When do you start making levels?

ceaselessly

Member
Your preproduction is done, you've got at least rough art assets in place, and most importantly, have spent hours and hours refining your game's basic mechanics. Is this when you start designing the levels themselves? Or have you done that already? Or is that not even on your radar yet?

I'm working on a 2D platformer right now and am pretty happy with how the movement mechanics and state machines are working. I finally figured out what the expletive I'm doing with vertical moving platforms, and I've programmed a few enemies and pickups. Is it time to start seeing these things in a context resembling actual gameplay?

I understand there's probably no one correct answer here, but would love to hear others' thoughts about this. Do you start designing levels out of the gate, or do you fine tune the mechanics first? Does implementing levels inspire you to keep working at fleshing out the programming, or does it distract from it?

What's your workflow?
 

Xer0botXer0

Senpai
Just my thoughts here:

I think as you work on levels you notice new features that you'd like to implement.
So maybe work on 1-3 levels and see how they go.

Perhaps you're breaking your levels up into stages, like you see in some games, you've got Stage 1 with 3 levels in the jungle biome, then stage 2 with 3 levels in a different biome.

Not really necessary but Isee a lot of games breaking them up like that and giving you stars for finishing them, like angry birds. if you haven't gotten all the stars you may feel inspired to revisit the old levels. Obviously new levels unlocking as you finish the old ones or progress in astory line(dont ask me about story lines though I've got 0 clue on em)

I've worked on one particular game for a while, trashing it and starting over and I've noticed that each time I started over worked on different things first.

Since it's to be an mmo I usually setup the handshake and such first, then character design, then if i recall the one time I first worked on player movement and designed the world generation and location transitions and then houses and inventory, another time I setup one room even skipping character design this time and worked on multiplayer interaction, workbench interactions and so on.

I feel like if there is an issue it's that you don't want to delve deep then feel required to backtrack.
 
According to my current and previous project, levels got/get done very late.

The previous project had a level-editor that was usable for defining attack patterns, but the backgrounds all had to be placed manually. And, the biggest error was not being able to start the game at a certain level, the game always had to start at the beginning. It was a rhythm style game, so if I tried to start the game at anywhere else besides the beginning, the timing would be all messed up. BIG mistake, made testing a nightmare. Make one change near the end of the game - had to be tested from the beginning again. I hacked in a kind of *fast forward* mode and infinite lives to try to make testing easier, but it still sucked.

Current project, building a level editor in-game where you can instantly switch between play-mode and edit-mode seamlessly. All level building is done in-game, so no need to re-compile when level changes are required. Level data is all saved to disk, and is shareable to other team members, so if one person builds a level, they can just send the level data files to someone else, and they can play/edit the level on their machine.

I'm doing the programming, so level design is not my main focus, building all the tools to make levels using the editor is. Still have features that need to be finished.

The artist/designer will start building levels proper once most of the level editor is completed.

I expect that there will still need to be some tweaking of the mechanics even once we start building the levels. The positive thing is that if any level changes are required, it can be done super quick.

Even in its unfinished state, its still possible now to make levels. When the designer does make a level or two now and then, it is very inspiring seeing the thing working. So building some early levels definitely helps motivate further development.

Current workflow is to prioritize the most important features, focus on developing and polishing the feature/level item. Once its implemented, test the feature out in-game to make sure its working properly. Then on to the next thing.

Like you said there's no one right way. The down sides of developing like this is extra effort required as we haven't locked down the level design first. This leads to sometimes big re-writes of certain parts of the engine to accomodate stuff that wasn't previously forseen. Having a level design from the start would have made it easier to know in advance what all the features that would be needed, and build the system accordingly.

The one advice I would offer is have a way to build and edit/change levels that is quick and efficient, and able to be shared easily. It will make designing and testing the levels very fast, meaning you can iterate your level design rapidly.
 

Khao

Member
Make levels when you're starting out, but don't make levels to include them into your final game design. Make levels to test your mechanics out. When you know what's works and what doesn't. What's fun and what isn't. What's easy to understand and what's too complicated. What's easy to execute and what's difficult. That's when you make your levels for real.

Also, don't make your first level first. It's actually a good idea to make your first level last, because by then, you'll know what your entire game can offer, and you can build your first level around the idea of preparing the player for what's coming. It can be good practice to start with your game at it's peak of difficulty and complexity, and build towards that moment backwards.
 

ceaselessly

Member
Not really necessary but Isee a lot of games breaking them up like that and giving you stars for finishing them, like angry birds.
Funny how our monkey brains can be tricked into replaying the same levels just to get an accursed star!

Current project, building a level editor in-game where you can instantly switch between play-mode and edit-mode seamlessly. All level building is done in-game, so no need to re-compile when level changes are required. Level data is all saved to disk, and is shareable to other team members, so if one person builds a level, they can just send the level data files to someone else, and they can play/edit the level on their machine.

I'm doing the programming, so level design is not my main focus, building all the tools to make levels using the editor is. Still have features that need to be finished.
I love this idea. I'm pretty mediocre at programming, so that's a little over my head. However, now I have another landmark to strive toward! Are you developing the level editor in Game Maker, in tandem with the game itself, or separately?

