Hazelnut Bastille Developer Talks About Building Upon The Great Design Of Legend Of Zelda



    Hazelnut Bastille places a lone heroine on an island of dungeons, danger, and potential allies. Offering adventure trough top-down action and puzzle solving, a crafting and trading system, and a series of weapons and unique tools, the game looks to have players enjoy a game that builds upon many of the design lessons of classic titles.

    Siliconera spoke with the developers of Hazelnut Bastille to learn more about what the developers have drawn from gaming’s past, and the developers were more than happy to share many details of what classic games have shown them, what they have taken from these lessons, and how they are building upon them and making them their own.


    What made you interested in crafting your own top-down dungeon crawler? What did you want to bring to this genre?

    Dennis Vavaro, Artist/Writer for Hazelnut Bastille –  So, let’s do a bit of scene building. 1990: I am five years old. My parents and I are living for a few months with a family friend of my mother, because our home burned down while we were all away one day due to a wiring error in the stove range. In middle of the floor, next to an upholstered cat tree, sits an NES with The Legend of Zelda already in the console.

    Not a few hours go by before it is staring back at me from the big screen. I look on as my family begins to explore Miyamoto’s cryptic garden of mysteries. To be fair, I had already seen a handful of games, but this was different. No one was telling you where to go. The character got gradually stronger over time. There were places where clearly you were meant to go, but somehow you didn’t have the means to get there (people often forget that LoZ is one of the very first Metroidvanias).

    Then there was the combat… rooms full enemies that dominated the screen, and together with scene geometry, restricted the movement choices of the player to the scenario the designers intended each time. Over the next 6 years or so, I think I must have beaten the LoZ over 40 times. I played it during that critical developmental period a young child has where they are learning the tasks they need to memorize to survive in their hunter-gatherer setting… as a result, I can recall the entire world map in my head from memory. Like a concert pianist, 20-25 years later I can picture the entire sequence of the game from start to finish, if I close my eyes.

    So that is a detailed autobiographical explanation of the personal connection people from our age group have with these games. But in the adult world of user-experience-design, creative deadlines, and finite resources, and where modern game engines can bring nearly anything to life, why did we choose this particular scheme of game? Something we have often cited in the past is the “scope vs depth” balance. There is a tension between what users demand, and what the surprisingly conservative games industry wants to provide. Users demand innovation and very wide scope. They want to experience something they have never experienced, and they want to do many things while there. The industry wants to produce low-risk projects with predictable development costs and financial outcomes.

    In some ways, even as an indie, we made a similarly conservative decision to work on a project with a predefined scope, where we could enumerate the steps and thus have a realistic view of the whole development cycle from start to finish. And despite this narrow scope being the opposite of what players consciously seem to demand, it allows us to create the thing most people want without knowing it: Depth.

    The indie games sector is now infamous for hooking users with seductive art, and then years later delivering an experience with no real impact or design. In a way, this is even to be expected: the things that make an indie game marketable and the things that make it a rewarding user experience have almost no overlap. This is the great tragedy of indie. It is very easy to make a game that resembles successful games of the past. Very few indies have the understanding and background to actually create something that successful on a design level to match the art.

    By working in a well-defined scheme of game with many precedents, and with an art style that is very quick to produce, we are able to shift our time heavily into being design-focused. There must be time to create assets, and reject them when they doesn’t serve the whole. There must be time to iteratively craft multiple level layouts, and determine the weaknesses of each. There must be time to pass the users sample material, and study their experience of it, not merely as a bug-testing QA, but as a design tool.

    So we have chosen this format of game because it has been developed over decades, with numerous ideas to build on and learn from, and because its art-style is simple to implement, and represents ideas as basic symbols. These factors allow us to focus all of our energy feverishly into design rather than art and assets, which is where it must be in order to produce a memorable user experience.

    Hazelnut Bastille places players in a world full of people to meet. Who are some of the characters they will come across, and what help will they offer the heroine?

    The player finds themselves amongst a group of people living on an island as expats. Common to everyone is that they all left their society of origin to escape something that also defines them on a deeply personal level. It is a little like the “Island of Misfit Toys”. Everyone has something they are hiding from, something they need, and perhaps even some measure of redemption to be gained. There is the great military hero of towering resolve, who was always absent, and unable to deliver those closest to him …the great social observer, who deeply understands others but can’t live amongst them… the natural mother-figure whose core womanhood is rejected… the creatives who are unwanted in a society of scarcity, and several more.

