How to Kill a Hedgehog: Level Design and Death in Classic Sonic

​

Good news: playing videogames is not, in fact, fun.

Jesper Juul makes a very compelling argument in The Art of Failure about how our experiences with video games are often negative, and purposefully so: we want to be challenged and we want to fail sometimes, even though we may sometimes not take it lightly as we would like. But then, why do we keep playing them? And why do we keep saying we do it for fun? His pivotal argument is that enjoyment of games has a central component of escaping failure. Beyond satisfactory motion play, secrets to uncover and general pleasantries, which are also important, we seem to enjoy having a personal journey of avoiding the inadequacy that failure brings. Even if we never fail, the idea that we could have failed, but didn’t, seems to be constitutive of the very idea that we are playing a game and not watching it unfold.

As a matter of fact, there is a proposition to be made here: if fun was so one-dimensional as not to accept failure as an integral component to it, video games would be highly inefficient in providing it. We’re often screaming, cursing or at least clicking our tongues at the games we play, and SEGA would be better off just selling us drugs. Failure is what gives us a sense of ownership over our experiences: everyone succeeds more or less the same, but each one fails in their own, personal way. And that special way only you could have played can be shared. It’s not a coincidence that amongst the most discussed classic Sonic stages, you won’t find Green Hill or Palmtree Panic, but Marble Zone and Death Egg will surely be there. You don’t even have to look back: Sonic Superstars has just been released, and all we can talk about are the final bosses! It’s not that failing is fun in itself: it’s that sharing our grievances, and what we thought had happened only to us, is a social and subtle component of fun that can’t be ignored.

Even for a purely personal experience with the game, failure plays an important role: since it is what a player will probably avoid the most, the prospect of failing is the element of a game that is most likely to drive a player to change how they act towards the game itself. That is, avoiding failure is how we are encouraged to explore a game’s mechanics, behaviors and places more deeply. Struggling at a game is how we get familiar with it, and how we remember it.

Following this sentiment, you can find an informal "Fun Scale" that has been going around since at least 2014, having been devised by outdoor activities enthusiasts to explain why we keep doing things that are not exactly fun, or like talking about them even if we would never do it again. This fun scale proposes there are three types of fun:

• Type I Fun, the kind of activity that is enjoyable while you're doing it. In Sonic, that would be breezing through the slopes in Chemical Plant, hitting that delicious extra hit on a boss that you took a risk for, etc.
• Type II Fun, the kind of activity that is fun in retrospect. In Sonic, that's beating a hard level, or finally killing that boss that gave you hell;
  
• Type III Fun, the kind of activity that is only fun as an exceptional story to share. In Sonic, that's getting a game over in Sonic 2's Death Egg even though you took your time to get more than 30 lives throughout the game (I'm so sorry @raphael_fc);

This "typology of fun", although informal, is congruent with Hirokazu Yasuhara's own theory of fun, about which he talked back in 2017 in a lecture he gave exactly about how to make games fun. He proposes a cycle, based on desire and execution, in which "fun" is defined as the "release of tension" between desire, plan and the result of an action through practice. The stage of "practice" only exists as long as there is "failure".

​

From this cycle, we can more or less define a framework of fun that happens not in spite of failure, but that embraces it. From Carol Yas' theory of fun, we can draw a general roadmap that reads a little like a bait and switch:

• The player takes interest in reaching a certain point, a decision that is a negotiation between what they want, what they feel they need most immediately and what they know how to do. You can influence that, but you can't determine it;
• They assess how they're supposed to do it with the limited time and resources they have, using the stage geometry to gather enough speed or momentum, however little that might be;
• Stage elements make the player's attention go to back and forth between patterns and individual elements on the screen as they execute their movements;
• Then, three things may happen:
  • They succeed, leading us to some good Type I Fun and, if they've failed before, some Type II too;
  • They fail! But by failing they get a new perspective on what they could do different, even if it's purely on execution, leaing up to a Type II Fun experience later on;
  • They fail, and they hate every second of it. But, if the story is good and there is some positive connection to the game already (the establishment of which this framework admittedly can't account for), they will at least experience some Type III Fun related to the game.

In this guide, we're only concerned with Type II and Type III, and how Carol Yas' "stage of Practice" comes to be in Sonic. So if you want to make a truly memorable fangame that will be talked about everywhere, one of the many things you’ll have to do is to kill Sonic. And you must actually want to kill him, leaving the player with that sense of inadequacy on one hand and, on the other, the prerogative of escaping this feeling that they’re lacking somehow. It will always be up to the player if they want to take up your challenge or not, but that doesn’t mean you should fear making them fear you.