It's actually a good idea to make your first level last, because by then, you'll know what your entire game can offer, and you can build your first level around the idea of preparing the player for what's coming.
Agreed! It's like writing a story by starting with how you want it to endā€”you always know what you're headed for!
 

Xer0botXer0

Senpai
I remember when I was younger I decided to make an ingame level editor so I checked out age of empires level editor to see how that's done, and then took some tips from games like prison architect..
I remember another time when I worked on an ingame level editor for a multiplayer game which was supposedly real time. I never got to adding all the features I had in mind.. too many other important things, I couldn't fit it all in my brain lol.
 

JackTurbo

Member
I've not made a platformer before. But what I tend to do is initially make a playground stage.
An area that I can test the games mechanics in to get a feel for things.

Personally I don't feel confident doing level design until at least most the players abilities and mechanics are in place. Level design seems so intrinsically tied to player mobility that I don't think I could do it without having the player pretty well defined.

So for me level/environment design is a very late stage job
 

pixeltroid

Member
Do you start designing levels out of the gate, or do you fine tune the mechanics first? Does implementing levels inspire you to keep working at fleshing out the programming, or does it distract from it?
You cant isolate the mechanics from the levels. Certain movement mechanics (such as wall jumps etc) can only be tested out in a level, or at least a part of it. Of course, you don't need a fully decorated level to be able to test the mechanics, but IMO a "finished" level (with tiles and backgrounds) somehow feels more inspiring.

Level design is AS important as the mechanics, and also needs to be tested for issues. And in order to do that, you need to create a level.
 
Are you developing the level editor in Game Maker, in tandem with the game itself, or separately?
It's all being done in Game Maker. So the game and editor are all in the same build/Game Maker project.

I've also made it so that the editor is de-coupled from the game itself. The Editor can be disabled and then just the game itself will run.
 

pixeltroid

Member
I'm working on a 2D platformer right now and am pretty happy with how the movement mechanics and state machines are working. I finally figured out what the expletive I'm doing with vertical moving platforms, and I've programmed a few enemies and pickups. Is it time to start seeing these things in a context resembling actual gameplay?

Sounds good. I think you should now create a test level and throw everything in there to see how it all works together! :)
 
I always make levels. How else can I test my games? During the making game phase, the levels are more to keep track of mechanics and enemies, and ideas like a museum.

I also hear to make the first level last. As you will know everything to be added, that and you want to be at your best making a level when everyone gets a first impression of your work

Legend has it Mario 64 was literally Mario in a single room for the first year of development while they hammered out Mario's movement since 3D tech was rare and cumbersome at the time. Then once that was done, they added levels.

As for making levels, I highly recommend basic geometry first before you make it all pretty. It's more effort to switch around pretty.
 

ceaselessly

Member
Sounds good. I think you should now create a test level and throw everything in there to see how it all works together! :)
Can you guess what I spent the last few hours doing? :D

(Edited to get rid of the unintentionally rude eyeroll emoji.)
 
Last edited:

Yal

šŸ§ *penguin noises*
GMC Elder
Also, don't make your first level first. It's actually a good idea to make your first level last, because by then, you'll know what your entire game can offer, and you can build your first level around the idea of preparing the player for what's coming.
It's sound advice, but I'd say it's more like, "by then, you have all your cool special effects set up, so you can bombard the player with them so they get a good first impression". I find that tutorial levels that try to teach you EVERYTHING at the very start of the game tend to be pretty bad, they drag on for a long time, patronizing you all the way, and then when you gotta use the new abilities half a game later, you've forgotten how to. It's more enjoyable to have a "press forward to win" level where things explode everywhere and there's cool water and fog effects, and game journalists are gonna quit after that level anyway so you gotta make it look good.
 

Khao

Member
I find that tutorial levels that try to teach you EVERYTHING at the very start of the game tend to be pretty bad, they drag on for a long time, patronizing you all the way, and then when you gotta use the new abilities half a game later, you've forgotten how to.
That's kind of the point in a way. If you know the player is not going to need the wall jump until level 9, you know you don't need to teach it to the player in level 1. But if you know they're going to need to constantly dash and double jump and whatever through levels 2 and 3, you'll probably want to make sure they leave the first level knowing how it works.
 

Yal

šŸ§ *penguin noises*
GMC Elder
Or even better, start the level with a jump that can only be passed with a dash / double jump, but there's no penalty for failing it. The player physically can't even enter the level until they master the dash / double jump, so once they do reach the real start of the level, you know they can use it properly. (You could use a signpost or a slowly-fading-in tutorial text to tell the player what they're expected to do if it's not obvious). 2D Mario games almost always does this these days, and all the way since Super Metroid a good metroidvania forces you to use a newly acquired ability to leave the room you obtained it in (so the designers know you picked it up and know how to use it when you leave the room). Players are like goldfish, they've got really short memory and will suffocate if you don't give them air to breathe even though they look just fine in their little glass bowls.
 
Top