Method and exciting perspective

This is a post-mortem for my ludumdare 31 entry


Getting ready to dare: Setting a personal objective

No wonder that every 4 month I -among many- keep getting back to the jam : The Ludumdare is a fantastic opportunity to learn and improve while having fun, and the community is always ready to give back (I tried to do it as well).
For every LD I participated in, I tried to focus on a specific topic, be it making my first game, trying specific concepts, experimenting with procedural generation, improving my square-only graphic style, delivering an experience more than a game…
This time my personal challenge was a bit different….

You see I noticed a strange pattern in my projects: In the past 3 months I have been working on a personal project… and I have nothing to show. I produced design documents, assets, code (a lot of code actually), but I have nothing to show. Because what I really did is pieces of a jigsaw puzzle that do not fit together.
So what am I doing wrong here? Why do I get lost in refactors, doing-only-the-cool-stuff -and a bit of over-engeneering- while not suffering from those problems in my LD compos?

During a LD I work in a very specific way: I have a sheet of paper which always has a list looking more or less like this

  • feature 1
  • feature 2
  • levels
  • feature 3
  • feature 4
  • levels
  • assets
  • title screen / game over screen / level select screen / option screen
  • levels
  • fancy feature 1
  • fancy feature 2
  • levels
  • etc…

Do you notice a pattern? I have milestones in this list (usually were I write “make 1-2 levels”) which means that as long as I reached it I will always have a playable game, no matter how far I can get down the list. I also have a clear prioritization of the features and I don’t lose time on assets while I don’t have something playable.

This simple list, which I never used outside a LD compo, brings me focus and ensure that I have a playable game to deliver. I never actually stopped to think about it until 1 week before the compo, but I really think this is a strong foundation for a method I could use on bigger projects.

So I had my goal for this compo : evaluate this method. I was ready to enter the jam…


It begins… Theme annoncement

If you are reading this you and I probably have something in common : tons of ideas. Not enough time.
Like some of you, I have attempted a LD knowing which game I would make before the theme was even announced. I have to admit that the result was always disappointing
While this is not the purpose of this post, I will take the opportunity to share a small trick with you: I keep a record of all my game ideas. I always have a notebook with me, and I write any game idea (or mechanic idea) I may have. Right before the theme announcement I read through it. With a bit of luck, one of those ideas will immediately come to mind when I see the theme.

I have used this trick twice now, and each time I knew within 5 minutes what game I would make. I also think those two games are my best, probably because those ideas were both well-though and matched the theme.

So here was my basic idea, as written in the notebook :

the player controls various weapons with a command-line and the situation is displayed on a very simple console display. (spaceship, or command quarters)

A perfect match with the theme.


The original idea first came to me while reading a novel by R.Heinlein, “Citizen of the galaxy”:

 (yes, i am french)

At some point in the story, the hero works as the fire-controlman of a spaceship. His job is to assist a computer with the complicated calculation required to hit long distance targets. While the concept of a human calculating stuff more quickly than a computer is very… old sci-fi… I found it really cool.


Let’s dive into it

First thing first: I quickly put together a list of features, prioritized them, set clear milestones. I came up with something like this:


I don’t have a screenshot of the original version, this is what it looked like after the compo


  1. On the left, I have the actual list of milestones. It is not explicitly stated, but after each milestone I intended to have a playable level.
  2. Then for each milestone I have a list of features I want to include.
  3. My color code is pretty simple :
    * red = blocking something else
    * yellow = good enough
    * green = does exactly what I want
    * purple = bugged

I put any asset needed in there as soon as I encounter it. I do the same for the bugs : If I don’t intend to fix it right now, I put it on the list. So the 3 lists on the right side a filled while I dev.
You can see I also use the color code in the asset list : A yellow asset is a placeholder or something I don’t like.

Choosing an idea and setting up those lists took me 20 minutes, no more. I had clear objectives, a step-by-step list of things to do, and fallback plans in case something took too much time (and it did). How awesome is that?


So? How did it go?

Really well. I never got lost developing unnecessary stuff, had time to eat and sleep, had early versions of the game, everything worked as intended really quickly, didn’t have any major bugs…

… except for a side effect I hadn’t anticipated.
Until 3 hours before the deadline I had nothing more than this screenshot:


Yes I had a bunch of functional mechanics, but the (little) content that I had was hard-coded and was not enough to make an enjoyable experience. So what I did was a feature-lock. I dropped what I was working on, and I began to produce content only.

