Around the end of July Psyop began working on Nightmare: Malaria - a physics platformer set inside the body of a young girl who is dying of malaria. Players would control that girl, floating, bouncing, and swimming inside her own bloodstream - with death lurking at every turn.
The game is a pro-bono initiative to raise awareness both for preventing malaria generally, and for a specific charity, the Against Malaria Foundation. We were allotted 3 months to concept, prototype, and release our title for iOS and Android, with a target completion date of November 1.
The project began with an attempt to define success for all of the relevant stakeholders. For the Against Malaria Foundation, their goal was, and is, to prevent malaria through the purchase of insecticide treated bed nets, which means money. For the development team at Psyop, the goal was to create a visually stunning game. Early inspiration came from Peter Singer's TED talk on effective altruism, and it led to a particular line of reasoning.
However, there were some restrictions. One of AMF’s unique strengths as a charity is that 100% of donations they receive are used to purchase nets, less only nominal transaction fees. In keeping with that mission, direct monetization through either IAP or direct sales was not appropriate, due to the 30% cut taken by Apple and Google. Neither would it have been appropriate to reinvest proceeds in advertising; instead it was deemed the ideal product was one which could attract a dedicated audience entirely on it’s own merits.
This ruled out a large number of games, since so many genres are dominated by games which directly reinvest proceeds into buying users, effectively creating profit through a sort of app store arbitrage. Given those restrictions, casual, mass market titles were abandoned in favor of a niche product: a difficult platformer, appealing to a core audience. The game would be entirely free - with the dual goals of educating users about the severity of malaria, as well as driving traffic to the AMF website. Since user conversion is primarily the responsibility of AMF, the success metrics were as follows:
From a design perspective, the most important message to be communicated was that bed nets are the most cost effective way to prevent malaria. Secondarily important was emphasizing the severity and dire consequences of contracting malaria in a developing nation. Based on those goals, the following were established as the gameplay priorities.
This initial design was satisfying - the fluid physics and bouncy cells were working well on mobile devices, but there were problems in early testing. The difficulty progression was very steep - and it needed to be so that players could learn the neccessary skills to complete the final levels. The controls were intentionally difficult to master, but players were tiring of training levels, yet frustrated as soon as their progress was blocked. This early balance between boredom and frustration merited more rigorous testing than it was given. Since release, two patches have been released to address feedback regarding early game difficulty and the perception of control 'slipperiness'.
The second problem was that the message was not resonating with the mechanics. Nets and mosquitos were incorporated early on in the design, but even late into production they felt more like a skin, rather than something significant to the gameplay. Furthermore, the levels suffered from a lack of goals and short term feedback loops. Without short term feedback, the smallest lulls in gameplay felt boring. Without goals, early levels were aimless.
Late in production, there were two major changes introduced to address the messaging and goals. The first change, despite it's simplicity, was controversial. Each time the player dies ( read: frequently ), an unskippable 1.5 second randomized message relating some facts about malaria was added. The purpose was to give death some meaning, to increase tension while playing the game, but most of all, to incorporate some real world information. Is a slightly frustrating death screen the best time to incorporate messaging? Probably not, but out of all the screens in the game, it was the most relevant time, and the goal was to relate the players own death in-game, to the millions of actual deaths worldwide. Either way it should have been addressed much earlier in the design.
The second change was more successful. The most important aspect of distributing bed nets, is that by protecting an entire community, malaria can be dramatically reduced, because the parasite requires both humans and mosquitos to reproduce. Once all humans are protected, the parasite cannot spread, and eventually will die out. This is important, because it forms the justification for bed nets being distributed by charities and governments - there are enormous gains if ALL people are protected, and the marginal cost of having one person go unprotected is very high. In economics, that is a classic example of a public good.
Incorporating that concept into the design, even if somewhat loosely, was important. Of course, it would be fascinating to develop an entire game around just the idea of parasitic transmission vectors, but given the temporal, monetary, and aesthetic constraints, exploring new mechanics was not feasible. Instead, a conventional level unlock system was chosen, where players need to save a certain number of teddy bears to unlock future levels. Once a level is complete, players see the bears they saved, safely protected under a bed net, and those left behind skewered and dying on a mosquito's proboscis. Each level contains 3 bears, and each subsequent level requires about 2 additional bears to unlock. The player is not required to play levels in order.
This mechanic was introduced two weeks before release, and it dramatically improved the game. The image of dead teddy bears at the conclusion of each level proved compelling enough that players would immediately begin replaying levels until they successfully collected all of the bears. This was a critical change in player behavior, that required no prompt, text, or tutorial, but simply an image that conveyed emotion. This permitted the introduction of key platforming skills, such as double jump, as early as level 1 - without resorting to tutorials or text. While some users will still find this confusing, it benefitted the players who relish in the joy of discovering the game for themselves.
The challenging secondary objective of collecting all the bears enabled the reduction in the baseline difficulty of every level. Because players perceived the collection of bears as their choice, they were not frustrated at performing the same task over and over again. Furthermore, as players progressed, they could collect enough bears to skip levels they found too difficult, and return to them once they had improved their skills. In one stroke, the content was extended, the difficulty curve lessened, meaningful player choices were facilitated, and the underlying motif of saving others was introduced. In many ways, all of these changes were not mathematically significant, but rather psychologically significant - they proved potent motivators in gameplay testing.
The cutting room floor
Just as important as the successful mechanics, are the unsuccessful ones. Initially there were plans for a whole series of physics-based level design elements - but one after another, they were cut in order to complete production on time. Most deserved their removal, as they simply overlapped with other mechanics, and did very little to reinforce the core ideas.
Initially the design document included three worlds, Blood, Brain, and Liver, symbolizing the progression of malaria through the body - however, one month before release, it was more important to polish the first two worlds, rather than add a third - especially the additional level design elements to justify more content were not complete.
One particular mechanic that should have been explored earlier in development was a timer. As it turned out, the game was a bit more difficult than initially planned, so once there was development time for exploration, it was apparent that time pressure would only exacerbate existing difficulty problems, and there wasn't time to rework the levels again. Nevertheless, adding a time element would have synergized well with the idea that the character's body is failing. A timer would also have created tension, without necessarily increasing difficulty, which could have been used to stabilize the difficulty curve. Since zero on-screen UI was a visual design goal, integrating a diegetic timer into the aesthetics of the game would have required a fair amount of planning.
Nightmare: Malaria launched on 'Giving Tuesday', December 3. In one week, there have been 102,000 players, playing 145,000 sessions, with an average session length of 5 minutes, 9 seconds. 75% of the players are using iOS, and 25% use Android - and more than half of the players are outside the US. There have been 40,000 clicks through to the Against Malaria Foundation.
The visitor flow clearly shows the initial release had difficulty progression problems, shown here as a steep dropoff at level 3. Version 1.1.2 modified the physics, as well as added intermediate levels to lessen the initial difficulty slope. Those changes are pending approval in the iOS app store, and have been pushed to Google Play. Hopefully these changes will improve both returning user numbers, as well as session duration.
In the next post, I'll examine some of the level design tools and processes we developed to create Nightmare: Malaria. Thanks!