What I can offer is a guide for everyone who feels the literature on Sonic often ignores enemy and hazard placement on levels. It’s a complementary guide to level design that does not intend to replace the ones we have (Trees' and Sparks'), but to enhance a particular aspect of it that is often overlooked.

Now, we won’t be talking about life systems or what happens after Sonic dies just yet, although I will say failure is at its best when it offers a new perspective on how to succeed, and that’s what will make us try again. This is a guide focusing on the level design aspects of failure and, while the effects of failure on general game design philosophy are definitely worth the investigation and discussion, it’s a much bigger issue.

An Anecdote

Spoiler: An Anecdote

It's important to bear in mind that this isn’t an argument that games should be harder. It’s simply a framework for level design that embraces failure as a dimension of fun, not an antithesis to it. This is best exemplified by my favorite platforming challenge from my childhood: the kitchen in my parents’ old house.

When I was a child, we had this rule at home: only one chocolate from the box per day, and only after dinner. In order to enforce this rule, my mom would keep the box of chocolates on the top of the refrigerator and get it for us after we were done with our vegetables. However, there was a huge cupboard beside the refrigerator that my brother and I would climb when she wasn’t looking to get the extra, forbidden chocolate. When she noticed we were getting too skilled at climbing the cupboard, she started hiding the chocolate box at the top of the cupboard itself, which was much, much harder to reach. We would still try, but we would get it much more rarely.

So one Easter, my mom bought a chocolate egg for us and, knowing we liked a) chocolate b) disobedience, put the egg at the top of the cupboard and said, rather unconvincingly, she would hate it if someone were to steal the chocolate egg from the top of the cupboard. And so we tried, because we had been prompted to do so, and so we failed. We didn’t mind much. We knew it would be hard.

Ten minutes later, the egg had been magically moved to the spot on the top of the refrigerator, from where we could more easily snatch and eat the chocolate egg.

Now, the point here is that the difficulty of grabbing chocolate from the top of the cupboard was always the same. However, in regular days, failure and success at doing so belonged to me and my brother. In that particular Easter, it belonged to my mother, and I didn’t think much of it. I didn’t frame the challenge: she did and, for once, she was the one to get frustrated that I hadn’t gotten what I wanted.

So pretending for a second that you really don’t want Sonic to survive doesn’t necessarily mean you’ll make a stage harder, but it does mean the player will have full credits for clearing it, and you’ll be better by not trying to control that feeling so much. It’s the subtlest of differences, but what are games without a masquerade and make-believe?

Typology of the Hedgehogicide

Spoiler: Typology of the Hedgehogicide

Classic Sonic traditionally conveys failure in two ways: a) directly setting the player’s progress back (such as in a Game Over), and b) suggesting their progress doesn’t meet a standard (by telling them they need all Chaos Emerald for the true ending, for example). We will only talk about how to set the player’s progress back here, and that relies on dealing directly with danger. We will help you with analytical tools to better understand the challenge you’re making the player face, as well as the challenges that exist in the official classic Sonic titles, so you may better understand why you like some aspects of the challenge there you like, or why you don’t like things you don’t like about it.

So if you want to kill Sonic the classic way, it can be done with five methods:

• Crushing: Sonic dies if he would be pushed into a space smaller than his hitbox;
• Time over: Sonic dies if the timer reaches 10 minutes;
• Falling: Sonic dies if he touches the bottom of the stage map;
• Drowning: Sonic dies if he spends more than thirty seconds underwater without touching air;
• 0-Ring Damage: Sonic dies if he sustains damage while he has zero rings;

These are the situations a player will most desperately try to avoid, and they are the ones that will make a player take their minds off their main objective: succeeding at the game and going forward within a stage. These major dangers, however, interact with each other and can’t be seen as separate pillars of Sonic-killing. A player is constantly evaluating what’s most dangerous to them at any given time, adjusting their desires and behaviors accordingly. They are less likely to look for the exact timing on a spindash to avoid being crushed if the timer counts over nine minutes than they would if it showed only three minutes. Likewise, a player with zero rings is more likely to expose themselves to damage, even if they’re underwater.

This constant motion of where the attention of the player is directed is a key component in excitement, so you should always be aware of to what major dangers you’re exposing the player, and actively choosing to frame them as more or less important in each stage. It helps that the skills needed to avoid each of these are distinct and are, sometimes, mutually exclusive:

• In order to avoid being crushed, a player often needs a keen sense of timing and the ability to execute movements with precision;
• In order to avoid time overs, one needs spatial and time awareness to evaluate if they’re close to the end of a stage. If they feel they’re close to getting a time over, they then need quick decision-making and focus on going forward a fast as possible;
• In order to avoid falling to the bottom limit of the stage map, a player needs spatial awareness of where the limit is and, if they feel it’s close, precision and timing are once again needed.
• In order to avoid drowning, a player needs to heighten their spatial and time awareness to gauge whether the time they have left is enough, and where to go if it’s not.
  • Unlike the danger of a time over, which make a player have to quickly understand what will get them closer to a mostly abstract forward direction, the space between an air bubble spot and the next is a self-contained challenge that forces a player to look and move around, bringing the player’s attention closer to Sonic himself.
• In order to avoid taking damage, a player may need a wide range of skills, from quick reactions to a sense of rhythm. In fact, the relationship between avoiding damage and common platforming skills is the bread and butter of the skill enhancement loop in Sonic. There will be a whole section about this, because you can’t adequately explain this in a single bullet point.