    The heroine shares this same mold, and is very much an antihero. Perhaps her great redeeming character trait is her self-reliance and bewildering resilience, but these are in fact largely born out of the Hellenic cleanse of tragic desperation. She arrives on the island by chance during the middle of a funeral honoring a lost hero. By some bizarre turn of events, that very night she takes up the impossible burdens of that same hero.

    Her unstoppable drive is mistaken for selfless charisma, and this energy quickly becomes a binding force between her and everyone she meets. At some point, she must reckon with the fact that her own goals, and what everyone wants from her are fundamentally incompatible.



    What drew you to add a crafting system to this adventure? What do you feel it adds to the dungeon-crawling experience? To games in general?

    The characters, beyond their involved back-stories, are also a hub of interest for another reason. Rather than go with the standard currency of many games in the genre, we chose to go with a trading-based resource system. The player, over the course of play, collects a number of commodities. These generally have immediate uses, or are at least precursors to things that are useful, so they are good to have in and of themselves, having intrinsic value. Every NPC has need of certain commodities which they are willing to trade for some rarer or more exotic items as well.

    Some of these rarer items may also be combined with other items to craft things like potions, special weapons, protective charms, and many other useful tools. There are also some very special items which can only be obtained after trading-up with various NPCs a half dozen times, which in turn may be priceless to someone else in the world. The trade mechanic is one of the main means the player has for breaking the ice with certain characters, and advancing their storylines together, and to open up quests and other features.

    On a design-level, this trading system has several uses for us. It creates a sense of provisioning and preparation. Oftentimes, the player will find a situation they cannot overcome except by going back and preparing for it by trading for and crafting the items they needed. They wouldn’t have sought these items earlier, because resources are quite finite, so they would have been used on what the player perceived they needed at the time. This need bolsters the notion that the player is taking an active role in overcoming the challenges of the world, rather than merely going through the locations in sequence, because B follows A and C follows B.

    A system with so many options also allows the player to express their unique play style: do they want to rely on superior weapon consumables, or on health potions and protective charms? Would they like to spend a huge number of resources pursuing a special unique item, or save those resources and use them organically throughout the game? Will they deny themselves access to one item if they pursue another?

    The trading and crafting systems combine together this way to give every collectible value. Everything can be converted into something else. Some necessary items require planning and several steps at different locations to obtain. It keeps the player thinking about exactly what is in their inventory, instead of collecting some item and forgetting about it entirely for the next 30 hours.

    What unique tools will players get to use as they cross the world of Hazelnut Bastille? What neat ways have you designed for players to interact with it?

    With item design, we were really focused on making sure that each of them, rather than being merely contextual, offered a way for interacting with puzzles, a way for traversing progress gates, and especially that they have some kind of combat application. There is nothing worse than finding a shiny new toy, only to find that all it is good for is opening a certain door that it happens to be the oddly-shaped key for. Players feel cheated in that situation.

    Some of the more interesting ones (without spoiling everything!): a tool which allows the player to transform into the denizen of a two-dimensional realm to avoid some hazards, and even visit this realm at times (think Flatlanders); a tool which allows the player to split energy into its two polar opposite base components, and use each for various purposes, as well as being a weapon; a tool which allows the player to both traverse distances and retrieve objects; and a tool which allows the player to create a very special projection which can deflect forces and control the flow of enemies in a room.

    On top of this, there are a number of “flavor” items in the form of rings which the player can obtain, often though side-quests, which though they only provide slight changes of under 15% of any given stat, have great impacts on how the player experiences some situations. For instance, though they only get 10% faster with one particular ring, the difference in their movement speed vs that of a certain enemy might triple as a result, making them substantially more evasive.


    What challenges do you face in making an appealing dungeon to work through? What thoughts go into enemy groups, room setups, puzzles, etc to make a player enjoy them?

    While Hazelnut relies on complexity through the combination of simple elements, every single room should offer a special problem which the player is equipped to solve, but may require some novel thinking. If you stop giving the player new situations to deal with, they quickly grow bored because their inner life is being ignored.

    So, in terms of combat, every fighting-oriented room should do one of the following: introduce a new enemy; place a known enemy in a geometry that in some way specially favors it (players tend to find this one a bit frustrating, hehe); present that enemy in a number high enough that it alters how the player must deal with it; combine several enemies in a way where their strengths synergize in a special way; place enemies in a configuration that makes them near unbeatable unless some puzzle-mechanic is utilized; create a situation where the player must use a certain combat technique or move in order to avoid taking damage; or create a situation where an otherwise simple enemy is combined with environmental hazards that restrict what the player can do.

