Self Help

Spelunky - Derek Yu

Author Photo

Matheus Puppe

· 38 min read

Here’s a summary of the key points:

• Spelunky is a platformer inspired by roguelikes, a niche genre known for randomized levels, permadeath, and consistent rules for interactions.

• The roguelike genre began with Rogue (1980), which featured ASCII graphics, turn-based gameplay, and dungeon crawling. The definition of a roguelike is debated but usually includes randomized levels, permadeath, and shared rules for players/monsters/items.

• More recent roguelikes like Mystery Dungeon and Dwarf Fortress have added graphics and accessibility but maintained complexity. Dwarf Fortress in particular brought more mainstream attention to roguelikes around 2006.

• The author sees roguelikes as a childhood influence and was excited to see the genre gain more mainstream interest.

• The author compares the creative process to assembling a jigsaw puzzle, with pieces coming from various inspirations and life experiences. Sometimes a single missing piece can connect many other pieces; Spelunky was the result of many puzzle pieces coming together, including the author’s interests in platformers, roguelikes, and level design.

• The initial idea for Spelunky came from thinking about the randomization of Rogue combined with the platforming of Super Mario Bros. The first prototypes were made in Game Maker, using simple sprites and tiles.

• The random level generator was inspired by cellular automata, using simple rules to create complex results. It took time to refine the generator to create challenging but fair levels.

• Release of the initial prototype, called Spelunky Explorer, led to feedback that helped refine the design. The name Spelunky came from a sprite of a cave explorer.

That covers the key highlights from the beginning of the chapter on the origins and development of Spelunky Classic. Let me know if you would like me to clarify or expand on any part of the summary.

  • The author and his friend Alec worked together for two years to create their first commercial game, Aquaria, which was released in 2007. Aquaria was a success and won several awards, validating them as game developers.

  • However, the development of Aquaria was challenging and intense. The author and Alec had clashing work styles and personalities. After Aquaria, the author wanted to slow down but Alec wanted to start another big project immediately. They decided to part ways.

  • The author decided to use Game Maker, a simple 2D game development tool, to work on smaller projects. He was inspired by the short, creative games made by hobbyists in Game Maker. Making these smaller games felt more joyful to him, like planting a garden compared to planting an entire forest.

  • Game Maker also reminded the author of the first game making tool he used as a child, Klik & Play. Using Klik & Play at age 12, he made and released his first simple game, Trigger Happy, which received positive feedback and encouragement. This early experience hooked him on making and releasing his own games.

  • In summary, after the intense experience of developing Aquaria, the author wanted to return to smaller, more personal projects. Game Maker provided an outlet for that, allowing him to make short but creative games - his “gardens.” This return to smaller projects was a return to his roots, reminiscent of his early days using Klik & Play.

  • The author grew up playing games on old floppy disks he found from his parents’ graduate school days. Among these games were roguelikes like Hack and NetHack.

  • The author was fascinated by these old games, in part because their simple names and interfaces hid complex gameplay and possibility spaces within. The games felt like real adventures and mysteries to explore.

  • The author discusses Hack and NetHack in more depth. Hack, from 1985, has simple ASCII graphics but complex gameplay involving many keyboard commands and mechanics around eating, hunger, and survival. NetHack built on Hack, adding more content and tile graphics. The author shares some stories of emerging gameplay from NetHack, like rescuing and taming a troll, and the many ways you can be turned to stone by cockatrices. These games are praised for their vast possibility spaces and handling of so many possibilities elegantly without bugs.

  • The author admits that the only roguelike he remembers beating was Mystery Dungeon: Shiren the Wanderer. He was mildly disappointed to admit this to a questioner at a convention.

  • In summary, the author grew up with and was fascinated by some early roguelikes, especially Hack and NetHack, though he only remembers beating one, Shiren the Wanderer. These games made a big impact on him with their mystery, complexity, and possibility spaces.

  • NetHack is a popular roguelike that requires extensive knowledge and patience to ascend in. Ascending refers to retrieving an amulet deep within the dungeon and escaping. It can take 40,000 turns or more.

  • The author found NetHack too tedious to beat, preferring simpler roguelikes. After developing Aquaria, a complex game, the author wanted to make a simpler roguelike.

  • The author’s roguelike prototype had constrained item slots, action options, and small 15x15 levels to encourage simplicity. However, it didn’t add much new to the genre, so the author stopped development.

  • The author liked the action and tension of platformers but disliked repetition and memorization. The author liked the variety and meaningful death of roguelikes but disliked some repetitive tasks and memorization.

  • Combining platformers and roguelikes led to the idea for Spelunky. The earliest build from June 2008 already featured randomized levels, destructible environments, and the spelunking theme.

  • Almost all levels in Spelunky depend on the player destroying walls to access parts of the level. Destructible environments give players more agency and help with level generation.

  • An article about the 1983 game Kagirinaki Tatakai, where almost all levels were destructible, helped inspire destructible terrain in Spelunky.