Creating situations in which the player must choose how they absolutely don’t want to die creates a healthy discretionary space for the player, because they will find the situation they’re in is multifactorial and can try many different approaches to avoid that failure, even as Sonic dies. Speeding up in a section a player found to be easy buys them time to be more patient with harder platforming sections further in the stage, so as not to get a Time Over or sustain a 0-ring damage; upon dying by crushing, a player may wish to improve their timing to run past whatever crushed them, or they may wish to exercise their spatial awareness by looking for paths that completely avoid the hazard. Sometimes, the timing skills a player gets by working to avoid a certain hazard make it possible for them to overcome that same challenge!

However, you will notice it’s extremely uncommon, if not unheard of, for any stage to try to employ all five hedgehog murdering methods at once. The reason for this is that, if we follow Juul’s train of thought, inadequacy can only be properly parsed if a player can identify what they did that was inadequate — that is, a glimpse of the adequate must be envisioned. And that’s impossible, or at least overwhelming, if all skills are being tested and demanded at all times. Failure is a negotiation between the player and the game, but it’s the latter that is able to show elements that will make overcoming failure seem worth it. If the player can’t establish a hierarchy of skills or hazards that they were able to surpass, avoiding failure becomes meaningless. There is no story to tell, and as we said, being able to tell a story is crucial to the art of failure.

It follows that carefully choosing, for each level, the major threats that will be prevalent or even present at all is very important. If anything, it should be noted that the only stages in which more than one major hazards (except damage) are even present in the classic tetralogy are:

• Labyrinth Zone (Sonic 1): Crushing and drowning;
• Scrap Brain Zone, Act 1 and 2 (Sonic 1): Crushing and falling;
• Scrap Brain Zone, Act 3 (Sonic 1): Crushing and drowning;
• Chemical Plant, Act 2 (Sonic 2): Crushing, falling and drowning;
• Aquatic Ruin, Act 1 and Act 2 (Sonic 2): Drowning and crushing;
• Hill Top, Act 1 (Sonic 2): Crushing and falling;
• Mystic Cave, Act 2 (Sonic 2) Crushing and (for all that matters) falling;
• Oil Ocean, Act 1 and Act 2 (Sonic 2): Crushing and falling;
• Metropolis, Act 3 (Sonic 2): Crushing and falling;
• Tidal Tempest, Act 1 and Act 2 (Sonic CD): Crushing and drowning;
• Metallic Madness, Act 2 (Sonic CD): Crushing and time over;
• Metallic Madness, Act 3 (Sonic CD): Crushing and falling;
• Hydrocity, Act 2 (Sonic 3 & Knuckles): Crushing and drowning;
• Marble Garden, Act 2 (Sonic 3 & Knuckles): Crushing and time over;
• Carnival Night, Act 2 (Sonic 3 & Knuckles): Crushing, drowning and time over (thanks @SyntaxTsu!);
• Flying Battery, Act 1 an Act 2 (Sonic 3 & Knuckles): Crushing and falling (thanks @SyntaxTsu!);
• Sandopolis, Act 1 (Sonic 3 & Knuckles): Crushing and falling;
• Sandopolis, Act 2 (Sonic 3 & Knuckles): Crushing and time over;
• Death Egg, Act 2 (Sonic 3 & Knuckles): Crushing, falling and time over;

Which is a relatively low number of stages (except in Sonic 2), and even in the stages where they are present at all, they are very rarely employed together. You will notice many of them also belong in the latter half of the games they’re in, so a sense of progression is highly beneficial to the player’s understanding of danger and, ultimately, their own journey around it. This sense of progression can be established with an incremental approach to these major threats (i.e.: don’t crush Sonic in Green Hill), which ties in nicely with the thematic progression of your game, but also with a clear distinction between the way many ways a player can be killed: again, there are only so many ways to succeed, but there are infinite ways to fail.

