Shelled Post-Mortem

Shelled Post-Mortem

Shelled! was a game idea conceived of in January of 2005 after working with the GarageGames Torque Game Engine (TGE) for almost a year on another project, a zombie ballet game that was eventually cancelled. Shelled! was envisioned as a quick side project that would take 4 months and would modify the engine in small ways to leverage its benefits and therefore be very quick and easy to finish. The philosophy was to keep it simple.

It was concept-pitched as a 3D Scorched Earth with the air-cam of Orbz and simple controls and graphics, in line with titles such as ThinkTanks. At the time both Orbz and ThinkTanks were nearly the only games on GarageGames and both seen as role models of how to do a successful 3D casual game. Both went on to Xbox and saw wide distribution.

It seemed a sure-fire formula. I felt no need to prototype because the gameplay was already proven. Scorched Earth is well loved, and new players could be introduced. Plus, nobody's done a 3D Scorched Earth that was (in my opinion) pick up and play. There are many clones both 2D and 3D, but none that really nailed it. The market was out there for an easy to play 3D Scorch, and I planned to make the game that captured it.

A ROCKY START

I launched into production quality art from day one. I reasoned I was getting it so cheap (at great discount from the artists for interest in the work) that I felt it wouldn't hurt if I had to throw some out, or had to iterate it at further expense later. In addition, I felt the art needed to be strong early on to properly convey what I felt was so special about the game's potential. Another additional benefit was looking the part as much as possible when trying to recruit further for the team.

In hindsight, one of the biggest mistakes was not prototyping early enough and feeling too confident about the play mechanic in my head. Another huge mistake was production art too early in. Art and polish assets should only go in when absolutely necessary, as it gives you more room to creatively move about with less things breaking as a result, being less "boxed in," and less backtracking as well. As said, this did have the benefit of getting momentum in team building, but there are probably better ways to accomplish that.

So 4 months total rolled by and there was no way it'd be done by then - we barely had the tank working, let alone multiplatform release candidates complete with DRM! Having production art so early was deceiving and made the game appear to be more finished than it actually was. Then 6 months total, then 8, then things were looking pretty good, we had a few bugs and untested things but we had got a heck of a lot of features. Problem was we were focusing on features instead of player experience.

OUT-FEATURING THE COMPETITION

Early on, I created a grid showing comparative games (Scorched Earth, ThinkTanks, clones) and features those games had (number of weapons, levels, tanks, etc) and made sure that my design compared well to the original, its clones, and similar indie games. I didn't try to best everything - Scorch had interesting variations like bouncing bounds available; ThinkTanks had not only different tanks but with varying stats - but I tried to hit the basics and then some.

Our cool stuff, in less than a year: playable 3D game demo, 25 weapons including weapons that transform the ground, 20 page physics lesson custom written for us to teach kids (or whomever) how some science of physics and trajectories work, 5 tank upgrades including shields and a bigger cannon, 30 options hooked up to a GUI for player created custom games, 4 gorgeous skyboxes and grounds from Tim Aste (which was later donated to TGE 1.5), great high-quality music tracks, snazzy HUD work from Magnus Blikstad, gorgeous sci-fi loading screens, wind and custom gravity listing all the planets in order with pictures, little turtle men and borg turtle men, team play, we even had a 2D color matching mini-game with fantastic 3D art (which later split off as Sploidz). Boy, we had it all! Except, um, tight fun core gameplay (whoops!).

Two more months and we got everything set. There were some bugs and an ill-considered last minute control scheme change. We didn't make a build for IGC 2005 but heck, we'd been working on this for nearly a year, so we thought let's show our latest build anyway, warts and all.

The game hit with a thud, and not just the rough spots. Many were supportive (at IGC 2004 I didn't even have a game, so any game was progress), but there was definite criticism. It was direct and it was hard to hear. But it was necessary to learn how to build a better game. Intuitively I knew what had to be done, but the criticism was the boot in the ass. I had already spent a year and multiple thousands on it, but I redoubled my efforts to get it done right.

TRIMMING THE FAT

As 2005 ended (1 year ago) and through early 2006, dramatic changes took place:

25 weapons were cut to 9, the physics lesson was cut, all tank upgrades were cut, all custom options including working team play were cut, both Tim and Magnus's art were cut, the sci-fi loading screens were cut, wind and custom gravity were cut, both turtles were cut, the super-cool mech tank was cut, the 2D mini-game was cut, the controls and firing sequence was totally changed to be greatly simplified, and the art style committed to cartoon, where it wavered between sci-fi and toon before.

Big changes! Note that all this art and code was completed contract work that had already been paid for. In fact, it cost me to have the programmers remove all this stuff and gut the game to a still playable state!

Once the dust settled, the game immediately played better. Though cautious of the cuts beforehand, we knew we did the right thing once we saw the proof of this new build. We were bolstered to press onward.

SUPPORTING THE FOUNDATION

For all of the Winter and Spring of 2006, work was spent on refining this new game that emerged, iteration upon iteration, polishing and polishing, getting more and more confident that we were getting it right, this time focusing on the core game. In late spring we had our first beta test and the general response was great (and consistent!):

"Overall a nice game that - in my opinion - needs a bit of polishing, but definately has potential. Keep up the good work!" "Man this is so much better than it was. Overall it needs more polish but in general it's fun to play." "this is a very fun game!" "Awesomely fun, the feeling when everybody is shooting shells is just wicked. " "It obviously needs more polish graphically, but I think you have a winner." "If you wanted to capture the feel of Scorched Earth, you succeeded, Shelled is way more fun." "I just wanted to let you know that I've found myself playing the Shelled beta more than I anticipated"

This is a totally random assortment that I just pulled from my email. Enlivened by the beta feedback, as well as generally good feedback from GarageGames staff, we continued to polish and refine month after month. The carrot seemed a perpetual "few months" away.