  • The name “Spelunky” originally referred to the spelunking theme and setting. It was meant as a temporary name but stuck after positive feedback.

So in summary, the author aimed to combine the action and simplicity of platformers with the variability and consequence of roguelikes. Early builds of Spelunky already featured the key elements of randomized, destructible levels and a spelunking theme. The idea was inspired in part by an obscure 1980s action game with destructible environments.

  • Spelunky follows in the tradition of adventure stories about exploring dangerous places for treasure, like Indiana Jones. These stories inspired many video games.

  • Themes like this help convey abstract rules and make games more compelling. The theme of exploring caves fits Spelunky’s procedural level generation and emphasis on environmental destruction.

  • Spelunky’s level generation follows some rules:

  1. Levels go from top to bottom, with an entrance and exit.

  2. There is always a path between the entrance and exit that doesn’t require special items.

  3. The generator tries to avoid creating areas where the player can get stuck.

  4. It aims to seamlessly connect the edges of the 4x4 room grids that make up each level.

  • The level generator first picks a top room as the entrance, then creates a path to the bottom, marking each room as either on the path (1-3) or a side room (0). Side rooms may be inaccessible without destroying walls.

  • The destructible environment means side rooms can still be reached, giving players more flexibility in how they explore levels.

  • Following clearly defined rules and a consistent theme allowed Spelunky’s design to emerge naturally without compromising any part of it. The various elements of the game reinforced each other.

That covers the main points around Spelunky’s themes, level generation, and design. Let me know if you would like me to clarify or expand on any part of the summary.

  • The game Spelunky utilizes procedural level generation to create randomized levels. This makes level design easier since the developer only needs to ensure there is a path from the entrance to the exit. Side rooms that happen to connect are a bonus but not required. To players, dead-end side rooms will seem like regular walls.

  • Each room runs a script called “scrRoomGen” that builds the room from tiles. It first selects a random room template from 6-12 options based on the room type. Templates are strings of characters where each character represents something in the game, like walls, ladders, platforms, etc. For example:

0000000011 0060000L04 0000000P11 0000000L11 0000000L11 0000000011 0000000011 1112222111

  • “Chunks” are groups of 5x3 tiles that provide more randomization. There are chunks for floors, ceilings, vines, etc. The script places chunks in the room based on the template.

  • Another script called “scrEntityGen” populates the level with monsters, traps, items and treasure. It checks each tile and randomly decides whether to place an entity there based on percentages. The percentages were determined through trial-and-error. For example, in the Jungle, there’s a 1/60 chance of a mantrap, 1/60 chance of a caveman, 1/120 chance of a fire frog, and 1/30 chance of a regular frog.

  • The developer worked on Spelunky in secret over the summer of 2008. He felt pressure to release it but it was still buggy. He posted about it on TIGSource, an indie game developer community, to get feedback. He had been following TIGSource since 2005 after becoming disillusioned with the mainstream games industry. TIGSource showed him the thriving indie scene.

  • The randomized level generation is an example of the kind of programming the developer enjoys - simple but engaging. Though unrealistic, the levels are fun and avoid trapping the player.

That covers the key points on how Spelunky generates randomized levels and populates them with entities as described in the passage. Please let me know if you would like me to clarify or expand on any part of the summary.

  • The author developed the game Eternal Daughter using Multimedia Fusion, but the tool and its community stagnated over time. When the tools’ limitations were reached, the community also declined.

  • The author took over the indie gaming blog TIGSource after its founder had to abandon it. Through TIGSource, the author connected with other indie developers and helped define the “indie spirit.”

  • The author developed the game Spelunky and first shared it with a small group from TIGSource. Their enthusiastic feedback spurred the author to quickly fix bugs and add new content.

  • An important addition was the shopkeeper and shops. Shops are common in games but Spelunky’s shopkeeper made them dangerous. Players are tempted to steal valuable items but attacking the shopkeeper is very risky, especially for new players.

  • The shopkeeper has four main states: idle, follow (when items are picked up), attack (when items are stolen), and patrol (in wanted levels). The shopkeeper is adept at finding the player in attack mode, though the AI is basic. Shoplifting has high risks but high rewards, testing the player’s boundaries.

  • Experienced players seek to find the limits of a new game, even if it means frequent death. The author aimed for Spelunky to have many high-risk, high-reward scenarios to tempt such experienced players.

The shopkeeper’s AI uses simple rules to get around and attack the player:

  • Occasionally turn towards and shoot at the player if close enough
  • Jump over obstacles or up to the player’s level
  • Do nothing if the player is below and within range

The jumping behavior makes the shopkeeper seem erratic and dangerous. This shows how simple mechanics can achieve thematic goals.

Releasing the game publicly led to a mix of reactions. Some loved it, some wanted to love it but found it too hard, and some already disliked it. Negative feedback stands out and is hard to take, especially publicly and personally. However, it’s important for improving the game.

