Indie game development is such a funny thing. I can't claim to know how the process of creating a big budget video game works, but I'd struggle to imagine how one could be made with so many hands involved without a rigid list of objectives and tasks. Indie teams work almost in opposition to that work ethic, flying by the seat of their pants, and that can lead to a product that hardly resembles where it began. There are benefits to both sides, and I'm more than satisfied to be in this field, but there are limitations as well.

In my opinion, being driven by feeling and being your own boss go hand-in-hand with creating some of the greatest and most expressive art in the world. They are also synonymous with projects being shaky, a lack of a quota leading to volatile progress as new ideas lead to changes in direction. The results are always a labor of love, a creation only guided by passion, but one usually marking the end of a very erratic road.

All of this isn't to say that these paths are always turbulent, or that such turbulence can only be painful. My own experiences with the development of Battle Craze were chaotic, but rarely were they ever not fun. The twists and turns are deliberate, with new ideas striking us far after we began to follow the initial plans we had jotted down. We are afforded to choose when we want to extend our development, often to capitalize on opportunities we only decided we didn't want to miss way down the line.

We planned to develop 20 playable characters. After a lot of spontaneous ideas over the course of several years, we ended up with 72.

One amusing facet of a game designed by Janis (lead artist and designer of, well, practically everything) and myself (the coder, in charge of game balance and designing some movesets) is the number of times we have come to the same conclusion, but through different inspirations. The "Craze Gauge" and "Craze Cancels" were birthed during a discussion in which she threw out the idea of Marvel vs Capcom 3's X-Factor, but without the X-Factor. I told her it reminded me of an idea I had for another project which was essentially Street Fighter V's V-Trigger, minus the V-Trigger. The end result is really just All-Purpose Cancelling similar to Guilty Gear's Roman Cancels, but the trajectory of that idea was far more jagged than one would assume.

Another feature that cropped up in the same vein is "Side-L."

Our characters were all originally designed with one movelist, the one that would eventually become Side-C. At some point, while working on our sixth character and final character before the first beta release, La Justicia, we had started spitballing ideas for an alternate set of moves for each character. The inspiration was pretty natural for us; I was pulling from memories of the unlockable "EX" movesets available in Guilty Gear XX #Reload, while Janis was remembering a similar feature in ArcSys' other big 2D fighter, BlazBlue. It could be likened to quite a lot among fighting games, from Super Smash Bros Ultimate's Echo Fighters, to the retro Clone Character of old, to the Moon Styles available in Melty Blood.

Any way you slice it, the goal remained the same: create unique-feeling playstyles for existing characters, reusing as many assets as possible to minimize our workload. We had ambitions to create 20 unique characters, so designing new functions for each with mostly their existing visuals was no small task, but it was one we were convinced would pay off. We bounced proposals off of each other, eventually introducing fun ideas and gags ontop of gimmicks that we had yet to explore with our roster as written. Even a silly prospect for the future, an evil variant of the quirky Sherm, was added as a Side-L variant.

I volunteered myself to be responsible for filling out the Side-L alternatives where we hadn't yet planned anything. My initial inclination was really to just try to work in the also-recent Craze Gauge addition into new traits; an act that was not fruitless, but haphazard in the grand scheme of things. At that time, we had very little idea of how the gauge would control the flow of battle, so its importance was ethereal to us.
My proposal for the game's protagonist and most beginner-friendly character, Ajit, was the addition of enhanced special moves. You see this in all sorts of forms, from Street Fighter to Mortal Kombat, and more recently in King of Fighters. This would, however, be tied specifically to the new Craze Gauge, rather than the usual "Meter" that such moves usually associate with. Ajit would function similarly in Side-L as he does in Side-C, but trading in his powerful counter Craze Technique for these metered special enhancements. Additionally, he could charge up his Gauge faster through his new Craze Technique, and gained a similar damage boost through usage of said resource.

I got cold feet about this decision right before the first beta release, however. Looking back over the options one Side had over the other, it was starting to feel like I had made a mistake. Side-C had a quick and strong counter, one that buffed his damage for the duration of the round if used well. However, Side-L had double the special moves thanks to his empowered attacks, could charge up his Craze Gauge on a dime, and gained the same damage boost by simply using the resources he was building. In retrospect, it ended up being a lot more fair than we expected, but simple math made us believe that Side-L was much too strong. Thus, we pulled back.

In the first public release build, Side-L Ajit had no EX moves. It was an eleventh hour decision, but not a difficult one to enact. The ability to access those attacks was simply removed, and any mention of them was scrubbed off the command list. I tend not to clean up code when it goes unused, however; I quite like the little reminders, the artifacts left behind from ideas long abandoned. So, even though EX moves weren't available to use, they remained intact, just waiting to be dredged up.