ANOTHER RE-DESIGN?

By summer the game was pretty much in its final state and was shown in earnest at the Northwest Games Fest in Portland. The title drew little attention and didn't place in the top 3 games, of which only 5 games entered. People that played it generally seemed confused and weren't playing as it was intended. The controls threw people despite their greatly simplified format. People wanted to fire straight, despite it being a trajectory game. It still wasn't adding up, at least in person.

Frustrated though willing to do whatever it took - even a second radical design change - I started on the next iteration of the game, which would eliminate the uniquely blended first and third person mechanics and strip it into a more traditional strictly third person mechanic. In addition, you would no longer directly pilot the tanks, but rather control by proxy, lowering the learning curve further through more simplification.

CROSSING THE FINISH LINE

Perhaps another re-design would have made for a more simple and fun game with the widest possible audience. But part of making this game was getting a game of this scale completed under a certain amount and within a reasonable timeline, and that budget was over (topping $8k when I planned to spend $4-6k) and that time was up (22 months in dev when I was planning for 12-16). We finished up networking functionality and last-minute bugs, then released it as a free game (Nov 2006).

The decision to stop the game where it was at was a hard but good decision. It was about as fun as it was going to get. Further refinement could tighten things further, or it could loosen them, or it could just be different but no better. You don't know until you try. This game was in late production and now was not the time to be prototyping core gameplay mechanics. Now was the time to finish.

FREEING MY GAME

The biggest question fielded since its release is this: why free? Clearly given the beta feedback (and feedback since release) the game will have its fans (and already does). However, clearly from the in-person playtesting, Shelled is not for the broadest audience possible (despite graphical and thematic cues of broad appeal).

That being the case, Shelled could make, say, a few thousand at best? Maybe earn half the cost of the game back in a few years, optimistically? On the other hand, it could be given away free, potentially spread like wildfire and mod'ed like crazy, and be worth far more in reputation than the money could buy. A half million lifetime downloads? (Like the free Scorched3D has had) Hell yes, sign me up! Versus 500 lifetime sales? Maybe I'm too all-or-nothing after putting so much into the game, but for that little potential return, I'd rather take the risk and make the game free in an attempt to maximize its reach.

In the bigger picture, all along when I would run imaginary sales numbers by my friend, he was supportive but always added that even if the game didn't make a lot of money back, that the game was "money in the bank" for its worth in opening doors (and others, like Jeff T, echoed this sentiment). And open doors it has, as I'm now involved with GarageGames. Had I not produced this game (along with Sploidz), I would never have been involved with the company, community, people, techonology, or industry as closely as I have, and if that were the case, I wouldn't have had the experience required to do what I'm doing now. So open doors it has, just like I was told all along. I have everything I need, so I have nothing to lose by making it free.

LESSONS LEARNED

That's the story of Shelled. My 3 years of DIY gamemaking isn't vast knowledge, but it's certainly hard-won. Advice starts with this one: make games. The more you make the more you learn. The more you learn the more you can apply to future projects. One of those will be successful, or perhaps many of them, but not likely your first one or even first few. Whatever other mistakes you make that are more specific, none of them matter if you stop making games. Make games period and you are bounds ahead of the rest.

Onto specific mistakes, some previously mentioned: getting production art too early; failure to prototype (!); failure to prioritize game features; overcomplicating the GUI, controls, game concept, game objective, among other things; failure to focus on core fun; failure to have first public playtest early enough (10 months into project!); failure to have next playtest soon enough (8 months later!); failure to have a sufficient quantity of total playtests (only 4 in 22 months!).

We also did a lot right. Foremost, we (the whole team) were open enough with the game to consider and execute on bold changes which improved upon the game tremendously. Had we not been open to such dramatic change, we would have ended up with something that not even we enjoyed. Another way of saying this is that we accepted that some mistakes were made, and we were committed to doing it right even if it meant more work.

Other things we did right: trying to err on the side of simplicity (with mixed success), looking at what other games in the genre did right (key in initial design), and taking player feedback into direct account for design change considerations (absolutely key for ongoing improvements). That, and about an oil tanker of perseverance.

As final advice on game-making, I cannot suggest strongly enough to make not just your first game but your first few games as simple 2D games. The craft and process that goes into making a 2D game is exactly the same as a 3D game, but more managable and with less requirement for expert design to achieve even a base level of fun. I love that I have had both 2D and 3D experience, but had I planned it beforehand, I would have spent $8k and 22 months making four 2D games, one after the other. On the other hand, my passion was for this project, and it is hard to deter passion, which is why the GarageGames 3D engine will always sell itself.

WRAP-UP

Since there were so many assorted artists, there will be no art post-mortem, but I'm sure Gary Preston (lead programmer) will put in a word or two on the technical side at some point for a programmer's post-mortem. Shelled was quite the long-haul experience for him. Though I've already said it twelve times to Sunday, it bears repeating: thanks to Gary for sticking with it and being a great lead and making this game possible and as high of quality as it is.

Mass-market or not, I had a lot of fun making this game, and the struggles don't cancel that out. So many nights I'd come home from work and download the latest build, or check out a new art asset, and the air crackled with excitement. I'm absolutely proud of the accomplishment and of the game we made, and I learned a lot. I have a blast playing the final product, and know it will find its audience out there. For a game to find its audience - what greater return is there than that? That's what drives me to make games.

Well that's it for this one. Be sure to grab the game free if you haven't already. There's a-shellin' good time to be had!