One angry player’s feedback showed the game was too hard, but not for the reasons assumed. Fixing other issues actually made the game easier without compromising difficulty. Feedback should be analyzed like a doctor diagnosing a patient. Players know their symptoms but not always the underlying cause. Other developers can also miss the real issues by focusing on details.

Each player’s feedback tested and shaped the developer’s vision for the game. A vision evolves and solidifies over time with input. Spelunky’s vision became more solid through interactions between creator and audience.

In summary, the shopkeeper AI uses simple rules to seem complex. Public feedback, especially negative, is hard to take but important for improvement. Interpreting feedback requires determining underlying causes, not just surface issues. A game’s vision develops and strengthens over time with audience and creator interactions.

The author discusses their lifelong passion for video games, which began in the 1980s with Nintendo and arcade games. They were particularly inspired by The Legend of Zelda and its sense of exploration and discovery. However, the author feels that recent entries in the Zelda series lack that sense of mystery and feel too engineered, comparing them to theme parks.

The author argues that for an adventure game to feel truly adventurous, the game world must feel indifferent and self-sufficient, not solely designed around the player. They cite an essay called “Saving Zelda,” which argues that Hyrule should ignore Link more and have its own sense of place. The author agrees, saying that if a world only exists for the player, both the world and the hero will feel artificial.

When creating Spelunky, the author tried to recapture the feeling of real discovery by making the game world partially random and uncontrollable. They focused on their own vision instead of what others wanted, using the simple Game Maker engine to concentrate on game design. The random elements ensured that even the creator couldn’t fully control the player’s experience.

Like Shigeru Miyamoto, the author spent a lot of time outside as a child, but was most attracted to virtual adventures like The Legend of Zelda. They also frequented arcades, which evoked a sense of danger and chaos alongside the flashing lights. The author sees arcades as liminal spaces between the virtual and the real.

The key ideas are:

  1. Recent Zelda games lack the sense of mystery and adventure of the early entries. They feel too engineered and theme park-like.

  2. An adventure world should feel indifferent and self-sufficient, not solely designed for the player. If it only exists for the player, it will feel artificial.

  3. Spelunky aimed to recapture a sense of real discovery through randomization and focusing on the creator’s vision.

  4. Arcades represent a liminal space between the virtual and the real, evoking a sense of chaos.

  5. Like Miyamoto, exploring virtual worlds and real worlds as a child inspired the author’s game design.

  • Arcade games were often violent, strange, and sexualized. The author visited arcades as a child and played violent games like The Punisher and Mortal Kombat.

  • Arcade and home gaming experiences were very different. Arcade games were meant to be played in short bursts, usually 30 minutes to an hour, while home games could last hundreds of hours. Arcade games had to be immediately engaging or people would stop playing quickly.

  • The author became frustrated with mainstream home games and preferred indie games and games from his childhood. He particularly liked the challenging nature of arcade games.

  • In developing Spelunky, the author aimed to recreate the arcade experience without worrying about quarters. Beating Spelunky takes around 30-60 minutes, similar to an arcade game. The ghost enemy induces time pressure like in arcade games.

  • Timers were common in arcade games to prevent one player from dominating a machine and to keep crowds moving. Hard timers kill the player immediately while soft timers make the game harder over time. Timers add challenge and force players to make quick decisions.

  • The arcade model had downsides, like the poor continue system that allowed players to keep putting in quarters to continue. This gave arcade games a reputation as “quarter-munchers” that were greedy. However, good arcade games could be beaten on a single credit, showing they rewarded skill.

  • High score systems showed that arcade games rewarded skill and mastery. Your score reset when you continued, so only achieving a 1cc or one-credit clear allowed you to get a high score. Experienced players avoided continues to get 1ccs.

  • The author made the mistake of credit-feeding in Metal slug: 3, spending many quarters to make little progress. Continue systems allowed casual players to see more of the game but were meant to extract money. Regular players still spent many quarters but over more time.

  • In Spelunky, the author aimed for a middle ground. He avoided hand-holding and the opaque continue system of arcades but allowed progress through the Tunnel Man, who offers shortcuts to later levels so players can practice.

  • The entrepreneur Tunnel Man will build shortcuts between areas of the game for a price. Players can donate money across runs to eventually unlock the shortcuts.

  • The designer wanted players to realize that the game is best played from start to finish without shortcuts, but didn’t want to explicitly tell them. The solution was twofold:

  1. The basic design of the game incentivizes playing from the beginning. Players can get essential items and prepare better for later levels. Skipping levels means missing these opportunities.

  2. A secret chain of events called “The Chain” makes shortcuts less useful in the long run. It culminates in accessing a secret level called the City of Gold.

  • The Chain consists of four major items, one in each area of the game, used to obtain the next item. The first is the Udjat Eye, found in a chest in the Mines. It reveals the entrance to the Black Market in the Jungle.

  • The Black Market is an expanded shopping area with seven shops. It always contains an Ankh, which revives the player after death and is required for the next part of the Chain.

