaS-watcher, but consider myself a long-time casual player of games if that makes sense.
Anyways, here's my thoughts on your level/edit.
- I hated the jump just before the half-way point. We can't see the koopa we're meant to bounce off until we jump. Don't make 'leaps of faith' unless you have a good reason and the jump is a simple one.
- liked the hidden platforms going to the Lakitu.
- the row of koopas we needed to bounce off, one by one, is rather tricky.
- the entire level can basically be skipped by flying through it, rendering the 'puzzle' and challenges null and void. Maybe having the dragon coins at ground level to encourage folk to go the 'hard route' or having a secret exit at ground level or sky-bound obstacles would work well. Or maybe ensure that none of your platforms are long enough to build up the speed to fly.
- the goal floating in mid-air looks weird and glitchy. I know that you're experimenting with the tools, but having the goal on a really tall platform wouldn't have changed the level itself, but would have made the positioning in keeping with SMW's internal rules. If you want to be a pro level designer, you've got to pay heed to those internal rules that every good game establishes.
- If a player reached the mid-way point then dies, they can't use the spring for the exit (without going back and doing some superhuman feats...)
The entire level seemed a bit patchy and bare. Two sections - the gaps with winged koopas - were tough, whilst most of it seemed like areas between them, rather than parts of a level in their own right. It may be worth extending a theme more fully across a level and ensuring that you don't have 'dead zones' that just seem like areas to traverse, without much happening.
I got some fun out of it, but it fell short of the quality of good professional output, and those are some reasons maybe why.
I'm just some random guy. Don't let my words get you riled - I have my opinions but they're only mine.
Thanks for the analysis Bezman. My levels are being designed as a change of pace for the typical 2D Mario crowd and will test timing, puzzle-solving and situational analysis. It's great to see a casual POV though.
Bezman wrote:
- I hated the jump just before the half-way point. We can't see the koopa we're meant to bounce off until we jump. Don't make 'leaps of faith' unless you have a good reason and the jump is a simple one.
Good point. I've moved the Koopa over one block to the left so that he is within the draw distance.
Bezman wrote:
- liked the hidden platforms going to the Lakitu.
The Lakitu cloud has been removed in favour of a high road, because I discovered that the whole puzzle could be skipped just by flying. There should be other ways to skip it instead.
Bezman wrote:
- the row of koopas we needed to bounce off, one by one, is rather tricky.
Play around with it. If you get Mario up to the maximum P-speed and take a full jump you need only bounce off the middle one.
Bezman wrote:
- the entire level can basically be skipped by flying through it, rendering the 'puzzle' and challenges null and void. Maybe having the dragon coins at ground level to encourage folk to go the 'hard route' or having a secret exit at ground level or sky-bound obstacles would work well. Or maybe ensure that none of your platforms are long enough to build up the speed to fly.
The get-out-of-puzzle free is for advanced players (or those who have already played through the level) as a means to an end from a speedrunning or short-for-time perspective. I never (with some exceptions) want the player to feel forced into a puzzle they can't or don't want to solve. Just because a certain path is the "proper" way doesn't mean it should be the only way.
Bezman wrote:
- the goal floating in mid-air looks weird and glitchy. I know that you're experimenting with the tools, but having the goal on a really tall platform wouldn't have changed the level itself, but would have made the positioning in keeping with SMW's internal rules. If you want to be a pro level designer, you've got to pay heed to those internal rules that every good game establishes.
I have replaced that with a pipe, followed by a short (and unexpected) run to the exit.
Bezman wrote:
- If a player reached the mid-way point then dies, they can't use the spring for the exit (without going back and doing some superhuman feats...)
Yeah, that occured to me too. The midway point's been removed
Bezman wrote:
The entire level seemed a bit patchy and bare. Two sections - the gaps with winged koopas - were tough, whilst most of it seemed like areas between them, rather than parts of a level in their own right. It may be worth extending a theme more fully across a level and ensuring that you don't have 'dead zones' that just seem like areas to traverse, without much happening.
This particular level was designed to be paced in a certain fashion, lulling the player into a false sense of security and then hitting him hard with a particularly tricky jump or other obstacle. That was entirely deliberate.
Bezman wrote:
I got some fun out of it, but it fell short of the quality of good professional output, and those are some reasons maybe why.
The level's been drastically changed (and some new ones have been added). I will be uploading the new ROM some time tomorrow for further analysis.
My current project: Something mysterious (oooooh!)
My username is all lower-case letters. Please get it right :(
Joined: 2/28/2006
Posts: 2275
Location: Milky Way -> Earth -> Brazil
at least zip it...
"Genuine self-esteem, however, consists not of causeless feelings, but of certain knowledge about yourself.
It rests on the conviction that you — by your choices, effort and actions — have made yourself into the
kind of person able to deal with reality. It is the conviction — based on the evidence of your own volitional
functioning — that you are fundamentally able to succeed in life and, therefore, are deserving of that success."
- Onkar Ghate
Bisqwit wrote:
- the entire level can basically be skipped by flying through it, rendering the 'puzzle' and challenges null and void. Maybe having the dragon coins at ground level to encourage folk to go the 'hard route' or having a secret exit at ground level or sky-bound obstacles would work well. Or maybe ensure that none of your platforms are long enough to build up the speed to fly.
The get-out-of-puzzle free is for advanced players (or those who have already played through the level) as a means to an end from a speedrunning or short-for-time perspective. I never (with some exceptions) want the player to feel forced into a puzzle they can't or don't want to solve. Just because a certain path is the "proper" way doesn't mean it should be the only way.
But flying through it feels so unsatisfying and isn't a particularly interesting thing to do!
If there were 2 ways through the level and each had its own 'style' of challenges (e.g. maybe one needing us to fly through small gaps, the other needing somewhat precise jumping) then I could see your point. But flying as is doesn't feel like an 'alternate route' - instead, it reminds me of Rainbow Bell Adventures, in which to fly is essentially voluntarily skipping all the action.
It's not having an alternate route that I disagree with - rather, it's having one that is so uninteresting that rankles.
I think that if someone doesn't like the style of the game, rather than give them an absolute 'get out clause', just appreciate that they may not be the audience for the game.
If you want multiple routes, give each a modicum of challenge. Or use the map-screen or warp-pipes to implement multiple routes.
theenglishman wrote:
The level's been drastically changed (and some new ones have been added).
Cool! I'll look forward to checking it out!
On a slightly-related note, how did you do edits like that? Is there a tool in existence to make the job easier? Is there some FAQ, guide or other documentation that you used to help you get started? How long did your first level take you?
I'd love to give something like this a go if there's not too much coding-type-work involved.
I'm just some random guy. Don't let my words get you riled - I have my opinions but they're only mine.
