Tears for Spheres: Developing "Grudgeball" in the year everything changed

Jumping Growth

Three guys in a room, somewhere in Buenos Aires, Argentina. That’s what HeavyBoat looked like back in 2010, when it was born after its founders (David, Juan and Leo) decided to leave the company they worked for. As an amateur musician, I soon joined them on a freelance basis, making music and sound FX for their games. A failed attempt at creating an original IP was traumatic enough for the guys to decide we’d focus on work-for-hire from that point on. So with a little luck and a lot of work, three years later we’d done more than 20 flash and HTML5 games, mainly for Cartoon Network and Disney.

It was appropriate that “Jumping Finn”, our fairly successful Adventure Time flash game, would be the one to take the leap into the mobile world, expanding into a version called “Jumping Finn Turbo” (or “Super Jumping Finn” depending on where you live). But the scope got considerably larger in our next mobile game, based on another important Cartoon Network IP: Regular Show. “Best Park in the Universe” was released in May 2013, by far our biggest project up to then.

HeavyBoat’s steady growth meant that soon after “BPU” was released, the company moved to its fourth office in four years. Only this time, it wasn’t a simple relocation. We had reached a point in which it seemed we could simply get stuck doing the same thing forever – or at least until our clients got tired of us. So if we really wanted to keep growing, there was no small step we could really take: the only real option was to try to jump forward ourselves and go for broke.

As a result of that decision, HeavyBoat went from around 10 employees to almost 25, in about six months. The goal was to have three teams that could run in parallel, each producing a mobile game. Coincidentally, this big experiment was to be bookended by the production of our next big Regular Show game, eventually known to the world as “Grudgeball: Enter the Chaosphere”.

This is the story of Grudgeball’s development – but more than that, it’s the story of how it got trapped under the weight of the business risks HeavyBoat took during that time, our most difficult period to date. In the end, the game managed to dodge the mistakes we made along the way and became quite successful – though our big plans for the company wouldn’t turn out as well.

Deathball

We only had two things to start from: we knew we’d return to Mordecai and Rigby’s bizarre universe (if you’re not familiar with Regular Show, go watch a couple episodes now!), and that the game was expected to have “universal” gameplay, with simple control schemes. Early talks included ideas for an endless runner, a management game, a racing game, a dungeon crawler - even a “Lost Vikings”-style puzzle game. But then our guys came up with the basic “dodgeball” idea: controls would be kept simple by having characters move on their own, so players could concentrate on the most positive actions: shooting, blocking and counter-attacking. It would be somewhat reminiscent of the old Windjammers for NeoGeo arcade (hey, many of us are thirty-something, but don’t hold that against us), only this time with three-player teams, as CN was understandably interested in having a lot of character presence in the game.

The recent expansion of HeavyBoat (henceforth HB, for simplicity’s sake) gave me the chance to hand over my sound & music responsibilities, and focus on my actual area of expertise, which is writing. The first “action-packed dodgeball game” pitch document we presented to CN back in November 2013 featured the working title of “Dodge the Chaos”, and a basic storyline about a dodgeball match gone wrong. Skips, the all-powerful immortal yeti, would throw the ball “with such power and fury” that it managed to rip space open and tear a portal into a dimension where dodgeball was the main form of warfare. If you haven’t seen the show, trust me – it’d have fit in.

The original prototype, with a post-apocalyptic “Mad Max” style theme.

Together with a fairly close-to-the-end-result prototype, the document helped sell the idea – but not too easily. By the end of 2013 one of HB’s founders, Juan, was flying around the world trying to strengthen our relationship with our clients and searching for some new possibilities. In a chat, he told me people at CN were concerned that the basic concept we had might not be as wide-appealing as intended: making what would essentially be a sports game didn’t sound wacky enough for a Regular Show game, especially considering there was actually a dodgeball-themed episode about to air.