    For a concrete example of combining enemies in a medium stress situation, we have a particular room from the demo with an open floor: the three blue crabs it contains will idle a while, then will move three discrete times. This means that they can cover a wide area of possible floor spaces around them quickly, since the player has no sense of where they will go. This creates a circle of mental tension around every single one, where the player knows they should not stand. Then we have our jellyfish enemy which has two key traits; it fires 4 diagonal projectiles, and it can’t be attacked with the sword while it is charging up or it will shock the player. The crabs in this room will set the agenda, since they are controlling where the player can go at any given time, meaning the jellies are free to pelt the player with endless projectiles until the crabs are dispatched. This is a pretty simple example, but you can picture how it intensifies from here as more elements are added!

    In terms of managing the player’s gradual learning of concepts, we are applying the tried-and-true formula of introducing a concept in a clean and focused context, then reinforcing it, then giving an interlude, and then testing it later in a high-stress situation. This ties in largely with the level layout for the map as well. What you generally need to do is to create a sequence where the player cannot really avoid learning what you need them to learn in more or less the proper order. But then you also want the player to have a sense of agency. They must feel as though they are making choices about where to go when, and get the sense that they can actually get lost, while the level is in fact gating where they can go. Super Metroid is a superlative example of how to design this way.

    Basically you draw up the full set of lessons the map should teach. You end up with 12-15 things. Then you divide those things into pools which can be taught in any order. For instance, you want the player to learn things 1-5 before going on to 6-10, but they can be any order, as long as they learn 2-5 after 1. You then put your progress gates around the level to ensure that they have exhausted all the ones in the current pool before going to the next one. It can get pretty challenging balancing it this way, but you can give the feeling of freedom and choice while also keeping the player on a steady curriculum.

    Later on, you open up shortcuts and new passages to minimize the need to repeat spaces unnecessarily. You also lay things out so that any given objective on the map is under a maximum number of rooms away from the start. If you want to build tension, you can put an objective an extraordinary distance away, with many severe obstacles between. It can be a challenge balancing all these factors at once!

    Likewise, what thoughts go into creating an interconnected world filled with secrets and buried dangers? What work goes into making a grand world filled with hidden items and pretty places to explore?

    These games seem to benefit from a world which is heavily open from the start. Link’s Awakening is one counter-example, where things open like a cloistered nautilus shell, one ring of chambers at a time; on the whole, though, the experience of suddenly being able to go anywhere, and contend with a world of options, is a big part of it. It really lends the player the feeling that they are finding their own way without guidance. While that is never fully true, since a good level-designer is giving the player a whole slew of hints to gently guide them on, getting lost in the ocean is part of the process.

    I am actually an architect as well, so I have sort of a special take on site and building planning. In laying out a building, you first come up with your program, with square areas for each room, and a desired set of adjacencies. These are generally placed first, and then you place what are known as “communicating” or “service” spaces, like hallways, antechambers, closets, mechanical spaces, etc in the intervening spaces to connect them in a logical fashion (all the while, staying conscious of how the structure will be roofed, and what massings you create on the outer facade).

    The design of these open worlds is very similar. You start with the terminal and hub locations. Hubs will be passed through very frequently, and should have the majority of the player’s repeat amenities, and terminals are places the player will visit only a couple times. You tentatively place them all down on the map, and begin to craft regions of the map between them that serve as heavily-traveled highways in loops. The player will traverse these again and again. They tend to produce only basic drops, but they also offer low resistance to the player except at a handful of special choke-point moments. These should generally be circuit-shaped and interconnected, to keep the player from getting where they are wanting to go in an immediate straight line, but also offering multiple paths to any destination. These spaces can also take on somewhat of a regional character at different locations, but also use a common language of assets and massing.

    The next layer of spaces between highway circuits and terminals are transition spaces. These tend to be much tighter, and offer higher resistance, and express regional tileset character to a greater extent. The player would find a lot of “arena”-like spaces in these transitions, or spaces where the environment offered puzzles impeding traversal, or a higher occurrence of Metroidvania-type item gates.

    One rule that I find seems to work across all level-design is that if you want to get the player lost, avoid long axes of straight lines, and force them to turn often. Without trying very hard at all, players immediately seem to lose track of where they are, and this builds immersion.

    Another aspect of it deals with how the shape of spaces constrains enemy encounters. For instance, we have one particular, fairly straightforward enemy that fires a projectile at the location the player was a few frames back. With lots of area to maneuver, this enemy is trivial. But constrain the player to move in orthogonal passages, and the thing becomes deadly. We have other enemies for which the opposite situation holds. In this way, enemy design puts strong design-pressures on things like overworld spaces as well.


