Or: How to salvage a project that went horribly wrong

Usually, when I do a game jam, the development process is pretty smooth. I put together a plan and execute on it, have it done early, do some testing with friends, and relax when everyone is scrambling to finish. There’s always stuff that has to be left by the wayside, but for the most part the final product looks like the original plan and I’m happy with it.

I don’t mean to brag but I pretty much have this down to a science. Or so I thought.

The development of what became In The Middle Of The Night was not like that. It was slow, poorly planned, tumultuous, infuriating, and started out with building a completely different game. It was a fucking mess, and looking back I’m sometimes amazed I got a coherent game out of it at all.

After the first Magical Girl Game Jam back in April, I kept thinking about magical girl games. In addition to ideas for a sequel to Shattered, I came up with two ideas I slammed down in a frenzy. Over the next few months I’d occasionally think about them and work on the ideas a bit. When Magical Girl Game Jam 2 came around I already had a few ideas to start from, develop around the revealed themes, and build into games.

Well, I mean, I had to pick one, I couldn’t do them all, and that’s always hard for me to do. In the days leading up to the jam and after the theme reveal I decided to play with the ideas, draw some key art and write out a synopsis to get a better feel for each concept.

I also did finally actually watch some magical girl anime! Well, I watched one episode of one show. I had more lined up, but after one episode and some TVTropes and Wikipedia articles I decided to yolo it and start working on stuff. That was probably ill-advised in retrospect but jumping into things without fully thinking them through is going to be a recurring theme here.

It’s magical girls, IN SPACE!

The plot and setting were fairly well cemented by the time I started working on it. Some time in the past, it was prophesized that a time would come where a great evil would return but there would be no magical girls left to fight it. One was sent forward in time…

Fast forward to a space opera future where humanity has just emerged from a devastating war. Tyra Kemppainen (at first a Marine, later Naval Special Forces) is investigating a colony that has mysteriously gone dark when she is attacked by a strange apparition, saved by a girl with magical powers who, dying, passes them to Tyra. She must learn to use her new powers and fight back the evil which has emerged and is now destroying what is left of humanity…

I think someone in the comments drew comparison with Pretty Cure Star Twinkle when I started talking about my concept and, yeah, pretty far off base there. It would have taken more inspiration from Halo, Mass Effect, and The Expanse.

Originally, it was just going to be one magical girl sent into the future, and she was going to either die and remain as a spirit or remain alive but lose her powers, acting as more of an advisor, source of exposition, and character foil and not an ally in combat. But at some point I figured more magical girls is better, upped the count to four with three retaining their powers and structuring the game around finding all the magical girls before confronting the final boss together.

The planned gameplay would have been very similar to what I eventually built, for reasons I’ll get into later. Lots of Halo influence, some Doom and Quake, some from gun-and-magic mods and retro-style shooters. This is actually the concept I ended up picking, at least initially. I still think it’s pretty cool. But holy shit did I bite off more than I could chew.

It’s magical girls, but set in the 1930s! To be honest it was always kind of vaguely defined, I started with the setting, then the characters, with the plot coming after. Something something fighting evil monsters from some alternate dimension. Maybe a section in Chicago and one in Paris. Possibly academies. Not much more detail than that. I’ve toyed with doing period pieces before and I thought it would be interesting and fun to do it again.

I never got to the point of figuring out what the gameplay would be for this one. After throwing around the concept a bit, I was still really eh on the whole thing, so I decided fairly quickly to go with something else. However, the character designs I came up with for this would be recycled into influence the designs for The Last Magical Girl when I decided to add more magical girls.

I’ve been using a lot of songs and albums for codenames lately. Both codenames (The Howling and Our Solemn Hour) are Within Temptation songs. The overall codename, April Rain, comes from Delain’s second album. Their first was Lucidity, which was the codename for my first Magical Girl Game Jam game.

I almost did Shattered 2 for this jam, and up until about a week before the theme reveal that was my firm plan. But the jam organizer and most of the participants seemed to want real magical girl games, and Shattered 2 would have been really pushing it. I played with the plot and setting trying to get it closer to proper mahou shoujo (bear in mind my understanding of such is still kinda shit) and came up with something I was kind of happy with.

To be clear, I didn’t think it would get removed or anything like that, but I didn’t think it would do well either. Mostly I wasn’t happy because it felt like a forced compromise- trying to bodge a concept into something it really isn’t, a square peg round hole situation.