  • Dying with the Ankh in a level with a Moai statue transports the player inside the statue. There they find the Hedjet, which prevents dark levels and grants access to the City of Gold.

  • Each item in the Chain has a purpose beyond progressing to the next one. The Udjat Eye reveals hidden items, the Ankh revives the player, and the Hedjet prevents dark levels. This avoids the “too many locks and keys” problem where progress feels artificial.

  • The Chain was designed to feel integrated into the world, not like an obvious sequence of secrets. This makes its discovery more meaningful. The designer focused on integrating the secrets, not on how many players would find them.

  • The Moai statue and Ankh in Spelunky require skill, luck and dedication to obtain.

  • Obtaining the Hedjet by sacrificing the Ankh at the Moai shows mastery of the game.

  • Defeating Anubis to get the Scepter, which unlocks the City of Gold, is a tremendous achievement.

  • The City of Gold, inspired by myths of El Dorado, contains gold that can raise your score tremendously.

  • Reaching the City of Gold shows mastery of Spelunky’s mechanics and uncovering a hidden secret.

  • The Chain of items exemplifies Spelunky’s tensions and discoveries. Players can discover it gradually through play.

  • The Chain provides challenge without difficulty settings. Difficulty settings have problems: players must choose without information, and some will choose an easier setting they don’t need.

  • In Spelunky, players start by unlocking shortcuts, then beat Olmec, then can try the Chain. Everyone plays the same game.

  • The Chain provides permanence in a random landscape. Randomness gives longevity, but permanence makes the world feel real.

  • Players today want to be comfortable consumers rather than art contributors. They want games to “play the way they want” rather than appreciating a creator’s vision.

  • The best games come from mutual respect between creators and players. Creators have a vision; players trust and are guided to it.

  • Nintendo taught players through level design, not hand-holding. Spelunky guides players with features like the ghost and City of Gold.

  • The joy of discovering a creator’s vision through play is fundamental. It’s revealed gradually, not spelled out. This extends to cuisine, where omakase means trusting a chef’s vision.

  • Fast food’s “have it your way” is the opposite: indulging players rather than guiding them.

In summary, the key ideas are: 1) the Chain requires mastery and uncovers a secret; 2) difficulty should be built-in, not chosen; 3) the best games have a creator’s vision revealed through play; and 4) this “joy of discovery” should guide players, not indulge them.

  • In 2008, Jon Blow released Braid, a successful puzzle platformer, on Xbox Live Arcade. He offered to help Derek pitch Spelunky to Microsoft for XBLA.
  • Derek was initially hesitant but became interested after seeing other successful indie games on consoles like Castle Crashers and World of Goo. He put together a design document outlining some new features like Challenge Mode, Endless Mode, and local co-op multiplayer.
  • Microsoft greenlit Spelunky for XBLA but was concerned about Derek completing it himself in a year. Derek foolishly claimed he could program it himself if needed.
  • Derek’s friend Andy Hull, who he knew from an online game making community, offered to help with programming. Despite Andy being less experienced, Derek chose to work with him because they had a good relationship and team dynamics.
  • Derek compares finding teammates to dating, emphasizing the importance of shared vision, communication, passion, and long-term commitment. For a commercial game, teammates matter even after release in marketing, updating, porting, etc.

In summary, Jon Blow helped Derek pitch Spelunky to Microsoft for XBLA, Microsoft approved it but was skeptical Derek could finish alone, and Derek chose his friend Andy as a programmer over more experienced options because of their good team dynamics. Derek believes having a good team is critical for commercial game development.

The author had some difficulties in explaining Spelunky’s vision and appeal to newcomers, especially to the mainstream Xbox players. Spelunky was originally designed to be challenging and unforgiving, with the levels and events generated randomly. Much of the joy of playing came from discovering secrets and overcoming difficulties by yourself. The author was worried that adding explicit rewards and hand-holding systems, as suggested by the producer Kevin and other Xbox managers, might undermine this core experience.

The author agreed with the view that “tangible, expected, contingent rewards reduce free-choice intrinsic motivation.” Providing rewards for already fun activities may reduce people’s natural motivation and turn the experience into a grind. The author was concerned that systems like leveling up, experience points and collectibles might do more harm than good.

However, the author thought an “adventure journal” might work. It could track interesting stats and facts about players’ accomplishments to motivate them in a less superficial way. The journal could build on the existing player profile and collect more stats, e.g. how many times players have played and died. The author outlined a plan to flesh out the journal to make it a motivating feature without undermining the game’s spirit.

In summary, the key tension was between staying true to the game’s vision and appealing to a mainstream audience. The author wanted to avoid excessive hand-holding but also needed ways to motivate casual players. A well-designed adventure journal seemed a promising solution, providing motivation through tracking accomplishments rather than giving out rewards. Overall it’s about balancing the needs of players who already love the game and those who need more encouragement to stick with it.