After the game went public, the next on our character roadmap was Ikue, in addition to filling out some Side-L slots that were yet to be completed. I remember a lot of this coincided with Ikue's development, as at least a few of the details involve her just by happenstance. One was the proposal for Side-L Ikue, another hastily-crafted kit that made use of the Craze Gauge in distinct ways. The plan was to allow her to enter an empowered form, which we affectionally refer to as "an Install," named after Sol Badguy's Dragon Install. This would briefly grant her a window in which she would be overall stronger, with plans to incorporate something akin to Street Fighter Alpha's V-ism to justify the weighty cost.

Amusingly, the original Ikue, Side-C, also featured EX specials, but they were more traditionally tied to the Adrenaline Meter.

But, with new developments came more and more ideas. The move from MUGEN to IKEMEN as our engine gave us the tools to switch between Sides on the character select screen, which was a bit closer to how we ideally wanted it to be handled. As I was implementing said features, a thought kept coming back to me as I remembered all the unused Side-L Ajit content. It was all fully functional, just deemed a bit too strong. There wasn't a reason we couldn't use it in some way.

I made a third slot for Ajit, forcing the player to hold a specific input to use this secret new Side. Unfortunately, the limitations of this system made it less of a surprise to my fellow devs, as picking 'Random' had the chance to fall on this otherwise hidden character. Given the roster at that point, there couldn't have been more than 12 outcomes, so it wasn't too uncommon to see him pop up.

The joke was that this character was the original proposal of Side-L Ajit, EX moves and all. However, given that it was simply a stronger variant of a character, it had be stapled to a significant drawback. If the additional options were deemed powerful enough, the movelist could be seen as more viable, and I wasn't excited about the idea that a gag character could overpower his balanced counterpart. I made certain that wouldn't happen when I slapped on a reduction of his base health, which was a whopping 50% drop.

The name Side-K came pretty naturally. Side-C was the purest form of a Battle Craze character, thus 'C' for Craze. A lot of my own takes on existing Battle Craze assets were labelled with 'L' for 'Lucid,' a work name for the story Teri was written for, thus Side-L. This newest Ajit variation was just a hastily-made joke, so I thought the name should reflect that: 'K,' for 'Kusoge,' a term that gets thrown around in the fighting game community for janky but fun games. I'm pretty sure at this point, I had vowed to make any other joke sides into other unique letters, but that never came to pass.

A quirk of this secret method of character select is how it automatically chose the character, rather than simply place the cursor over their invisible icon. At first, this caused a bit of an issue when it spontaneously crashed the game upon entering a match with Side-K Ajit. He worked in testing, but the character select seemed to be causing problems somehow. It quickly became apparent why.

MUGEN, IKEMEN, and fighting games before a certain point almost always chose your character's color palette by what button you pressed when you selected them. Battle Craze adhered to this rule out of necessity, given the alternatives were clunky and unappealing to us. When you use one of the 'non-face' buttons to select your character in IKEMEN, however, the engine gets confused because no palette is picked, and it just crashes. This normally wouldn't be possible, but if you set a hidden slot to be selected via any of these buttons, the game tries to push ahead without having a palette chosen.

The solution was to just put a face button, I.E. Light Kick, into the command to select the secret character. IKEMEN reads the command for Side-K Ajit's slot, and also registers the LK press as Ajit's default palette, solving the problem. Unfortunately, I'm quite a stickler for options, and I never use a character's default look if I'm given the choice. It also would've been weird, in my eye, for any usable Ajit palette to be inexplicably tied to this new Side-K playstyle.

So I did what anyone would do: I just made a weird, in-human palette out of the blue.

A fun piece of trivia about our character development is that everyone starts as a copy of someone else, usually Ajit. This has the effect of another character's sprites, more than likely for when they're being hit, showing up on in-progress characters. Given palettes are indexed, meaning that every color corresponds to a numbered slot, you'll end up seeing a lot of versions of Ajit with someone else's colors mixed up over his.

Sometimes, this erroneous application of colors leads to visuals that aren't too shabby looking. I remember thinking fondly of Ikue's 8th palette (holding Taunt and pressing SK when selecting her) on Ajit, in spite of the rather unfortunate trade of skin tone. We hadn't really needed a new palette for Ajit before this, but I saw a golden opportunity here. With a bit of tweaking, I created the color for Side-K Ajit, while taking the color of his arms and applying that as his overall skin tone. Of all things, this decision was inspired by the weirdly inhuman palettes of the fun and fascinating Jackie Chan in Fists of Fire.

The abnormal nature of the new look for Side-K Ajit began to birth a sort of strange lore among us in the development server. It seemed to simply take a life of its own as we started contemplating ideas that the unnatural Ajit was a sort of in-universe clone, a strange non-human that copied his skills but at the cost of being as brittle as tissue paper.