By the way, Shattered 2 also follows a similar pattern for codenames. It’s Mother Earth, from the album  the song The Promise (which it’s somewhat inspired by) is on. Shattered 3 is Will To Power, and Shattered 4 is Apex. You can take as little or as much from that as you want ;)

It’s hard to say whether Shattered 2 would have been a better choice in retrospect. In terms of scope it would have been smaller than The Last Magical Girl, and reusing assets from the first Shattered was a possibility, but it would have still been a significantly larger game than what I ultimately put out. And of course there was the whole compromise aspect to it which likely would have made a confused end product.

I struggled a lot trying to choose between The Last Magical Girl and Shattered 2. I think if I had been happier with my plan for Shattered 2 I would have gone that direction, but as it stood I was kind of hesitant on it. I figured The Last Magical Girl had more potential, but had some reservations on that, too. At one point I said “okay, if I think I can make Shattered 2 work I’ll just go for it”, but again I wasn’t sure about it so I reneged on that. At another point I almost just flipped a coin.

I’m not sure why I ended up going with The Last Magical Girl, but it was an impulsive, spur of the moment decision. Once I’d decided, I immediately started working on it, before even making an outline, so I couldn’t go back.

It was I think a week or two in when I started to second-guess myself. I finished the outline and realized what I had was a truly massive game. It called for something like 5 levels, all with massive set pieces, about a dozen characters all needing animation frames, long visual novel style interludes between each level and full-on CGI cutscenes.

Meanwhile, actual progress on the game was a lot slower than I thought it would be. It took me the better part of a week to do the layout for one level, which still needed scripting, and I hardly had any of the sprite/2D assets done. At this rate it would never be finished on time.

So I did what I normally do, which is massive requirements scrubbing. Cut everything you don’t need, then cut some more, until it’s manageable. I reworked and rewrote the outline, but I was only able to shorten it a little bit. At the same time the more I took out the less happy I was. Sure, I could do a complete game, but it would be very thin, weak, and missing a lot of the stuff that I really wanted to get in. I also felt it would still be a lot more work than I really wanted to put in.

I think I made a few mistakes here. One was starting with a such a big plan in the first place. Usually I’m more careful to plan with an appropriate project scope, but I think the longer jam period lulled me into a false sense of security and I let my imagination run wild. I also didn’t really design the game in a way that could be easily scaled down- for the story to work, it pretty much needed every level and every character. Tearing anything out would totally trash the flow and pacing of the game.

The other was assuming I’d be progressing as fast as I did when making Shattered, a false assumption because of much different conditions. I was a lot busier in real life this time, and even if I wasn’t I don’t think I could have sustained the kind of jam/crunch pace I’ve done for shorter jams for a month and a half. I also had a lot of trouble with motivation, maybe because of a lack of urgency from the longer jam period, maybe because of other factors. Long story short, I was moving a lot slower this time.

A few people suggested I do a demo, but I wasn’t willing to go this route for a few reasons. I hate leaving things unfinished, especially if they’ve already been released or announced. I guess it’s kind of a personality thing. I also felt I wouldn’t ever actually get back to and complete the game because of time constraints and other things getting in the way. Meaning it would just sit there, taunting myself and everyone else…

While some of the people in the chat are endless fonts of work ethic and motivation, I often struggle by. The Discord history is rich with my rapidly hammered words of utter panic.

I actually kept plugging away at The Last Magical Girl a few days longer, pondering my options. I could hack it down and try to make it work, but I didn’t like that option as I’ve said before. I could do a demo, but it would remain a demo forever. I could switch back to Shattered 2 or another concept, but that would mean starting over from the beginning.

I wasn’t willing to throw away what I had (because that’s something that drives me nuts too), so I ended up designing a new game around what was already built. In fact, Suckerpunch is quite literally built upon The Last Magical Girl- I just copied the project folder and there are still many references to AprilRain (the original codename) in the project. This is why it’s such a weird game mechanically- I’d already mostly built the Halo-ish gunplay and gun-and-magic combo stuff at this point. I don’t think I redesigned the combat at all- there was some stuff that didn’t get implemented, and some that did after the split, but it’s mostly all plans for The Last Magical Girl.