Death is a formless entity

Spoiler: Death is a formless entity

The concept of failure is in a different realm from that of the setback in that failure is not a mechanical phenomenon, but part of a psychological circuit that is originated in desire. A player wants something, but fails to get it and has to either try again or is forced to want something else. It’s a disconnect between what the player wants, what the player needs and what the player is able to pull off. Since it is construed by the player, subjectivity plays the most important role here: what each player values the most when they engage with each element of a stage will determine what they look for and how they approach it, so allowing them the margin of error to reconsider what they should do empowers them even though Sonic may die. If they’re the ones to build the hypothesis for what they should have done, they are more likely to try again and engage with the game on their terms.

Desire, too, isn't a fixed, objective phenomenon. It changes as the player makes their way through the game, and failure itself can shape what the player wants from a stage. It's swayed by the player's very particular relationship with the information they have about a given stage and Sonic itself, their own general predispositions as a player and how they perceive their own skill at that moment. In this sense, the concept of "path" in a stage is difficult to conceive: the idea of linear segments and intersections may account for how a player gets from point A to point B, but definitely not why a player does so in the way they do, and isn't really conciliable with a cycle of fun that's so directly tied to what a player can see on the screen and how they organize their priorities.

Fortunately, a trio of researches have proposed a framework of analyis for 2D levels back in 2008, in which they give us the concept of rhythm groups. These are short, non-overlapping sets of components in a stage with readable beginnings, middles and ends. The authors decided to name these little clusters like that because they saw that there is a cadence to each challenge, as if it was a musical phrase. Their internal rhythm does influence how the player will behave towards them insofar as they can read these sections, and then the idea of a path will arise from the connection of various rhythm groups. This is useful because the perspective of dying, avoiding death and changing desire patterns, as well as the incremental information about a stage map that comes with multiple playthroughs (some of which, too, are brought up by dying), makes it so that the concept of path is a function of the player, not of the stage map. Consider, for example, Marble Zone Act 3:

​

The information a player can acquire about elements in a stage, and how relevant that information is to the player, makes non-linearity possible even when the structure is normally considered fairly linear. Small moments, such as seeing a desperately sought-for checkpoint right above the player, can make them take the risk of jumping at a single retracting spike. Sometimes, recklessly and haphazardly so, leading up to failure. The rhythm of the player changing according to the rhythm groups, and the reasons why they do it, is more important to understanding how a player evolves in Sonic than the mere existence of separate paths in a map.

The relationship between pacing and perception of difficult being described here isn't anything new: Nicollet points towards difficulty almost as if it was a function of time and uses many examples from Sonic as he does so. @HEDGESMFG talks about it too in the replies of this very thread, and it only goes to show the importance of cadence, how and why it changes, to the idea of failure.

So if you consider a sequence of crushing hazards that move in a certain order and at a certain speed, and a sequence of retracting spikes that move in that same order and speed, the theme of the challenge to be perceived by the player is similar, even though the threat isn’t. How a player might deal with that specific kind of challenge, too, will change. For the player to have a satisfactory construction of the skills they want to have in order to clear a stage, it’s important that a stage seeks to highlight a certain vice in gameplay. Amongst the ones that are possible, there are:

• Mechanical imprecision: the inability to control Sonic adequately and land a jump or a move on the right place, at the right time;
• Slow reflexes: the inability to move away from incoming danger quickly enough;
• Bad tempo: The inability to move in the same rhythm as the stage’s elements;
• Hesitation: The inability to grasp an opportunity to reach a safe position, not exactly due to slow reflexes but to not taking the decision quickly enough;
• Recklessness: The inability to gauge the risk of an action relative to known or unknown off-screen elements;
• Reading imprecision: The inability to move according to a single element’s behavior, doing exactly what it wanted you to;

You will notice, again in Marble Zone Act 3, that some rhythm groups are very similar to each other — not only in the elements they use, but in their cadence. The fact that a stage will use similar rhythm groups, changing their general direction and sequences, points towards a notion of pacing that belong to the level as a whole. In Marble Zoe Act 3, we highlighted 6 rhythm groups that repeat themselves, with changes in direction, punishment for failure and even some elements. Fundamentally, though, they build up a certain mechanical theme.

Each stage adopts a specific mechanical theme to threaten Sonic with, primarily, even though it can and will, of course, probably have all of them even if you don’t design is as such. This is on the player, after all, and your focus is on the skill you would like them not to have, not on the ones you would like them to have. Thus, it’s this theme that will make the danger in your level unique and memorable. The nature of each specific threat is less important than the skill (or the lack thereof) that they, in conjunction, kill Sonic for. And all elements in a stage, not just the hazardous ones, tend to one side or the other by behaving in different dispositions regarding their relationship with Sonic’s presence.

