Oasis was the winner of the IGF’s 2004 Game of the Year and Innovation in Game Design awards in the web/downloadable category, but the history of this project was anything but glamorous. It was created by Mind Control Software of San Rafael, California, a company that has specialized in game design since its formation in 1994. The company’s first project was Alien Logic, an action/RPG based upon SkyRealms of Jorune, a paper-and-pencil role-playing game designed by the company’s founder, Andrew Leker. Mind Control Software has historically split its time between development for hire and its own independent aspirations. Oasis fit squarely into the latter and eventually saw the light of day in April 2005, with PlayFirst as its publisher.

Oasis is an Egyptian themed, turn-based strategy game played on a 10x10 grid shrouded in a mysterious fog. The player has 85 turns to move his character, the Scarab King, throughout the playfield in hopes of rebuilding a civilization that will survive the inevitable barbarian onslaught that arrives after the 85th turn.

The levels, consisting of desert, mountains, cities, an oasis, and barbarian entry points, are randomly constructed according to some fairly complicated heuristics. The goal of each level is to both find the obelisk, which is hidden in one of the oasis tiles, and to defeat the barbarians militarily. The player is rewarded with a magical glyph each time he or she succeeds. Players win the game when they have obtained twelve glyphs.

Oasis has been described as Civilization™ meets Minesweeper™, as it has the quick, clicky feel of Minesweeper, and the empire building aspects of Civilization.

Introduction

Oasis started out as a quickie design that I scrawled in my design pad on a particularly slow day back in mid 2002. The dot-com bust and the sluggish post 9/11 economy created the opportunity for concentrated design time, which I seized upon. The first draft of the design felt very compelling. It emphasized discovery, resource allocation tension, and oddly enough, an emotional misdirect that I hoped would add to the game’s mystique.

The first person with whom I shared the design principles of Oasis loved it. He didn’t think that the game would be commercially viable, but he said he’d be the first to play it if it was available. This was to be the catchphrase of Oasis, “I don’t know who’d buy this game besides me and my friends.” As a businessman, I thought it best that I put Oasis on hold, probably indefinitely. Then, in fall of 2002, I described Oasis to my mentor in the game industry, William (Bill) Dunn.

He immediately offered to pay for a prototype, which shamed me into creating one post-haste. I created the first Java prototype of Oasis over the Thanksgiving weekend in 2002. Oasis was fun, right off the bat; it just needed a few years of work to polish it. I spent every spare hour I had on Oasis, turning over some of the work to a programmer that Bill paid to work remotely on the game. Combined with the art resources of the company, we had a sharp looking Java prototype.

Buoyed by a sense of confidence in my little Java prototype, I made my boldest move. I emailed a copy of Oasis to Marc LeBlanc, a friend and former co-worker at Visual Concepts. Marc and I taught game design at each Game Developers Conference to about one hundred industry folks each year through a tutorial that Marc had devised.

There are two things to know about Marc LeBlanc: 1) he’s exceptionally smart (almost cripplingly so), and 2) he’s exceptionally opinionated. I thought carefully before hitting the send button. He might hate Oasis and waste a chunk of my life with long rants or a flame. In fact, I did receive the rants, but they were aimed at constructively improving the game. He was bitten by the Oasis bug. Marc joined Mind Control Software in the summer of 2003, intent on creating a C++/DirectX version of Oasis in time for the September 1st Independent Games Festival application date, and winning an award for Oasis.

What had started as a slow afternoon was turning into a significant investment of time, resources, and brain share. All bets were on Oasis.

What Went Right

1. Design Principles. I should say up front that I’m a fervent believer in Formal Abstract Design Tools (see Gamasutra article) and I teach the MDA game design framework each year with Marc at GDC. Those concepts and that framework were extremely valuable when deciding how to implement a feature to achieve a desired result. It aided communication and sped up the tuning process significantly. That said, the goals that I had initially set for Oasis came from no framework; they were bits of inspiration and wisdom gathered ad-hoc from many sources. This appears to be a truth of game design, that process and theory can help when things go wrong, and they can help a designer pinpoint the relationship between a rule and a result, but inspiration and experience are hard to express and even harder to teach. Thus, I can’t justify the core principles chosen to create Oasis, but here they are:

2. Playable from the Start.The first Oasis prototype took two days to create. I coded it in Java because I had fallen in love with the JetBrains’ IntelliJ editor and wanted an excuse to play with it. That first version of Oasis was an ugly beast, as the programmer was also the artist, but the basic game was fully playable. Over the following weeks additional features were added, such as Technology, and the ability to play more than one level without restarting. However, starting with a playable game had numerous benefits.

It was challenging to cut the feature set of Oasis down to something small enough to implement in a weekend, but a larger prototype might never have been made. The promise of the initial version propelled the project forward.

A philosophy of prototypes that I’m fond of goes like this: as a designer, you’re in a dark room and your game can be found somewhere on the walls if only you had a flashlight to search with. Game code is that flashlight and the batteries come from the energy of your programmers. The moral of the story is to find your game early, even if it means creating a throw-away prototype.