The result: A single mission. Everything works as intended but the game has only one mission (a tutorial actually, I couldn’t expect the player to guess which commands he was supposed to type)


Is the method working?

Yes. I don’t have any doubt about it.
While I didn’t expect it, this feature-lock situation I got into is actually pretty cool: I had very little time and had to rush the game but thanks to that method I didn’t have any last-minute-ugly-and-shameful-patch to do. All I had to do was producing content (levels). And that could not break the game.

Of course this is a first draft of the method. I can already see some problems with it:
– the lists are inaccurate (you can see I left some entries untouched in the early milestones
– the lists are too basic: I actually had a more short-term TODO list on paper (especially for the command-line input, which had sub-features like auto-completion)
– dependencies between features are not visible
– the color code is very simple and was not always accurate
– I am not sure how this could work for a team.
– feedbacks (and requirements for a feature, and design-stuff, and….) are not well-integrated
– …
So I will keep working on it (and will probably write about the results). I already did some research and found something pretty similar, and much more detailed.


What about the game?

As a result of the method, even if there is only one mission, the game feels very complete and don’t have any broken feature.
But it is not the game I wanted to make.

If you re-read the idea that was in my notebook, you can see it was all about weapons:

the player controls various weapons with a command-line and the situation is displayed on a very simple console display. (spaceship, or command quarters)

Instead I have scanners, sensors and communications… far from the “fire-controlman” I wanted the player to be. To sum it up, I ended up with a simulation of… whatever her job is:



Anastasia Dualla simulator 2014

Yes, I was a little disappointed in the game itself.

And yet, it works.

Since I didn’t have all the features I was hoping for, I had to be a bit creative with the way you could use commands. And the result surprised me in many ways:
– I ended up with mechanics far from what I wanted, but I was still excited about it during the dev (and still am a month later)
– I expected the game to be fast-paced and all about quick reactions (basically this kind of game, but better). Instead, I ended up with a really good concept for a narrative game.
– I thought a game about scanning ships would be dull and boring. It turns out that the play-testers (family and friends) were really liking it. More surprisingly, people who I though would quickly-test-it-and-tell-me-i-did-a-good-job-to-get-rid-of-me were the most enthusiast. When someone tells you “You know… of all the game you made, I think this is my favorite” several days later, it really brighten your day.


What if…

The more I think about it, the more I believe the punch-line I came up with (“Paper Please meets Battlestar Galactica”) hits home. I have a really good material to make a good sci-fi game. And by that I don’t mean a game with heroic moments, laser shooting, planets exploding, teleportation and fancy stuff like this… I mean that I have a way to deliver a “real” sci-fi story (the ones that present a setting and make you think about it) with mechanics that serve it. I could totally make a full game out of it…

Not so fast!
Having good feedbacks from your family and friends is not enough, even if you really believe it could work. Let’s not jump to conclusion and collect some real data:

I put together a post-compo version of my game that I updated multiple time. In it I…
– Added one mission which is not a basic tutorial
– Added a lot of common-0 sounds (I couldn’t do that in the context of the compo, but the difference is incredible)
– Reworked (very quickly) the GUI to tackle the most common feedbacks I got
– Fixed minor bugs

Those change are there to give a clearer view of what the game could be. Below I added a poll to ask people if they would be interested in the game and a field so people could leave me an email address to be notified if I actually decide to make a commercial version of it.

Here are the results:

  1. ld31-pollEven if I consider any played-but-didn’t-vote as a “no”, I have a positive ratio
  2. I was not expecting much of it, but people actually left me their mail! (no screenshot for that one, obviously).
  3. I have more plays on the post-compo version than I have votes on the LD site.
  4. Some people (not coming from the ludumdare site) contacted me to tell me I had a really good idea
  5. I was told from several sources (once again, not only on the ludumdare site) that I should make a full version of it.

So should I make it?
– I have an Idea I believe in, and that other people like.
– I have a method that will help me to stay focused and to respect deadlines
– I have time, and money (I didn’t talk about that? Well I do)

So yes, I think I should. It may not have the same name, the same context or the same look, but it will probably happen. Stay tuned!