Early on in the development of Spelunky, Derek Yu made several key decisions to improve the game for Xbox Live Arcade:

  1. He added a journal system to help guide new players. The journal belonged to an explorer named Yang and contained tutorials, hints, and lore about the world. However, Yu didn’t want the journal to hold players’ hands too much, as that could hurt the experience for more skilled players.

  2. Yu struggled to get Spelunky running on the Xbox 360 dev kit. He enlisted the help of Andy Hull, the lead programmer of Spelunky, and other engineers to troubleshoot issues. Eventually, they decided to ditch the Braid engine they were using and build a custom Spelunky engine from scratch. This was a difficult decision but allowed them to understand all parts of the code.

  3. Yu saw the Xbox Live Arcade version of Spelunky as a “remake” or “fan game” of the original PC version. Remaking an existing game allowed for more creative freedom than developing an entirely new idea. Yu was eager to improve the game’s graphics and add new features while keeping the core mechanics intact.

  4. Some of the improvements Yu made include:

  • Updating the pixel art graphics to high-definition 2D art.
  • Fleshing out the game’s lore and backstory.
  • Adding new items, enemies, terrain, and other features while balancing the difficulty.
  • Streamlining the interface and menus.
  • Polishing the level generation to create more interesting layouts.
  • Improving the game’s music and sound effects.

In summary, Yu stayed true to the spirit of the original Spelunky but essentially remade the game to take full advantage of the Xbox 360’s capabilities and appeal to a wider audience. The end result was a massive improvement over the first version of Spelunky.

  • The pixel art style of the original Spelunky was outdated and didn’t suit modern HD displays. The developer wanted to update the graphics to HD for the reboot.

  • The movement and controls in the original Spelunky were also outdated, feeling “floaty” and hard to control. The developer wanted to improve the controls to feel smoother and more responsive, especially the jumping. Improving the controls made the game easier to get into without making it easier overall.

  • The developer upgraded the enemies by making them faster, smarter, and adding more difficult variants. The level generation was also updated to increase the difficulty and add more enemies on later levels. This made the game more challenging overall despite the control improvements.

  • The developer aimed to make the world feel more “alive” by adding new monsters and variations, even if their AI was simple. The human mind is good at finding patterns, so adding variations and relationships between monsters created the impression of an ecosystem and societies. Even simple behaviors combined can feel complex and coordinated.

  • An example story shows how simple AI behaviors combined to create an emergent complex situation. A tiki man picked up a boomerang in a shop and walked out, causing the shopkeepers to go berserk and attack everything around them. Though the systems involved were simple, they interacted to create a memorable story.

  • Randomness and procedural generation provide “free value” by creating endless variations, but to be truly compelling they need to be combined with a complex, interconnected system of rules and behaviors that is worth understanding. When done well, even simple stories and events within that system can become meaningful.

  • The soundtrack for the original Spelunky was provided by George Buzinkai and Jon Perry. After enlisting them, the developer was approached by Eirik Suhrke to provide music for the reboot.

• Eirik Suhrke is a Norwegian composer who co-ran an independent music label and community called Pause.

• Derek Yu initially wanted dynamic music that changed based on gameplay for Spelunky on XBLA but scaled back that idea due to limited resources.

• Eirik aimed to mix FM synthesis, sample-based tracker music, and live band music to match Spelunky’s retro yet modern feel. The music received some criticism early on, especially the jazzy tracks for the Mines. Eirik adapted by pulling back on some elements, especially in early areas.

• The final soundtrack has 47 tracks and, like the game, took time to win over players but is now beloved. One track, “Adventure Begins,” was removed due to criticism but later restored after players requested it back.

• Derek believes finishing projects is an important skill. He wrote a guide on finishing games with 15 tips, including choosing a good idea, starting work, prototyping, focusing on fun, managing partners or working alone, anticipating grind, using deadlines, pushing forward, maintaining health, avoiding endless restarts, saving ideas for other projects, cutting content when behind schedule, scaling down rather than up if quitting, and recognizing the challenge of finishing the last parts of a project.

• Derek values finishing as a skill that has enabled many opportunities in his career, starting with his first finished game, Trigger Happy.

  • The author started making small games with a program called Klik & Play, which led him to create bigger and bigger games, culminating in Aquaria and Spelunky.

  • Although the author abandoned many projects along the way, finishing and releasing games gave him motivation, confidence, and experience. Finishing a game is hard, but it gets easier with each game completed.

  • The author observed that talented creators often struggle to finish projects because they are perfectionists who frequently restart their work. Repeatedly redoing early parts of a game makes it less likely the full game will be completed.

  • For Spelunky, the author and his partner Andy did little pre-planning or tool-building, instead focusing on simply moving forward and completing the game. While Andy focused on technical details, the author focused more on design and keeping the project progressing at the right pace. Their collaborator Kevin provided high-level guidance.