And then, Janis dropped this on us out of nowhere.

Enter Kahl D. Stone, a charactered explanation of the mysterious Side-K Ajit. His design stemed from the palette derived from Ikue's 8th color, his name begins with a 'K' because of Side-K, and he's a rebirthed idea from a previously scrapped concept. He is an odd mad scientist, enacting strange experiments that involves molding golems out of clay. In a flash, what started as an excuse to use cut assets led to an extension of the game's lore. It also began what would push the roster's size even further.

What happened next was potentially obvious as we started to contemplate what other Side-K characters could end up being like. Even if our first instance of this kind of character involved very little in the field of new assets, we had a lot of fun pitching random ideas for gag versions of existing characters. Eventually, we started writing them down, with the intent to only do them if I found time between character spriting cycles.

Incidentally, Side-L Ikue didn't really pan out like I'd hoped.

I ended up hesitating on giving Ikue the more unique benefits of a Gauge-driven buff state, out of fear that it may be too easy to go horribly wrong as far as balanced is concerned. That resistence caused a bit of a dud, leading us to workshop an entire overhaul of the Side before the update containing Ci-Ci landed. The value of Gauge was starting to become clearer, and the reliance on it to make a character unique while depriving them of its natural uses (the Craze Cancels) just seemed to be an undesirable trait. As such, Side-L Ikue was reborn with an entirely new playstyle, even if she maintained something of her old charging ability as a Craze Technique.

I wasn't going to just remove the old features entirely, though. Why do that?

Side-K Ikue was revived as soon as it died, simply having her old Side-L properties transfered over to the new Side. We also applied the now established Side-K color scheme to her, which, may I remind, is just one of her pre-existing palettes but tweaked to look good on Ajit. Our second Kahl clone character, and it had already come full circle.

The most tragic thing about Side-K Ikue is how she was balanced. Side-K Ajit used moves that were cut out of fear that they were too strong, so it came at a cost. Side-K Ikue was made *stronger* than the way it was released as Side-L due to using a movelist that, frankly, wasn't very good and no one actually enjoyed using.

The train was slowly inching off the rails now. The debute of Ci-Ci didn't come with the two variants expected, but instead he launched with all three, saddled with his Side-K clone right out the gate. This time it was simply a more volatile deviation of Side-L, who had to maintain a moisture meter to keep his attacks strong. They had practically the same moves, but Side-K Ci-Ci used its own health pool as a resource, and got stronger the closer it was to death. It wasn't very good. But, that was something of the point: Side-K was just a gag, easy to throw together and not meant to be competitively viable.

It didn't stay that way for long.

From this point on, Side-K wasn't a natural occurance born from thrown-out code, but something we actively enjoyed making. We contemplated all sorts of humorous ways to re-use sprites for entirely different attacks, pitching joke movelists that sounded good enough that I eventually decided to implement them. Side-K Wallace was his Side-L kit but with even more armor, and a drawback of building his Guard Gauge up when he absorbed damage that could eventually lead to a Guard Break. Side-K Sherm was a motion shotoclone, much like his inspiration, but with new sprites drawn for his Craze Technique that mimicked Street Fighter 4's Focus Attacks.

A few months later, we had released our 14th character. 10 of them had a Side-K clone.

The wheels had completely come off. If you consider each Side as its own selection, our roster was made up of more than a quarter of hidden joke characters. The strength of these fighters was more of a consideration as well, given how random select worked -- after all, no one wanted to get a useless character by chance.

More prominent was how it had affected me as a coder. Enabling a half-baked version of Ajit with a health modifier was something that took very little time. Concocting an entirely new kit was something else entirely, in that it was more or less just a repeat of what I was already doing with Side-L. We didn't want to play a bad character, nor did I want to waste time preparing new moves entirely as a gag. Effort was going into these creations. They no longer warranted being tucked away behind a cheat code. Players were already well-aware of how to access these variants anyway, so what was the point of trying to keep them a secret?

Thus, Side-S was born.

The decision was made while we were developing Side-K La Justicia: just officially add these third Sides to the roster. They would have their names changed to break the tie with Kahl and his clay amalgamations, in order to prevent one-third of our entire selection being weird clone golems. They were just yet another arsenal for the fighter to choose from, but decidedly focused on "unusual gimmicks" to make them really stand out from Side-C and Side-L.

That alone brought up an issue, one which required a solution that in turn created a surprising cascade of effects.

As mentioned, we were in the middle of making La Justicia's Side-K counterpart when this change was put into effect. Our resident grappler was, of course, a dedicated babyface in her lore, as evidenced by the title. The pitch for an evil clone was too good an opportunity to pass on: make it fight like a heel. We tapped into the essense of a wrestler the audience loves to hate, with "overselling" and taunts met with a booing crowd.