War Vs. Sports

Ironically, I was worried about the other half of the concept. Recent Regular Show games had been built around the basic “War” idea: there was “Paint War”, “The Great Prank War”… were we really going to do a WAR too? I suggested a different setting: the show is known for its 80s references, but had never really used the “futuristic sports” subgenre, seen in movies like “Rollerball” or “Tron”. So I pitched it in my next draft of the story: it had Pops (a cheerful old fellow) putting the other characters through some group exercises “to strengthen their teamwork spirit”, until an innocent ball game would get out of hand, and a serious, technologically advanced version of himself would emerge from a portal to kidnap the whole group into the future.

The guys at CN liked the basic setup of Pops’ compliment game and the futuristic timeline, but were still wary of any kind of sports metaphor. Fortunately, once I explained my concerns about overusing the war theme they quickly agreed, and we started going back and forth with the story during January 2014, in true collaborative fashion. Working via message boards, we re-wrote each other keeping whatever we liked from the latest version. They came up with the idea of Pops’ tears giving life to the compliment game’s ball, and I suggested he buried it so that it’d grow underground for centuries, influencing a whole city filled with gentlemen who used “the GAME” to run wild. Even the idea for the final boss was there early on - though they wisely simplified my over-complicated motivation for Future Pops to kidnap his younger self.

There were still a couple of issues to solve, but everyone was happy with the story. On March 7^th I happened to be in Los Angeles, so I was invited to pitch it to J.G. Quintel (creator of Regular Show) and his team, which I happily did. It went great: they got the references immediately, it didn’t contradict anything they were planning, and they really seemed to enjoy it. J.G.’s comments were short and wise: don’t flood the game with too many cutscenes, just make it fun!

HeavyBoat’s Buried Treasures

“Grudgeball” would be the third and last game to be released from our initial plan to develop as many mobile games at the same time - but that doesn’t mean the other two were any smaller. Although they couldn’t be more different in gameplay (one was a classic “snake” game and the other a runner), both had a protagonist called Jake, and would eventually get an eerily similar name: Adventure Time’s “Treasure Fetch” (another mobile expansion of a previous game of ours), and Jake and the Neverland Pirates’ “Treasure Trek” (based on the Disney Junior TV show).

While discussions progressed on the context and storyline of our new game, the first few months were dedicated to recreating the original prototype and defining the core gameplay. Yet slowly but steadily –and largely unbeknownst to us, in those optimistic days-, every unresolved issue and delay that the “Jake” games suffered started snowballing into “Grudgeball”. The Adventure Time game, in particular, had been in development for so long it seemed determined to mirror its main character’s elastic powers. Consequently, by the time it was finally done, another long-gestating project (to be released as of today) was just being greenlit, so with “Treasure Trek” also in its final stages, “Grudgeball” became the victim of some sort of middle child syndrome. Its brothers were getting all the attention: the elder ones as their release dates approached, the new baby as it needed to define its concept in order to get underway.

Doubling the number of employees meant there were lots of new people thrown into production almost as soon as they walked through the door. So delays were to be expected, but things started to get a little out of hand. We realized we were still not enough, and started bringing in people to save our troubled projects. Some of the new guys had little to no experience, so they had to learn along the way: the inevitable process of detecting personal strengths and weaknesses (so each one’s responsibilities could be properly defined), would have to be done during production.

And no area suffered these changes as much as our Art Department.

The Lost Boys

HB’s Art Director, Leo, worked on the concept art for the pitch document, but soon he would find himself stuck in Neverland as “Treasure Trek” was nearing completion. Discussions then started on who should be Grudgeball’s Lead Artist. One of the candidates was a very talented guy who did the first round of cyberpunk concept art once the futuristic setting was approved, but he unfortunately didn’t have enough experience to lead the group. Eventually, he got transferred to the “baby” project I mentioned before, only to find himself trapped in endless months of redefining the basic game concept – but that’s a whole other story.

The original court design we had in the game, closer to “Aliens” than “Blade Runner”.

For about three months, Grudgeball’s Art team was almost exclusively comprised of two artists who’d just entered the company. They did a great job, but couldn’t count on the guidance they should have had. Finally, around march, one of HB’s most “veteran” artists (who had been working on the Adventure Time snake game), requested the job.

The Art guys had a Leader… until he decided to leave the company a couple of weeks later.