  • The author originally planned additional modes for Spelunky like Challenge Mode and Endless Mode, but cut them to focus on completing Adventure Mode and multiplayer. Multiplayer was intended to be a casual, enjoyable experience rather than a competitive one.

  • In summary, the key to finishing and releasing Spelunky was avoiding perfectionism, redundant rework, and feature creep; keeping the scope limited; and maintaining steady progress toward completion. Releasing the initial game provided motivation to continue improving it.

The designer aimed to make the multiplayer experience accessible for players of all skill levels. He introduced two main features to accomplish this:

  1. Allowing players to pick up and carry each other. This enabled skilled players to help less skilled players through difficult areas. It also allowed players to benefit from items like jetpacks and capes even if only one player had them. This made the multiplayer experience more collaborative.

  2. Allowing dead players to come back as ghosts. This prevented players from becoming spectators for long periods if they died. Ghosts were limited but could still interact in minor ways, like blowing puffs of air. Living players could revive ghosts by breaking open coffins. This allowed players to keep reviving each other and stay in the game.

These changes made the multiplayer easier by allowing players to help and revive each other. However, they were balanced by limiting resources like bombs and ropes, the possibility of players harming each other accidentally, and the ability for players to kill or grief each other intentionally. The designer saw this as fitting with the game’s dark humor and mechanics.

The damsel-in-distress mechanic allowed players to regain health by rescuing a helpless woman and carrying her to the exit. The designer acknowledged this reinforced a stereotypical trope but saw it as fitting with the game’s pulp adventure theme and mechanics. In the final release, a “changing room” allowed players to play as the damsel after rescuing enough of them. However, the damsel retained her original helpless depiction.

The designer received criticism about the damsel mechanic reinforcing sexist stereotypes. However, some saw it as silly and irreverent rather than deeply problematic, especially given the game’s light tone. The designer understood the criticism but thought the mechanic achieved the right tone for this game. However, he aimed to add more diversity to the characters in the XBLA version.

In summary, the designer made several choices to facilitate an accessible multiplayer experience. However, the damsel mechanic received justified criticism for employing a sexist stereotype, even if done in a lighthearted way. The designer took this criticism seriously in aiming to add more diversity to later versions of the game.

The author decided to add more character diversity to Spelunky for diversity’s sake. This included adding new player characters of different genders and races, like a Sikh character and a female character inspired by Marion Ravenwood from Indiana Jones. For the damsel in distress, the author added options to rescue a “hunky” man or a pug dog instead of a woman. These additions were generally well-received, especially the pug damsel.

However, Anita Sarkeesian criticized the game in her Tropes vs. Women in Video Games video series. She said that while the additions were good, the female damsel still reinforced negative stereotypes about women, while the male and dog damsels did not do the same for men and dogs. The author appreciated the criticism and said it made him want to think more about the issue, even if others reacted negatively to Sarkeesian’s comments.

The author discusses how easy it is to rely on cliches and stereotypes in games without realizing it. He wants to ask more questions about these kinds of choices in his own games going forward. While art should explore offensive ideas, not every idea in a complex game is equally vital. The damsel could easily have been something other than a stereotypical princess.

The author’s perspective has also been influenced by having a young daughter. He notices issues like the tendency to default to male pronouns for ambiguous characters more, but has also seen more games, toys, and media with multidimensional female characters recently. He appreciates those who challenged his designs in Spelunky, as it has helped open his eyes to these kinds of problems.

In the end, the author says that the last 10% of work takes 90% of the time. Fixing issues like stereotypical or cliched designs and characters may seem small but can require a lot of effort. Overall, the summary conveys the author’s journey toward becoming more aware of diversity and representation issues in his own work, and his commitment to continuing to improve in the future.

• Project estimates nearly always underestimate how long it actually takes to finish a project, especially the final stretch. This is because we make estimates assuming perfect, uninterrupted work with no setbacks, which is unrealistic.

• Unexpected tasks, like polishing, debugging, and localization, add up and consume a huge amount of time. These tedious but important tasks are easy to overlook when making an initial estimate.

• Releasing on consoles requires an enormous amount of additional work to meet the platform’s technical requirements and policies. This includes extensive testing, localization, and ratings submissions which can take months.

• As deadlines approach, the remaining work seems to expand rather than contract. Anxiety sets in as you realize nothing is truly “good enough” and everything needs to be perfect.

• It’s often better to underestimate how long something will take, as a shorter timeline is more motivating and helps ensure the project gets off the ground. However, this needs to be balanced with the determination to see the project through to completion, no matter how long it actually takes.

• Expectations for how well a game will sell are hard to set accurately. Even highly-anticipated indie games may struggle to match the success of past hits. The reception of a game ultimately comes down to capturing players’ interest at launch, which is hard to predict.

In summary, finishing and releasing a game is an enormous amount of work that almost always takes longer than expected. But some amount of unrealistic optimism at the start of a project can be helpful, as long as you commit to doing whatever it takes to complete your vision. Success is hard to guarantee, but giving players a polished, compelling experience at launch is the best way to build an audience.

