This is the first of three short dev blogs I'll be doing as a retrospective of my experience creating the demo of my Ace Attorney-inspired murder mystery for the Otome Game Jam in 2023. My intention is that this should hopefully be useful for other people who might do game jams in the future... or entertaining for those who like to see other people make terrible mistakes. Or both! It's not a binary.

Part One: The Best-Laid Plans

All right, so first a little background on why I wanted to do this. I've been poking around the idea of doing a game for a few years, but I have a terrible habit: I'm never fully satisfied with things. I'll make a sprite and fuss with it, make a background and then change my idea for the style halfway through and make it again, that sort of thing.

So! As a kind of replacement for actually working on my flaws as a person, I thought a jam would be just the ticket. A meaningful goal, a set time limit, and a supportive community.

I was right about all of those things. Unfortunately, I was wrong about a lot of other ones.

MISTAKE #1 - Following the rules (even when I made them up)

The rules of Otome Jam 2023 were clear and really pretty lenient. Aside from it actually needing to have otome elements and being tagged appropriately, there was no hard and fast requirement that you only start working on it on the first day. In fact, they allow you to submit already open projects as long as substantial work is done during the jam period.

So why, oh why, didn't I start earlier? Still baffles me. I was determined only to do outline planning and character sketches and not start any of the actual assets until the first day. Again, why? These are the kinds of things I will be thinking about at 2AM probably until the day that I die.

So if you're in the same situation, learn from my terrible example: do what makes your life easier while still remaining within the rules of the jam.

MISTAKE #2 - A simple murder

Ah ha ha. So this is a bad one. I originally thought of doing a murder mystery because I love them, but also because I thought I was being so clever limiting it to a single place and small cast of characters. I knew I wanted the main character to be a detective, I knew I wanted the game (if possible) to include an evidence gathering system like Danganronpa or Ace Attorney, and I knew I wanted the love interests free to be either murdered or murderers.

I fixed pretty quickly on the main concepts: 4-5 love interests, each one with a separate murder and murderer. The fifth love interest ended up being ditched - I had the idea of a kind of himbo Watson sidekick for the MC, but it ended up feeling like that would cut into the opportunities for the MC to actually spend time with the LIs. So farewell, himbo Watson.

Despite my general failures as detailed above, I tend to be a fairly organized person. So I worked out the math! Four routes over the course of the game meant at least four murderers and four victims. If I didn't want them to repeat, or for it to be obvious who they were, I estimated I needed 12 characters in total, or 7 in addition to the main LIs and MC.

Below you can see the first spreadsheet I did:

I separated things loosely into a few different categories:

Art

• Sprite art
  • Expressions
• CG art
  • Variants
• Background art

  
  

  • Variants
• GUI art and associated elements
• Intro video (Didn't ever really think I could get to this one, which indeed turned out to be true, thus the low time allocated to it.)

Writing

• Common routes
• LI routes 1-4
• Bad ends 
• Coding (Calculated based on the amount of words in the main writing group above. This multiplier was low because I was planning to write directly in Renpy code, so the time allocated was mainly for correcting it.)

Music

• Background sounds
• List of potential music needed by mood

I based the idea on what would be necessary for the full game, and then lowered the estimates for the demo until it seemed feasible within the time frame. I think the concepts were fine, but the estimates turned out to be incredibly off-base.

For starters, I was estimating that because I would only get into two or three of the routes, I'd only need sprites for the four LIs and four other characters... until I realized that would pretty much give away who the important characters were for each route. And therefore I ended up needing all of these jerks to have sprites as well as the main characters:

Curse them and their fiddly outfits. Which leads into the third major planning mistake I made...

MISTAKE #3 - Self-knowledge and the lack thereof

Even typing on this topic makes me want to lift my hands up to heaven and shake them like a despairing peasant in a Monty Python movie. Because wow, this was a mistake I could have seen coming at this stage.

If you noticed from the spreadsheet, I was planning to write 50K words in just under 10 days. How did I get that number? Practice? Math?

...no, of course not. That is entirely based on the fact that years ago I did a challenge to write a 100K novel within a week. But as you might suspect, a few things were different this time around.

1. You can probably tell my writing style leans towards the verbose. Great for a novel where you can wibble on for 1,000 words about the setting of a scene, but this game was 90% dialogue. And dialogue can be hard to write, particularly when you're working out the character's voice as you go.
2. I wasn't just writing things out in text, I was writing them in Renpy code. I thought that'd make it easier later on, but it slowed down the actual writing quite a bit.
3. This is the big one. I didn't have a job back then! I wrote something like eight hours a day! And now, of course, I do have a job, and if I tried to write eight hours a day they would no doubt notice and have opinions. How could I forget that? Argh.