Oh, and of course, some designs were reused. Sakura was modified a bit but wasn’t even renamed. The Shades had not been drawn yet but I went with what I pretty much had planned for The Last Magical Girl. And Gina- especially in her magical girl form- bears some resemblance to Tyra, the protagonist of The Last Magical Girl.

I came up with a fairly cliched story based around my limited understanding of magical girl plots, a heavy does of 2020 zeitgeist, and what I knew I could fit into a limited scope. Main character is in a funk (which may or may not be actual depression), magical girl shows up to show her an amazing world and bring light to her life, or something like that. It’s definitely influenced by some of the weirder games of the first jam.

I originally intended to make it a lot more self-aware and lean on the fourth wall about its status as a salvage game but I ended up toning this down, for better or for worse. I did make the protagonist, Gina, a bit of an outsider to the genre just as I am and couldn’t resist having her throw shade here and there (although usually at the player’s discretion).

I gave the new game the codename Suckerpunch. No, it has nothing to do with that movie, it’s named for the Delain song, mostly because the story reminded me of the lyric sucker punch the demons from my dreams. Or something like that.

It would be a stretch to call the development entirely smooth, but I’d deliberately underscoped it and the relatively limited scope of the game along with the large amount of time remaining made it a lot less stressful to work on than The Last Magical Girl had been. I definitely had days when I just didn’t want to work on it or didn’t have time to work on it, but I could afford to have those days.

To give you an idea of how much smaller the scope really was, I went from having 5-6 levels with big set pieces to 1 level with some scripting. Each of those levels would have had an interlude, probably similar in length to the bedroom intro. There would have been 5 main characters, plus supporting cast and NPCs, instead of two characters (only one with world sprites) and two enemies. And of course the CGI cutscenes were dropped entirely.

Because I only had one big level- mostly recycled from The Last Magical Girl- I could afford to spend more time on details, which is something I really wanted to do but figured was not possible with the original huge scope. The bedroom especially is small, but full of details and interactivity, which was a lot of fun to build.

I wish I could have used the Colony level I’d mostly built for The Last Magical Girl, but I couldn’t find a way to fit it in. It’s still in the game (Warp M0_Colony) but serves no purpose.

Sakura has side facing frames (albeit only one set that is mirrored) and it’s actually the first time I drew anything other than front or (rarely) back facing sprites. The side sprites actually weren’t too bad, and felt like a huge hurdle overcome.

The name In The Middle Of The Night comes from a Within Temptation song, though it’s really just because that’s when this game takes place.

I decided to go with a slightly different conversation style than last time. In Shattered (and some of my other games) the conversations are in the style of Western RPGs where you pick an option and immediately get a response from another character. This time, it’s more visual novel like, where you pick an option and then get a screen of your character speaking. I like how it turned out, but it was definitely more work and easier to screw up than the other style.

I like to think the story has a proper beginning, middle, and end, following the pattern of the classic hero’s journey, but it really is a short journey with an ambiguous cliffhanger of an ending. To be honest, I thought I had a longer game and was actually surprised by how brief an experience it was when I playtested it in full for the first time. I’d deliberately kept the scope small, but I wish I’d left some room to extend it. The way the story flows and the game is built there’s really no way to pad it out, at least not that I could think of at the time.

So, it is a short and sweet experience, much more polished than Shattered in most respects, but I wish it had turned out a little longer.

There’s still some other stuff I would have liked to add and could have added. Better weapon sprites, especially for the assault rifle. A one-button kick attack. Sprites for casting magic. Character portraits with different expressions. A nicer transformation sequence. In the end I was tired and really needed to move on to other things, so I decided to call it a few weeks before the end of the jam and release the finished and mostly polished game I had.

Unfortunately, this is pretty much the end of the road. I don’t see myself doing a sequel to In The Middle Of The Night, nor am I planning any more updates. I really hate creating another dead-end IP when I already have so many, but I only have so much time. There are more stories to be told in this universe, but the same can be said for Ascension, Shattered, or Firefighter, all of which are much more developed.

Shattered 2 is likely to be my next game, though there are a few other possibilities I’m not ruling out. Most likely it’ll be announced on my Twitter or my website.

I think that about covers it, though I might think of something and have to publish an addendum again. I hope you enjoyed In The Middle Of The Night (if you haven’t played it, go play it!) and I hope you’re looking forward to what’s coming next!