  • Spelunky sold only 17,818 copies in its first week and 33,107 copies in its first month on Xbox Live Arcade. This was significantly less than comparable indie games like Limbo, which sold 78,044 copies its first month, and Fez, which sold 76,312 copies its first month.

  • There were many theories as to why Spelunky underperformed, including its high price point of $15, lack of marketing, the free demo satisfying players, and the declining popularity of Xbox Live Arcade for indie games.

  • The developer, Derek Yu, was disappointed that Spelunky did not sell as well as expected but was still proud of the critical success and cult following the game built. Sales, while not blockbuster, were still decent.

  • Development began on a PC version of Spelunky for Steam to release the game on a new platform and have another chance at major success. The conversion from Xbox to PC was easier than the initial development.

  • Derek Yu received an email from developer Doug Wilson describing a unique way of playing Spelunky with another developer, Nifflas. They would each play one turn of Spelunky per day while the other watched, raising the stakes by limiting themselves to a single daily run.

  • This style of play intrigued Derek Yu and matched ideas Spelunky’s developers had experimented with before, like a version where the game would delete itself after the player died for the first time. These radical ideas were hard to implement on Xbox but seemed more viable on the open Steam platform.

  • Steam was an attractive platform for Spelunky’s PC release due to the lack of strict requirements, easy updating process, and reports of strong sales from other indie developers. The Spelunky team was optimistic about finding success on PC after a modest debut on Xbox Live Arcade.

  • The author dislikes rigid boss battles with patterns that must be solved. They prefer bosses that feel like upgraded enemies and can be dealt with creatively. Spelunky’s mini-bosses like the giant spider and Yeti King are examples of this.

  • Olmec, Spelunky’s final boss, is different. He is very large and impervious to normal attacks. The only way to defeat him is to get him to fall into the lava below. This can be done by bombing a hole for him or exploiting his attack where he leaps up and then slams down, destroying floor tiles.

  • The Olmec fight uses destructible terrain and has a lot of variety based on the player’s equipment and other enemies. Although the goal is the same, the battle can play out many ways. Olmec’s size also makes the fight intimidating.

  • In the original Spelunky, beating Olmec ended the game. For the remake, the author wanted to add a secret link in the chain and a “true final boss.” They decided the player should go to Hell to fight this final boss.

  • The author’s concept of Hell was influenced by the Bible and popular culture but also other religions and mythologies. Their Hell is tailored to Spelunky, with lakes of fire, demonic enemies, and Satan as the true final boss. Beating Satan completes the Chain and is the ultimate accomplishment in Spelunky.

The key points are that the author aimed to make more creative boss fights, added a secret final area (Hell) and boss (Satan) to extend the game for experienced players, and drew from various mythological and cultural concepts of Hell to create Spelunky’s version.

  • The author chose Di Yu, the Chinese hell, as the punishment for nonbelievers in Spelunky. Di Yu allows for reincarnation and fit better with Spelunky’s design.

  • Hell would be the fifth area after the Temple. To get there, the player needs the Book of the Dead, which connects ancient Egypt and Hell. The Book of the Dead reveals the door to Hell in Olmec’s level.

  • Entering Hell requires standing on Olmec’s head as he sinks into lava. Hell’s visuals change to bright red. It features skulls, lava, traps, and demons.

  • Though Hell is difficult, the Jungle is actually the hardest area. Difficulty should ebb and flow rather than steadily increase. Players get mentally and physically tired, so difficulty should account for that.

  • Hell preys on the player’s emotions and anxiety. It looks scary and claustrophobic, the music is fast-paced, and succubi appear as damsels.

  • King Yama seems tricky but is more forgiving than Olmec. His minions and obstacles can be avoided, and he is vulnerable to bombs and shotguns. Beating Yama allows you to rescue Yang.

  • Though Hell seems very challenging, it may not be as hard as it appears, especially coming at the end of the game. The difficulty is meant to test the player’s emotions and decision making rather than just their reflexes.

  • The mystery box in Spelunky costs $8,000 but contains a random item that could either be very valuable or worthless. However, if you carry an unopened mystery box to an altar of Kali, you will receive a purple eggplant and unlock a secret song.

  • The eggplant was added late in development as a nod to “Totaka’s Song,” a simple tune that composer Kazumi Totaka hid in many Nintendo games. Derek Yu wanted to hide his own song, “Eirik’s Song,” in Spelunky in an obscure way. After rejecting many ideas, sacrificing a mystery box was chosen.

  • The eggplant was given as a reward to tie into a tradition of video games giving strange and unexpected rewards. The eggplant in particular has a history of appearing in many classic games, like Ice Climber, Wrecking Crew, and Kid Icarus.

  • To minimize work, Derek Yu made it so sacrificing the eggplant to Yama, the final boss, would turn Yama’s face into an eggplant and play “Eirik’s Song” again. He only briefly tested this and did not anticipate players wanting to do more with the eggplant.