    Hazelnut Bastille bears some clear similarities from A Link to the Past. What are you doing to really make this game stand out as your own?

    Hazelnut actually has many precedents: the Startropics games, Super Metroid, Secret of Mana, Chrono Trigger, Alundra, Terranigma, and the 2D Zelda titles, especially the original LoZ, A Link To the Past, and Link’s Awakening. I think folks tend to identify A Link to the Past most readily in the work for a few reasons. Of that set, it is the most well-known and played. It is probably the best regarded as well, and I think many players would like to see it recreated in some sense. This is actually not our goal; while there is a lot that it did better than almost any other game in its class, we did not set out to recreate LTTP. Of all the Zelda titles, I would actually say that the original LoZ is the one that shines through most readily, and a lot of folks old enough to remember both clearly will remark the same way.

    In studying our precedents, we sort of asked ourselves what aspects of each title were superlative, but coincidental, and which aspects were in fact the best possible solutions for certain problems. LTTP actually has the uncanny tendency to present the best possible solutions for a huge number of problems. Here are some examples: we chose to work in the version of topdown ¾ perspective that represents all walls of a room with equal emphasis, because it allows us to create dungeon rooms which don’t visually bias the player toward one wall or direction; LTTP has a tendency to use very heavy contours and borders which clearly declare collision space; similarly, it tends to represent objects as symbols rather than naturally in order to get objects to have clear orthogonal edges; it tends to render objects in the inner space of a room as being very short in proportion, so as to reduce visual conflicts with the very artificial form of perspective it uses; it has a system of environment tiles that allows for re-entrant corners, and this substantially increases the complexity of the dungeon spaces its tiles can support; it tends to favor very blocky enemy design, and this makes sense given the limitations of simple hitboxes.

    A lot of the design decisions of the background fabric of our dungeon spaces acknowledge that these techniques are in many ways optimal, and I think that this fact is at the heart of the dichotomy a lot of people have with the art-style: in one sense, they would like to see something completely new and different (which we also want), but in another, they also acknowledge, as we do, that some things have never been done better than they were in LTTP. And this is what we would reiterate as well; we did not set out to recreate the classic. We set out to create a work which learns from all of the best techniques past designers have used, and uses them as foundational tools for exploring our own ideas.

    In terms of what aspects of the work diverge substantially, I would actually say that level design is one of the bigger ones. Combat in LTTP tended to be incidental, and as interludes. It was often more puzzle-like: determine the best tool and approach for quickly dispatching the enemies in the configuration you found them. We use enemies very differently, more like original LoZ and Startropics uses them: they dominate the screen and deny the player from occupying most of the space; they are all moving different directions at once, and the experience feels sort of like occupying a clockwork contraption, and predicting the vector of all components at once in order to stay safe.

    The enemies themselves tend to be a bit less complex, but this is also an aid to us; the player will quickly recognize every enemy on screen, and recall what its exact capabilities are; this allows us to focus on complexity that results from compounding simple ideas, rather than complexity built into one asset, which tends to be a more fruitful technique for generating many novel situations at low development overhead.

    We are also placing a much greater emphasis on puzzles. In LTTP, puzzles tend to be simple affairs where the player must reach a simple conclusion about what order to perform a few steps based on quick observation. We are actually a lot more ambitious about puzzle design. We start in a similar way, by presenting the player with very basic situations with only a few elements, to make sure they have a context for studying each part; then we reinforce and test what we want them to take away about each puzzle element. After several iterations of this, the player encounters a monster of a situation.

    It happens in one room of the demo; invariably, when people see it, they glaze over and sigh. But then they notice they have seen and interacted with every concept it contains so far, and they must only grasp one or two more unique ideas in order to conquer it.

    There is also a big difference in terms of story-telling. Traditionally topdown, and particularly Zelda games, have been very event-driven, and this makes sense. Link is a doer. Something happens, and then he reacts to it. But he doesn’t really seem to do much of his own thinking, and he never expresses any emotion other than surprise it seems. Our story progression is much more character-focused. It is focused on the histories and needs of the various characters. It is focused on the relationship the heroine has with people around her. While games, by their nature, are active affairs, the story places an equal emphasis on what people feel as what they do. It emphasizes the degree to which people are truly reliant on one another.

    Alistair Wong
    Very avid gamer with writing tendencies. Fan of Rockman and Pokémon and lots more!

    You may also like

    More in News