So of course that went much slower during the actual process.

Luckily, it wasn't all bad. There was..

TRIUMPH #1 - Good company

Not mentioned in my previous list of grievances is that I'd pretty much decided to do this solo. I had a few different reasons for this, ranging from not wanting to pull anyone into something if I turned out not to be able to do it to the horror stories I'd heard about people fading out during game jams.

If I did everything myself, I figured I'd be the only one who I had to worry about delivering. (Note: it is an equally horrible feeling to be the one who is actually disappointing yourself! I was not right about this one. I suspect there would have been different stresses involved if I had secured a team, but doing it yourself is in no way the stress-free option.)

Still, I decided to participate in the call for people who might want to be involved on the fabulous DevTalk discord. I thought that if someone might be interested in the concept as outlined in my strange summary, that'd be better than approaching people blindly who might see my very niche concept and be appalled.

I actually had a few people approach me from this, but only one would become deeply involved in the project. That one was Shar,  who ended up being not only reading all of my hastily written scripts, but also providing fantastic and hilarious commentary as they did so.

Some of my favourite quotes from these comments:

• "I'M SMITTEN PLEASE BE EVIL..."
• "Not him getting dragged by a dead man..."
• "Build your harem, queen!!!" (Note: Shar is not referring to the MC there.)

And that was a triumph, because not only did I discover about a thousand differences between British word usage and the Canadian equivalent that I had somehow missed in my lifetime of mystery reading and have my scripts tremendously improved, but having someone else involved in the whole creative process would turn out to be both amazing and motivational.

TRIUMPH #2 - Planning to plan to plan

As already mentioned, I was determined to be a stickler about not creating anything before the jam officially started. But, as was also mentioned, I'm a planner. So I spent the time before the jam doing a few things that would really help me out later in the process:

• Searching for backup assets.
  • I knew that there was a chance I'd have to fall back on royalty-free music or other assets, and an equal chance that I would decide to do so far enough along that figuring out how to find and use them would eat up time I didn't have. Ditto for researching outfits and background options.
  • So I created a folder of bookmarks with useful sites for the jam, called 'Game Jam Prep". The folders look something like this:
    • Characters
      • 1930's fashion
    • Backgrounds
      • Manor house
      • Manor house plans
    • Music examples
    • Game jam articles
    • A link to all seasons of Hercule Poirot on a streaming site
    • Etc, etc, etc
• Reading previous game jam retrospectives.
  • I was intensely nervous about the whole process, so I searched everywhere for things people had written on doing game jams. Those would have been more helpful if I'd actually applied more of the things I learned there, but it was definitely useful in settling my nerves. 
  • The biggest thing to come out of this for me was creating several fallback strategies in advance. I.e., I would have liked to get through the first part of all three unlocked routes, but just in case I could not I had the backup plan of only doing two of the three. I had similar plans for the art aspects of things. This prevented me from panicking and let me start the jam knowing exactly what I wanted to prioritize.
• Investigating automation possibilities
  • I knew that I'd be doing almost everything myself, which meant that any time I could save on one aspect of something was time I could apply somewhere else useful. So I looked for ways I could speed things up.
  • I'll go into this more in the second part when I actually applied this, but the main things I found or made myself here were:
    • Solutions for auto-exporting sprite parts with the correct names and positions
    • A quick script to copy settings from one character for all other characters (useful when I kept changing character expressions later on)
    • Another quick script to check that all variables used in my Renpy script had been declared so that I didn't have to worry about that too much in testing

So here's my main takeaways from the stage before things kicked off, both for myself and for anyone else doing something similar:

1. Don't make life harder for yourself by inventing new rules to follow.
2. Even when you're trying to keep things simple, scope creep will get you. In retrospect, I should have just made the four LIs and had them take turns killing each other off.
3. Planning is great, but it works a lot better when you make sure you have an accurate - and recent! - idea of your capabilities.
4. Have at least one person to show things to and contribute insights as you go, since that way you don't have to force the people in your personal life to play visual novels against their will. They, er, don't like it.

Join me next week for part two, when all my bad decisions come back to haunt me. But in a fun way!

And here's a bonus pic of Nick from my twitter, just to top things off - I made it to say thank you to all the people who've so kindly downloaded the demo or commented here on itch.io, so it belongs here as well. ^_^