  • Players spent months trying to figure out the purpose of the eggplant through experimentation. Eventually, some hypothesized bringing it to Hell, but doing so seemed nearly impossible in single player due to needing to carry both the eggplant and scepter. Using a hired hand to carry one item could work, but the chances were very small.

  • Hired hands were added to make the single player experience less lonely but were unreliable and often did more harm than good. Beating the game with their help was seen as an extra challenge.

The key points are:

  • The eggplant and song were an obscure secret added late in development

  • The eggplant has a history of strange appearances in games and was meant as a silly bonus

  • Players determined bringing it to Hell was key but thought it was nearly impossible in single player

  • Hired hands were unreliable but beating the game with them added extra challenge

  • Having AI companions in games, though often troublesome, can feel rewarding when you’re able to keep them alive. In Spelunky, the hired hands are difficult to manage but an incredible technical feat. Giving them weapons or making them carry fragile items like the eggplant is dangerous.

  • Reaching Hell with an eggplant and a hired hand requires a lot of luck and chance encounters. The eggplant itself is easily destroyed, and the requirements to get one to Hell in single player seemed nearly impossible.

  • Players hacked the game to first reveal the eggplant easter egg in Hell. Two months later, the first legitimate eggplant run to Hell in co-op mode was completed.

  • A solo eggplant run was thought to take years, but top player Bananasaurus Rex completed one in a month by exploiting glitches. He would restart until getting an “eggplant seed” level with the right setup.

  • One glitch lets you break into the Moai with the ball and chain item, leaving the eggplant inside. You can then kill yourself, use the Ankh to respawn and get the hedjet, pick up the eggplant again and proceed. Rex also used a glitch to destroy the ball and chain when he no longer needed it.

  • To bring the scepter from 4-1 to 4-2, Rex freed the robot unlocked character to carry the eggplant instead of a hired hand. Unlocked characters normally only appear once, but resetting the save file brought them back. The robot was the best choice, found in the Mothership level.

  • The run involved a lot of skill and close calls to complete. Rex was ultimately able to get the eggplant to Hell in an incredible feat.

  • Spelunky is filled with secrets, Easter eggs and details that give the game personality and connect players. Creator Derek Yu added many secrets, even though few players would discover them.

  • Yu started making games as a teenager. Spelunky represents a period of his life and partnerships with others. He feels part of a bigger puzzle connecting influencers, players and those Spelunky has influenced.

  • Yu sees game development as putting together a puzzle with pieces representing thoughts, experiences and inspirations. The puzzle gets bigger with each game.

  • Spelunky blends games from individuals and big companies. Yu couldn’t imagine solo success when interviewed in 2006 but the indie scene grew.

  • Yu’s influences include roguelikes, Miyamoto’s views on exploring and “loops,” and a program that generates Spelunky levels.

  • Yu announced Spelunky on TIGSource and got feedback to improve it. Indifference to his early games motivated him. Miyamoto was also motivated by initial indifference.

  • Yu aims to teach through level design and give players joy in discovering how the game works. But players approach games differently, so outcomes are unpredictable.

  • Yu’s friend Andy helped finish Spelunky, using the “right” amount of technology to avoid delays. They wanted to make the multiplayer bombastic.

  • Yu changed the damsel in distress to be more inclusive after feedback. He generally tried to exceed expectations, though the “last 10%” of polishing took as long as the rest.

  • Yu sees each game as a new adventure, with more connections and secrets to discover. The landscape for developers has changed a lot since his early days.

Here’s a summary of the key points:

  • Spelunky sold poorly relative to Braid and FEZ in its first month. The article “Is XBLA Past Its Prime?” suggests the XBLA platform was declining in 2011. However, Spelunky grew into a cult hit over time through word of mouth and developed a dedicated speedrunning community.

  • Derek Yu describes Spelunky as a “disposable game” inspired by roguelikes that generate random levels. The randomization and permadeath create a lot of replayability.

  • The game is notorious for its difficulty and how long it takes players to progress. Only a small percentage of players have earned major achievements like beating King Yama. Players like Bananasaurus Rex have pushed the game to its limits through feats like solo eggplant runs.

  • The development of Spelunky was a collaborative effort. Many playtesters provided feedback, and programmers Andy Hull and Eirik Suhrke were instrumental in realizing Derek’s vision. The game also benefited from the support of other indie developers like Jonathan Blow.

  • A number of streamers and YouTubers helped bring more attention to Spelunky over time, especially through promoting the Daily Challenge mode. The speedrunning community has also kept interest in the game alive and pushed players to improve their skills.

  • In the end, Spelunky became a cult classic that spawned a larger franchise. Despite a slow start, word of mouth and the dedicated fanbase allowed the game to find success over the long run. Derek Yu pays tribute to all those who helped make Spelunky a reality.

Does this summary cover the key points accurately? Let me know if you would like me to clarify or expand on any part of the summary.

Author Photo

About Matheus Puppe