3. No Feature Creep. Oasis managed to survive the almost unavoidable tendency to add unnecessary features to a game. There were plenty of suggestions that we received from our beta testers that seemed good at first blush, but which fell apart upon further examination. It was amazing how few ideas really fit in, probably because Oasis is such a small, tight game.

Perhaps the urge to add features was mitigated because Oasis was not only fun, but challenging from its first incarnation. Historically, I have been most tempted to add features during development when my current project isn’t as much fun to play as I had hoped. This might explain feature creep within design teams: developers want their game to be fun and challenging during development, even in a feature incomplete state. A better approach might be to tune the game toward the target audience, which does not generally include people with knowledge of a game’s inner workings.

4. Stick to Principles. Oasis is a game that strays from several turn-based game conventions, but hopefully to great effect. After the player has taken their 85 turns, they have ten seconds to deploy their population along roads to cities of their choice. This was a controversial decision, as it created a time limit in what is otherwise a turn-based game. We kept this feature, as it creates a point of tension that leads to a climax as battle begins. In our minds, it’s not that different from a swing meter in a golf game. The exciting “moment of flight” in golf is predicated by a brief time-critical period where swing strength is established.

Another controversial area was combat depiction. For some time we toyed with the idea of a cutaway combat resolution sequence. There’s no doubt that we could have created a more beautiful view of battle, but it would have killed the graceful continuity that was achieved by keeping battles on the playfield.

There are several possible outcomes at the end of each level of Oasis:

The concept of losing a level militarily but not losing the game was controversial. Perhaps it was confusing to “lose” but be allowed to continue to play. However, this was a crucial part of the game’s design, as there was no way for us to easily tune the levels to be “fair.” Without fairness backing us up, we needed a safety net, which was the life force in Scarab King’s staff. This fit beautifully into the notion that clicking on the Oasis rewards the player with life points to be used in the case that the barbarians defeat the player’s cities. The conflict between ensuring a military victory and stocking up for a rainy day ended up being a signature design feature of Oasis.

5. Unexpected Help. Oasis improved significantly in its final months as production came to a close. Much of the game tuning had already occurred, thanks to the help of more than 6000 beta testers we solicited early in development. Their feedback on our message board not only helped us identify exploits and cheats, but helped us see holes in the tutorial and various aspects of the game’s presentation.

Our publisher also made some suggestions that improved the game’s presentation; the visual sparkle that accompanies the completion of a road network to a city is a good example. They also created a high score API that allowed players to see their global ranking at each difficulty level. Though we never played Oasis for points during development, it turned out that many Oasis purchasers were fanatical about their scores. Seeing daily, weekly, monthly, and all time high scores was clearly a boon to the community of Oasis players.

We knew from the beginning that audio would play a large role in the perceived fun of Oasis. We were lucky to have an award-winning composer, Michael Sweet, help out with the early sound effects, which was very helpful when we submitted a version of Oasis to the IGF judges. However, as time passed, so did the game’s presentation, requiring new audio. Luckily, a young composer, Harry Mack, a friend of one of our employees, created some sound and music for Oasis on the off chance that I’d like it. I loved it. His initiative paid off and we dedicated considerable resources into tuning the audio to completion.

Our family and friends believed in Oasis and helped play-test and tune the game. Their faith lent strength to the entire endeavor. We also received their financial support during some dark times when working on Oasis distracted us from pursuing consulting work.

What Went Wrong

1. Three engines. Oasis was originally built on a DirectX 9 framework. That seemed a safe bet because we hadn’t considered signing with a publisher that might require a lesser version. When we did sign with our first publisher, we had to migrate over to their DirectX 7 technology, which took a considerable amount of time. When we moved on to our final publisher, we had to strip out all of the previous technology and start from scratch. This was made considerably easier by the fact that Mind Control Software’s own DirectX 7-based Orbital technology had gone through two ship cycles.

There were certain benefits to using three engines, but overall, we spent more time on technology than was desirable. Going with DirectX 7 from the start, which was an industry standard for small games, would have been a better idea.

2. Learning Oasis. In creating Oasis, we were faced with the challenge of teaching the game to new players. This proved more difficult than we had imagined, as 1) There were more ideas that had to be understood than we realized, 2) We already knew the rules, and were a bit desensitized to a newcomer’s frustrations, and 3) Oasis breaks with convention in some of its rules, betraying some of our target customer’s expectations.

The game’s tutorial was extremely well crafted for its purpose. Each important idea was presented in the form of a level in the tutorial. The player would have to master a concept before they could progress to the next level. One downside to this approach is that expert players, those who need little hand-holding, did not have a straightforward rule set to learn from. Worse, the tutorial is extremely easy to play, creating the impression that levels can be won trivially. Some of these issues remain unresolved, but we did create multiple entry points into the tutorial so that players can refresh their knowledge of important rules immediately, without having to replay the whole tutorial from the start.