Juan thinks it was “probably the first big blow we got as a consequence of HB’s growth. Back then we were starting to think about how to communicate things to twenty-five people instead of six. I think he left partly because this wasn’t what he had in mind.” Grudgeball carried on without a unified view of how it should look like, and many decisions that should have been taken in those early days would haunt us til the very end of production. The game’s UI (User Interface) was particularly hurt by this situation, as each artist inevitably started doing his own particular take on any given screen.

It wasn’t until “Treasure Trek” was finally done that Leo could dedicate full time to steer Grudgeball’s visual design in a definitive direction. There was the expected friction, of course: the “new” person coming in to take over, unapologetically focused on finishing the game in time… you know how the story goes. But it was hardly the last big personnel change along the way.

Chaos by Design

Halfway into development, the other two HB founders started moving away from day-to-day operations, which lead to Grudgeball’s Game Designer, Fersis, being promoted to GD Director. Our QA Lead, Gustavo (Gus), was offered the chance to take his place, after showing interest in designing for a long time. The transition was handled as smoothly as possible, with at least a couple of weeks of Fersis and Gus working side by side, but it nevertheless slowed things down in an already delayed project. “When I took over from Fersis, the game was little more than a prototype. There were a lot of things still not defined, and I wasn’t very sure what I could change and what was already set in stone”, Gus remembers. “It was a ton of work for Gus; we threw the guy right into the ocean” Fersis admits.

As much as you’re willing to pick up from where your predecessor left off, bringing in a new GD obviously means bringing a new take on the project. Fersis had more of a turn-based structure in mind, where each side could perform a single action at a time; Gus was more keen on an action-based style. Somewhat inevitably, Grudgeball today has a little bit of both tendencies. Some team members wonder this mixture might have made the game too fast to feel truly strategic, and too slow to have a strong action feel. Would we have a more “solid” product with a single Game Designer?

“What tends to happen is that when there are constraints, the impulse is to ignore them and design a separate game, with features which don’t gel with the core gameplay” Fersis thinks. Gus’ view on the subject: “People with a programming background usually want to have a lot of control over their design. I do like turns and strategy, but I prefer adding a little chaos, which is one of the most basic forms of fun.” You might even argue the end result is the best of both worlds, but there are as many truths as people out there, and this lack of definition resulted in the core gameplay taking a long time to find its final shape.

Tools of the Trade

In any case, changes in team members were born out of necessity. Changes in technology, on the other hand – well, they were risky choices that our CTO describes as “an open wound. We didn’t quite know neither how to adapt, nor to argue against them.”

The project was built using the Cocos2d-x framework. At some advanced point during development, the team decided to incorporate tools we’d never used before: FMOD for sound and LWF to export Flash files into Cocos – the latter of which promised to be particularly useful for interface design. In retrospect, the brief time dedicated to evaluate and study both tools proved not to be enough. They undoubtedly opened up a lot of new possibilities - but also meant new problems. Lucas, one of our most experienced programmers, ended up dedicating a lot of time fixing issues found in LWF (and then sending them to its developers for the world to enjoy), while FMOD made the game crash so frequently that when time to experiment with it started running out, we made the tough decision to pull it out from the game and rebuild the whole sound using Cocos2d-x's library.

It wasn’t either tool’s fault in the first place, of course – we wouldn’t have chosen them if they hadn’t seemed like great additions from the start – but we just didn’t have enough time to study them and make them work well enough with the other tools we had. At least the LWF story had a slightly happier ending: it took months, but the artists could rely on it to make things that were not possible with Grapefrukt, and without the need of a Coder’s permanent assistance. All those different light patterns on the court reacting to what’s going on in the match, or animations running at the same time within different panels of a same cutscene – we wouldn’t have been able to do those kinds of details (nested animations) before.

Work on the second cutscene, from my own sketch to the guys’ awesome final version.

Deep into the Labyrinth

As 2014 went by, the game started moving away from the prototype, until it ended up returning to the original basic design - particularly in terms of controls. For example, the idea of giving each playable character a different shooting range was tried for a while, before going back to just being able to shoot in any direction. “It always goes like this” Fersis describes; “We work on the basics, move forward, and get to a point where we decide it’s not working. We take a completely different path - then we decide that’s not good enough either. So we go back to the original path, and solve the issue that had pushed us to move away from it. In between, we’ve lost sight of production.” Why has this happened to us more than once, I ask him. “Because we think we’re better than we really are.”