• Autonomous: The behavior of the object or geometry doesn’t change when Sonic is near, even if it only happens once;
• Reactive: The behavior of the object changes when Sonic touches or approaches it;

The periodicity of their behavior:

• Once: They will move and act only once, then either self-destruct, stop or return to their original state;
• Periodic: The object will move and act in a perceivable period that the player can adjust to;
• Random: The object will move and act in a random pattern, or a pattern that is way too complex for the player to quickly understand;

And the target of their behavior:

• Path-oriented: The object acts oriented by a predetermined path that can be inferred and read by the player;
• Sonic-oriented: The object locks onto Sonic and acts using that position as a target, specifically;

The amalgam of objects with different traits, plus the way Sonic can interact with them and the major threats present, makes up the mechanical theme of the stage, that is, the specific flavor of failure it will probably deliver, and the kinds of actions that will most likely result in deaths. Looking for a common theme among them helps a play parse their failures and hone ne particular way to use Sonic's speed. So when designing death in a Sonic stage, you should work to deliver a sense of danger even in the elements that aren’t dangerous. For example: when designing a stage full of traps, that focuses on quick reflexes and despises hesitation above all, having even normal floating platforms behave reactively to the player — needing to be triggered to start working the intended way, for example, especially if the player has to move backwards just a bit to catch the platform — helps the sense of danger to feel palpable even if the triggered platform leads directly to an item they’d love to have, like a shield. This is exactly how stages like Labyrinth Zone behave.

As a general rule, periodic elements with a predetermined path make up a stage or a section’s internal rhythm and force the player to look at the big picture to escape unscathed, as Sonic must match the general beat with his jumps or dashes of speed — and that is a challenge in itself. In turn, elements that act once, randomly or in Sonic's specific direction force the player to look at the area immediately surrounding Sonic. So, building a tempo-based section and adding a reactive element in the middle of it can disrupt this tempo, putting Sonic in danger and demanding quick reflexes and decision-making to keep him alive. Much in the same way, losing rings in a tempo-based section usually ends up killing Sonic, because as rings scatter, the player tends to disrupt their own pace to go after the lost ring, prompting Sonic to be hit again. The frequency and quickness with which you try to change the player’s focus from patterns to singular elements makes up an important element of difficulty and the mechanical theme you’re proposing too, so keep this in mind when playtesting instead of littering a stage with moving platforms and static hazards (or the other way around). The way you do it will have an effect on the player’s attention span and flow. After all, one element can have many functions depending on the player's behavior:

​

In this hypothetical example, the Orbinaut about to shoot its spikes as it sees Sonic tests different skills from the player depending on the direction he's coming from:

• The Sonic coming from the left is the only one that can safely an surely jump over the Orbinaut and completely ignore it, but, for that, he needs to not hesitate and have quick reflexes;
• The Sonic coming from above is the most likely to have their reflexes tested, as they will have the least time to react to the Orbinaut shooting its spikes at him;
• The Sonic coming from below is the least likely to be able to ignore the Orbinaut, but they also will probably have the most time to assess the timing between each spike;

Deciding the stage’s mechanical theme beforehand also informs your choice of elements and movement options for the stage, making each one kill Sonic in its own way and giving the mechanics of the game a tangible counterpart to visual theming. For example, casino stages usually deal in imprecision first and foremost, as they make Sonic bounce around; underwater sections, on the other hand, make controlling Sonic with precision easier, but anything but careful reading on hazardous objects will kill Sonic, as well as hesitating to jump to the next safe position when needed, because you may not have a second chance. Above all, maintaining an eagle’s eye approach to what threatens Sonic allows for emergent play and emergent deaths, which is incredibly important. After all, the same element can be harmful in different ways according to the context, and it’s up to the player to read the stage and the mystery of the off-screen. By knowing this, you can make conscious use of the various elements you can use to actually kill Sonic.

The Rodent Killer’s Toolkit

Spoiler: The Rodent Killer's Toolkit

Like we said before, the trick to killing Sonic effectively is to use malicious elements and still make it look like it’s the player’s fault for having them killed. There are two sides to doing this: failure being a byproduct of desire, giving benefits with one hand and hazards with the other is great for leveraging Sonic’s successes against himself. This is just a way to frame the simple principle of risk and reward, and make it actionable. Making a player have to constantly reconsider whether it’s worth it to go after shiny objects at all, or think up different strategies at different times, keeps the game engaging! Your toolkit consists of:

Malicious stage objects
The bread and butter of the afterlife ticket service for hedgehogs, these are objects whose only purpose is to damage Sonic. They’re the primary source for the feeling of danger because they are, or they should be, immediately readable to the player as hazards. Spike balls, spike nests in a pit (more about these later), lava, badniks, projectiles and all objects that can kill Sonic if he has no rings fall in this category. They can be further divided according to these traits:

• Countable or uncountable: a player can either always measure where the danger ends and estimate its relationship with the realm of the off-screen or not; lava or electrified surfaces tend to be uncountable. But not always! The fireblasts in Scrap Brain, for example, are countable;
• Vulnerable or invulnerable: whether the player can neutralize the threat or not defines its vulnerability. Green Hill’s Chopper (the piranha that jumps from waterfalls) doesn’t behave much differently from the fireballs that shoot up and fall back down in Marble Zone; however, the fact that one is vulnerable while the other is not makes the player behave much differently around them.