Our insensitivity to the frustrations of newcomers was probably predictable. Long development cycles tend to increase the divergence of spirit between game creators and their audience. Some of the focus test results seemed impossible to imagine, though video-taped evidence and careful observation confirmed them. The feedback that players receive in Oasis is quite different than that of a traditional mass-market ultra-casual game. Players may ponder why a city didn’t fight as well as expected and draw invalid conclusions. That’s rarely the case in a bubble-popper game.

Where Oasis breaks rule conventions for strategy games, it does so cruelly and deliberately. For example, the fields surrounding each city contain the lowest numbers of followers found in the playfield. Clicking on a field yields only one follower. However, games of this ilk generally reward players for finding and controlling all agricultural land. This created a valuable point of tension in the design: people in the surrounding land have taken refuge in the city; thus, deducing where a city is in the minimum number of clicks is advantageous. If we had gone with convention, it would have encouraged sloppy play. A tutorial level might have helped address this point of confusion.

3. Women and the war-game. Depending upon one’s definition, Oasis is either a war-game or an empire simulator. The player builds an empire and defends it against invaders. However, unlike most games of this type, there is no micro-management, levels are played in minutes instead of hours, resolution is dramatic and fast, and the game emphasizes discovery and building more than any other element. Thus, even though Oasis plays like a beautiful game of discovery and exploration, it looks like a war-game. Well, the war-game market is not dominated by women, and Oasis is distributed through web portals that typically cater to women in their late 30s and 40s.

The development team was at peace with this conclusion. There didn’t seem to be much that could be done to make Oasis more appealing to the actual market of potential female players. What we didn’t know was that many women who self-identified as hating war-games actually loved Oasis. Their claims that the game was incredibly addictive and satisfied their desire to explore and build came as a shock to us. We lost faith in targeting the female audience after listening to many women’s initial reactions to our little war-game. In retrospect, we should have worked harder to identify and fix what was causing such a strongly negative reaction.

4. Difficulty Ramp. When players migrate from Easy to Normal difficulty the game seems just as friendly because the first levels of Normal really are like the last levels of Easy. However, about halfway through a game of Normal players notice for the first time that there are mistakes that can cost them the level, and even the game. Typically, these are mistakes that they’ve been allowed to get away with while playing on Easy. Perhaps the game has trained them to play sloppily up to this point, and now they get their comeuppance.

The difficulty ramp in Oasis isn’t really that different than other games, but Oasis players express more shock and disbelief when their cities fall than, say, a player who loses in chess or a real-time strategy game. In fact, players have described the game as “betraying” them when multiple hordes of barbarians arrive on the playfield. What’s interesting is that multiple hordes are almost always preferable to a single, larger horde. Again, it is the shock that is so notable. It is as if the game creates a sense of comfort and contentment in its players that is betrayed on higher levels of difficulty. Some players quit when they reach this point. Others push through by tightening up their play.

Generally, the shock and horror of seeing one’s cities burn can be averted with very simple tactics: 1) Find mines early and place ten followers in each, 2) Find cities early and connect them so that populations will grow, 3) Find or deduce the barbarian entry point(s), 4) Search cities closest to the entry points for treasures, 5) Reinforce the cities closest to the entry points that have the strongest weapons. 6) Don’t waste turns clicking on fields because they only contain one follower each.

These tactics may be easy enough to accomplish, but it appears that the game might emphasize certain rewards that throw players off of these goals. In particular, the Oasis is a beautiful blue region that makes pretty sounds when discovered, and turns a brighter color when fully explored. That enticement alone is enough to distract players the moment they find the Oasis. The discipline required to resist is one of the game’s strengths, but evidently, it may teach players to play for immediate gratification that does not promote survival.

Oddly enough, the inclusion of advisors, like the Engineer and the General, can make Oasis more challenging to some players. When advisors were first implemented and I reported this finding to Marc, his response was, “That’s impossible. Advisors do nothing but make the game easier to win. Period.” My response was indicative of side effects we have witnessed, “When an advisor is available, I want to earn it so badly that I play poorly.”

Enter the world of a game that rewards players to their deaths. This is a wonderful strength to those who stick with the game, but not for the initiates.

5. Insufficient early platform testing. The inclusion of both 2D and 3D modes in Oasis significantly increased the difficulty of the development and QA process. Oddball graphics cards misreport their feature sets and capriciously violate standards. This tale of woe is mentioned frequently in the What Went Wrong portion of postmortems, so we were warned, but our initial success with the graphics cards we had on hand was good enough that we didn’t start testing the less common cards until late in QA. In an attempt to fix some very obscure problems, we ended up introducing some quirks on some less problematic chipsets.

Conclusion

Seeing Oasis ship was the culmination of a dream that started more than two years ago. We have experienced the joy of both seeing it win the IGF’s Game of the Year and Innovation in Game Design awards in 2004, and seeing it in the hands of the public. The difficulty in bringing this product to the market can not be understated. The resources that it consumed could have just as easily been spent on creating several less sophisticated and less fulfilling games. Only with the support and hard work of many people could such a game reach the market.

Hopefully Oasis will usher in a new era of light strategy games that can be played at a leisurely pace. There are limitless directions left unexplored, and if we have inspired others to take a crack at this space, it will all have been worthwhile.