Harsh self-criticisms aside, the truth was the core game still needed many adjustments to get right on track. Particular issues were brought into focus as Argentina was celebrating its Independence Day on July 9^th, and the first big batch of feedback notes arrived from CN. Thankfully, we were already working on some of the reported issues. Input was certainly not quite there yet, but it would soon get vastly improved. We also agreed that the single player mode needed variations to avoid getting repetitive, and boosters definitely needed some work. But there was a single piece of feedback that was of particular importance for us, mainly because it had also been the single most recurrent complaint raised at our previous Regular Show game: the fact that characters weren’t feeling different enough from each other.

A match from the Release version, with the Portals & Multicannon “court features” in action.

Features like Special Attacks, Counterattacks, and the Team Buff (which went through many permutations), were developed to find a way around this particular issue, yet most of us think it was the one thing we couldn’t really solve. “Court Features” worked great though: weird things that show up during the match, like barriers, portals, or the almost-unstoppable ricochet ball. It’s a good thing we also managed to squeeze these wacky effects into the “Versus” mode, considering playing against a friend in the same device is the one aspect of Grudgeball everyone always agreed was as fun as we hoped it would be.

Missed it by that much

Around June 2014, the long-awaited closure of the two “Jake” games meant more care could be taken of our beloved “Grudgeball”. Leo was able to finally lead the Art team, and Fersis found the time to turn his gaze at the game’s day-to-day affairs. That was when a careful look at the project’s Backlog (the long list of tasks required to deliver the product) revealed we would miss release date by a whooping three months.

In retrospect, it might have been the kind of panic the team needed to focus on getting the work done. The cold, hard numbers had spoken, and drastic measures had to be taken. Fersis restructured Grudgeball’s Backlog to make sure the team would stick to the game’s priorities - and then swiftly cut out everything else. “There were more boosters, obstacles, features, a whole help screen, Arcade Mode was going to be way more complex – all gone”, he remembers. “But once everyone saw the real problem we had, each area was able to find its own solutions.”

We obviously lost some cool stuff in the process (guess we’ll never know how the “invisiball” would have turned out), partly because the team hadn’t given boosters any kind of priority order: the ones you’ll find in the game are the ones we’d produced before these measures had to be taken. And as if working without a Lead Artist for months hadn’t been enough, the ever-expanding Art Department was put to the test once again, with most of the guys put into “timebox” mode. That meant each artist would get a fixed amount of time to pull off a certain portion of the game’s visuals: a character, a certain background, etc. No complains there: everyone agreed it quickly helped building a sustained production rhythm, avoiding endless iterations. We were finally starting to crawl out of the hole we had dug ourselves into.

The Last Few Bricks

For several reasons, release date was changed a couple of times. Originally, it was set around July, then it was moved to September, and CN finally settled on February 2015 so as to separate the game’s launch from that of other titles already in production. We couldn’t really keep a full crew working past October without complicating our own internal timeline, so the main Grudgeball team disbanded at the end of the month, while a few of us picked up the pieces and carried on.

Incorporating the localized lines of text was one of many remaining tasks: there’s a substantial amount of words in the game (in 10 languages!), so checking if every single line fit within the text boxes was no quick thing for the QA guys. That much was expected, anyway - there were more complicated issues to tackle. October saw the arrival of another set of tough-but-true feedback from CN: the game definitely needed more explanations and a better presentation of progression.

The Loading Screen – Before and After

Curiously enough, an important chunk of the adjustments both partners came up with in order to address those issues were laid over the game’s loading screen. For example, we incorporated around 30 gameplay tips I was asked to write to better show the extent of the game’s contents, and most importantly - the elevator you see on that screen (which is basically the moving court the characters stand on) got the animation it deserved. It seems crazy it didn’t move until then: you’d see it fixed in the current level. But it was the kind of key detail that got buried under the huge amount of small adjustments required at the same time, and CN’s feedback helped bringing it to the forefront.