They are more suitable for “guarding” valuable objects, making verticality visibly worth the hassle of platforming and killing the player when they act reckless while dealing with another major danger, such as moving around aimlessly to avoid drowning. In Sonic, you are much more likely to see more invulnerable and/or uncountable malicious objects in later stages, because they adopt a state of nondepletable danger that is useful for maintaining tensions high — which, in turn, makes it so that every skill the player has acquired throughout the game is tested twice as intensely.

​

Sonic games are peculiar in that, typically, malicious objects do not interact with each other. Badniks don’t get killed if they fall in spike pits or touch spike balls, or their movement doesn’t overlap at all. Fire-shooting enemies don’t rush to bodies of fire to recharge, and you (generally) can’t trick badniks that chase you into getting crushed in your place. I would say that this is because enemy encounters, in Sonic, are rarely cerebral puzzles to be solved — if ever. There’s little figuring out to them, except at the intuitive level of understanding their movement pattern or action timing. You have no option to use the stage’s elements against themselves. Rather, you can only use Sonic and his speed directly — what changes is how.

A general rule in Sonic helps you telegraph which objects are malicious an which ones aren't: beneficial objects almost never move. Rings and monitors are only seen moving at all in Scrap Brain (Master System/Game Gear) and Metallic Madness (Sonic CD), and even then there are other objects or geometries moving them. So the syntax of Sonic normally says that not all static objects are beneficial, but if something moves, it's neutral at best.

Malicious stage geometry
Objects or structures that can actively kill Sonic regardless of his ring count, such as crushing hazards or bottomless pits, are particular cases of the general rules of how you can kill Sonic. They automatically draw the player’s attention and must be properly telegraphed. One trick to make them worth the player’s attention is to associate them with otherwise easier or bountiful zones or routes in the stage map. For a stage with bottomless pits, it’s a good idea to keep most rewards in the bottom areas of the stage, and limit the number of hazards that can throw Sonic in them as he takes damage. You generally want 0-Ring Damage to be an incremental process, not a one-two asshole combo.

Spike nests on vertical walls are different from the spike pits we described in the last section because they are another example of malicious stage geometry and what they can convey: in general, malicious geometry is more interesting when it doubles as a useful platforming tool: If Sonic can stand on a spike nest on a wall and, from there, jump on a higher platform to access a different rhythm group, it’s on the player to avoid danger or successfully skirt around it to get what they want. This applies to crushing hazards as well and, perhaps more importantly, to regular stage structures: malicious geometry makes it so that the player has to read the level appropriately in order not to die and, more than that, get something nice in the way. For example:

​

There is a pattern to the electrical wires in Wacky Workbench that make it so that the floating platforms are either safe to stay on or almost hazards in their own right. And, again, there is a rhythm to them: the periodic behavior of the floating platforms and the periodic behavior of the wires makes up an internal rhythm to them that makes it so that the player who can gauge the timing of the section as a whole is more likely to stay safe. However, the Poh-Bees here can shoot at Sonic and disrupt the player’s attempt at matching the section’s beat. What’s more, the bouncing floor makes it very hard to be precise about it.

So the relationship between regular stage geometry and malicious objects makes it so that the player has to get a feel of the stage as a whole before proceeding. This adds depth to the gameplay and provides many steps for a player to master this particular section.

Malicious stage map structure

A stage that is long or less straightforward in its structure (with a looping Y axis, for example) can make the stage itself a time over hazard or can facilitate hazards in it. However, this is best paired with multiple secret rooms and paths, special stage opportunities or time-consuming platforming challenges that lead to useful, visible items such as 1-Ups (if a life system is present) or shields. If the player feels like they opted into wasting their time, it’s their fault, and they will leave the stage feeling like exploration is its own challenge. They might even want to play it again just to get everything! In this example from Death Egg, a player is likely to see how bountiful is the path beneath the light-beam-transportation system only after they take said system. Since the beam dances around, it wastes time. But a player who is low on rings may want to pursue those monitors anyway: that depends on the tension the player is feeling.

​

Another example of malicious structure is one that has a pervasive element that changes the game’s rules somehow. Water, besides being able to kill Sonic itself, changes how fast Sonic can go, which also enhances ordinary hazards. Gravity-reversing or a different orientation (a player has to go left or up to complete the stage, for example), while not being able to kill the despised rat outright, may make it easier for other hazards to kill him.