When all of Side-K became Side-S, they lost their tie to the Kahl golems, thus making them, well, them. That would obviously make Side-S La Justicia a heel, purposefully agitating spectators as part of her gimmick, which was at complete odds with her role. There were only two options from here: abandon the gag and retool the aesthetics entirely, or...

...Get creative. With a little bit of sprite editing, we were able to utilize palette trickery to cover La Justicia's face with a full mask, joined with new lore that turned her into an entirely new character. Granted, it was always intended to be La Justicia with a secret identity, playing the part of a heel as a gimmick within her wrestling federation. But, we had already transformed Sherm into an entirely new character, despite sharing a significant portion of the assets, so this was just another swing at that.

It began to get us thinking. Making these 'evil' alter-egos was relatively easy, and it was a way to implement characters who we couldn't justify taking their own slot on the roster. Evil Sherm was an idea for a unique fighter that turned into a Side when those were added, and there were a few more of those on the chopping block.

Reaper was our designated zoner. When Side-K came along, the choice to make their Kahl recreation a close-range rushdown brawler was clear. Janis already had this plan on the backburner from much earlier: an in-universe Reaper clone created by the villainous QUASARS that fought up-close and personal, inspired by the scrapped Kelvin character from Killer Instinct. When Evil Sherm and La Tirania were realized, Side-S Reaper becoming this unhinged evil twin was quickly put into play.

We were going to stop there, but I had remembered another canned clone character idea Janis had written down, and I wanted to take the opportunity this presented. She had dreamed up a Robo-Tsugumi, a play on the Guilty Gear character Robo-Ky, who was meant to be a far different beast from the boxing champion. I proposed we change Side-S Tsugumi into our fourth and final alter-ego Side. Initially seen as too divergent from the original plan, I sweetened the pitch in a way that didn't betray her idea: Robots can have different models. This Robo-Tsugumi would just be an earlier, more standard model.

From here, development continued as usual, but with thrice the gameplay consideration per fighter. Side-S would end up far disconnected from its Side-K origins. Side-S Ajit, Ikue and Ci-Ci were recreated entirely, and even established Side-S movelists were fleshed out over time. There was no longer room for wasted slots in our lineup, and we would be driven to buff and rework anything that fell below our standards. When each character was only one entity, that didn't seem like a hard task. When considering every side, the IKEMEN branch had 60 playable choices by the end of its development.

Now, Side-S is the breeding ground for any and all odd things we want to get away with. Even after IKEMEN, as we worked our transfer over to the new Game Maker engine, plans were made and executed to take recurring moves from other Sides and replace them with entirely new entities. The formerly 'complete' Side-S Wallace still used some of the same attacks from his other arsenal, but he had so many animations that went unused for this Side in particular. That sort of thing would so often make us think, 'why not have some fun with it?'

So now we do our level best to give Side-S as many distinct attacks as possible. Even in times when the functional difference is so minute -- what does a Super Headbutt do that a Super Punch doesn't? -- we want to maximize the feeling of Side-S being 'The Weirdo.'

Side-L didn't exactly stay the same throughout all this either. What initiallly started as a small shift in movelist between Side-C to Side-L became something much more evolved, somewhat naturally. At one point, Side-C Boundary was defined by creating stationary portals that he could either travel through or turn into an explosive. Side-L Boundary was just doing that, but the portal followed the opponent. With Side-S being an utter outlier who danced between different buff/debuff states ala Shulk's Monado Arts, it didn't feel right having such a small choice between the two other Sides. Side-L Boundary didn't feel like a unique character in a roster of 60.

At this point, our selection size was a point of pride. Though we were a team in the single digits, we had amassed a roster of 60 playable characters, with 12 more planned for the final product. We enjoyed creation more than anything, but we also didn't want that claim to be with caveats. You could make an infinitely enormous armada of fighters if you kept changing one character's color, and name, and then maybe swap some values around. When all was said and done, we wanted to have 72 meaningful choices for players to sink their teeth into.

As of time of writing, around 68 of these Sides have been realized, with plans for completing the collection already in place. We could have concluded with 24, and I would've been satisfied with that count, but we had more ambitions than previously accounted for. Development may have been completed before the six years this has taken, but it wouldn't have been nearly the expression that it has become.

While I can't exactly say I'm glad this has been such a long journey, I still would not trade it for a less meaningful one. A stricter plan may have cut these proceedings short, but that wouldn't have alotted us room to experiment, and to grow. We tripled our roster -- and therefore, our efforts -- because it's what we enjoyed doing. I only hope our expression finds its audience, to the people who would be amused by our colorful expression.

Oh, and we have to make sure that players scouring the roster of 72 characters for their preference is less daunting than the process of creating those characters. That's probably also important.