Aside from re-implementing the game’s music and SFX after pulling FMOD from the project, November and December were mainly re-balancing and bug-fixing months – and there were tons of them. By then, energy was running dangerously low. February would blissfully take the game off our hands, and into the world.

To say the things we truly feel

“It was the project in which they tried everything. Technology changes, team changes, administration changes – everything changed along the way. But surprisingly, against all odds, it turned out pretty well.”

Those are the basic feelings shared by the team, as eloquently put by Fede, the game’s QA Lead. Most of the guys are satisfied with the end result, but the single thing everyone agrees on is that the game looks awesome - so everything the art department had to endure at least bore its fruits. Little victories for HeavyBoat -considering our past mistakes- include the following: we finally have a final boss that feels like a final boss, cutscenes are way better than in “Best Park in the Universe”, and the volume of content is great – especially considering the wide variety of animated characters we managed to pull off.

The playable section of our colorful futuristic cast.

Regrets? We have a few, as the song goes. Some of the issues along the way we couldn’t really fix: it still can get a bit repetitive, characters don’t feel wildly different, and some information might not have been shown as clearly as it should have.

Ultimately, everyone is as happy as exhaustion permits. “I think it’s amazing that the game more or less looks just like (the prototype). That’s a first for us” says Juan. “It’s inevitable to say that it turned out pretty good considering the chaos it was to produce. Compared to ‘Best Park in the Universe’, we achieved what we set out to do”, Ale (our CTO) thinks. Fersis’ summary: “I like the way it turned out. It feels good, the art is great, the cutscenes are by far the best we’ve done, the tutorial is by far the best we’ve done (it’s inside the story, as it should be). It’s a cool game…”

“…but production was a disaster.”

Two Steps Back

So, the Grudgeball project was expected to define whether we had “grown” successfully or not. It was the last, definitive stage of a test to prove ourselves if we could really manage three games at the same time. So how did that go?

Well, we were clearly not ready. All three parallel projects took us longer than we had estimated, and release dates determined which ones would get the experienced people needed to pull them off in time. Unfortunately, in the end neither of the “Treasure” games did as well as expected, and development of our very next project wasn’t any simpler.

This chapter of HB’s history had begun alongside Grudgeball’s inception, so it was appropriate that it would close around the game’s release. That happened on February 19^th 2015, one day after 13 people - basically half of HeavyBoat - were told we’d have to part companies by the end of March. It was time to admit our grand experiment had failed, and a close look at the numbers offered two possibilities: to keep on until we all went bankrupt, or to react as soon as possible, go back to being a small group, and give people no longer in the company enough time to get other jobs, paying them every penny the law indicates for these situations – a practice not many companies around here honor, sadly to say.

So, by the time you read this, a small number of us will have returned to our previous, smaller office, back where we were a year and a half ago.

Out There

Thankfully, the great reception “Grudgeball” is having has cast some light over this difficult period. We thought it’d get buried under a mass of new releases or get a mild 3.5 rating, but instead, it’s been featured in the App Store and Google Play, it has a 4.3 rating (as of today), and has gone as far up as number two in Best Selling Games lists in under a month. Professional reviews have been really positive, as well as most customers’ opinions in the stores.

Who knows - it might even turn out to be the best-received game we’ve ever done. Yet another old story: you never know how people are going to receive your product, and if you’ve been long enough on the inside, you can’t really see beyond the pains it took to get made.

So thanks to everyone that made Grudgeball what it is: Gus, Fersis, Santi, Fede Donnet, Nacho, Ale, Diego, Nico, Fuego, Csr, Gera, Leo, Gonza, Julián, Fabi, Fede Pérez, Iván, Lucas, Jere and Juan N. Thanks to everyone at CN Digital for the opportunity and all the hard work, and to CN Studios for creating the show and letting us play with its universe.

Finally, thank you for reading as well - hopefully, I’ll be back in some time to tell you how we’re doing.

-- Guillermo Crespi

[email protected] / www.guillermocrespi.com

Images courtesy Cartoon Network/HeavyBoat