Another good principle to have is that, since gravity can only bring Sonic down, but the main way for Sonic to gain speed is by using downward slopes, the paths at the bottom of the stage should be more dangerous, in general. No free speed should remain unpunished, and since Sonic being faster means he’s more powerful relative to the stage, maintaining it should require active work on the player’s part. He should have to go slower so that he can go faster.

This brings us to one of the most important points about the stage map structure and geometry: blocks are your friend and Sonic's enemy, by nature. Curved surfaces help Sonic keep his movement, so their prevalence also changes how the player acts towards the objects you place. It's not by coincidence that stage progression in Classic Sonic starts with lots of curves and opportunities to be launched to the air curled, and ends with more blocky platforms and situations in which Sonic'll be in the air uncurled. This helps building tension and bringing new life to elements you've used before!

(A note: although a Time Over is technically a threat in every single stage in the games, the ones in which it is a threat in practice are very few and far between, and it’s the player that will put themselves in a situation that would warrant a death by time over for wasting time going after secrets and extras.)

The very special case of badniks

Spoiler: The very special case of badniks

Badniks deserve special attention because, for the most part, they are the only company Sonic ever finds in a game, other than Eggman. They are the stage's life, and make up the most volatile face of frustration in a Sonic game: because they can hurt Sonic, but can also be destroyed, they tend to draw the player’s attention towards them (unlike invulnerable malicious objects). So if the player died to one, feeling inadequate, the glimpse of success is almost inherent: a smaller objective becomes to simply kill them next time. People like killing badniks to neutralize the threat (and the noise they make when they explode is pretty satisfying too).

By being a depletable source of danger, they can change even a room that doesn’t move: as Sonic first enters an area, he may kill all badniks before engaging with the other elements, and this makes that area able to basically be experienced twice. And, because Sonic can bounce off them, a skilled player can use them as platforms that can only be used once to reach somewhere else, or kill them to use the whole area as a running ground. Their presence or absence, and how threatened the player feels by them, change the player's whole approach towards the stage. This dynamism adds depth to a stage, and must be applied consciously. They can freely either set or disrupt the tempo of Sonic’s actions, greatly conveying the mechanical theme you’d like your stage to have.

But good badniks are never alone in how good they are. Tying their placement, design and recurrence to the mechanical and visual themes of the stage, as well as its position in the game's general progression, culminates on giving a face to failure. And this unique relationship a player can develop with a threat makes it a character all on its own. Through hate, sure, but loving to hate something is important because it's a funnel to people's memories of a game. Again, shared contempt for an element, that people were able to overcome nevertheles and in an otherwise pleasant experience, makes people bond over the game. Which is fun — through failure.

Every badnik can be categorized according to some parameters, and they help us seeing through some of their purely visual cues to understand how they work in the context of the stage they are in:

• Behavior, which (like every stage object) can be autonomous or reactive;

(In which the yellow underline means an autonomous badnik and the red underline indicates the badnik is reactive)​

Technically, there is no completely autonomous badnik in Classic Sonic, insofar as they load at all when Sonic is sufficiently near them instead of obeying a global or stage timer. This is less important than whether they change their behavior when you can see they saw you. You can see Sonic 1, Sonic CD and Sonic 3 have an about 50/50 ratio of (semi-)autonomous and truly reactive badniks, while Sonic 2 and Sonic & Knuckles have a much bigger proportion of reactive badniks. This merely changes how danger is delivered and isn’t in any way relate to how hard each game is, but you should be mindful of these numbers nevertheless. After all, they make up an important component of rhythmic-and-disruptive motion we talked about, and how the attention shifts from movement patterns to each discrete move.

• Periodicity, which (like with every stage object) can mean the behavior is performed once, periodically or at random;
  
• Position, which can be water/airborne, ground or surface-based;

(In which the green underline means a ground-based badnik, the blue one denotes an air or waterborne badnik and the brown one means a surface-based badnik)
​

Some badniks can become different challenges depending on where they are positioned, even if their behavior is the same. Orbinauts of all kinds look fine in the air, but it's not unusual for them to be positioned close to the ground. Since they only ever move horizontally, if they do, they end up becoming functionally ground-based in this case — which means Sonic can kill them by rolling, or jumping over them becomes the only way to get past them without sustaining damage. It is probably fair to say that, all else constant, water/airborne badniks are more threatening the more a stage has opportunities for Sonic to be launched uncurled to the air, like springs or non-smooth surfaces. This means that, even if Pata-Bata and Ga have similar movement, Ga is much more dangerous because it's more common to be launched straight at a Ga by a bumper or spring in Collision Chaos than it is to run off a platform into a Pata-Bata in Palmtree Panic.

A badnik being surface-based means their presence is associated with some element in the stage for which the player must watch out. This increases tension even as the badnik may be absent: Choppers can jump out of any waterfall in Green Hill and, even though not all waterfalls have them, their surface-based position means a player is more likely to be guarded around every waterfall.

This is a smart way to integrate a badnik into the visual theme and increase tension all throughout a stage while avoiding littering the map with them unnecessarily. If an element on the screen — like walls or trees — can be associated with a particular badnik, the player is encouraged to be cautious around the stage as a whole, reading it like a congruent ambient rather than a disjointed set of elements. Allowing a player to formulate the stage's environmental rules in their minds is very beneficial, because it can improve how they understand the mechanical theme.

• Movement, which (like every stage object) can be static, pattern-oriented, Sonic-oriented or random;
• Attack, which can be done with a projectile, a body part or none;

(In which the red underline means a badnik that attacks with a projectile and the purple one means a badnik with a harmful or invulnerable body part)
​

Besides their own movement and existence, which can hurt Sonic if he comes at them unprepared and uncurled, some badniks can throw projectiles with movement that is indepenent or has a pattern of their own, or a body part that will harm Sonic even if he jumps at them while curled. That, or it reacts differently, like making Sonic bounce or simply acting like a normal platforrm. These demand that the player pays attention, because they can't be destroyed wherever whenever, making the player focus on their periodicity or anatomy.

A combination of both, as in the case of the Unidus, makes the badnik explode or detach its harmful body part as a projectile. This creates the unique dynamic in which a badnik attacking Sonic leaves them vulnerable or eliminates them, which is the closest Sonic gets to resemble combat. There are 15 such instances across the Classic Sonic tetralogy (Unidus, Bomb, Sol, Spiker, Asteron, Shellcracker, Slicer, Taga-Taga, Bigbom, Turbo Spiker, Star Pointer, Skorp, Iwamodoki, Chainspike and Spikebomber), and the reason why they are sparse is that they are so hard to be ignored that, for a moment, they become the center of attention of the player, but the best outcome is the simple neutralization of the threat, while the worst is death. They are mechanical themes unto themselves.

Mixing and matching these traits can give you very interesting objects, such as this asshole:

Hidrocity's Turbo Spiker is a reactive badnik that only acts once: if it sees you, it shoots its spiked shell and runs in the opposite direction. However, the spiked shell is fired off as a projectile — not in Sonic's direction, but at a fixed angle. So if it sees you, the one thing you can't do is to jump to kill it. It's the one thing that can get you hurt when around him, yet it's the thing most people might try for the first time.

But, of course, the challenge is not made of the badnik alone. Metropolis' Slicer would be much less hated if you could reliably jump over its homing blades. However, since they will usually see Sonic while he is going up the corkscrew gimmick, and they typically sit in tight tunnels that won't let you jump over them even if you react quickly and you know exactly what they're going to do, their margin of error ends up a lot thinner than it would be in, for example, Emerald Hill. As a matter of fact, Metropolis is a remarkably difficult stage and one of the reasons for that is that all three of its unique badniks have this combat-like quality, demanding full attention to their attacks in tight spaces. This makes the stage very mentally exhausting and, sooner or later, you might end up slipping due to a mechanical imprecision.

You are Eggman

Spoiler: You are Eggman

​

Good news! Bosses have a surprisingly weak definition in literature. Everyone knows what a boss is when they see it. But for all the discussion that takes place mostly since 2011, with discourse revolving around bosses no lnger being needed (1, 2, 3, 4), there is some difficulty in giving them a strict formal definition, even though they're such a particular phenomenon. Zwack's bachelor thesis does provide us with a review: Newman defines boss fights as "puzzles with only a single solution", a "singular problem" that is "completely self-contained [...] within the gameworld" with its own "magic circle", thus being "the controlled climax to a particular sequence of gameplay"; but this feels somewhat ethereal, because many events can be described like that and doesn't really provide us with many tools to understand how they work in the bigger picture. Some authors then understand they are is, or can be, a bottleneck that needs mastering skills you've learned throughout the "particular sequence of gameplay" to be beaten, but that's not necessary or inherent; Rodgers provides a vision that is more sophisticated, though not at all mutually exclusive with Newman's: it's "a dance between the enemy and the player". Yet, the idea that the boss is a climax implies a mechanical progression structure that — to our knowledge — isn't very well developed or understood. A climax of what? Stages only necessarily lead up to bosses by virtue of there being nowhere else to go, but there isn't a challenge progression that makes it cathartic or consequential in its nature. Certainly, it's not just a question of the boss being a harder challenge than the rest of the sequence whose end it "guards". And in any case, is it always the position of the challenge that matters?

In the context of Sonic, we typically understand bosses as encounters with Eggman's bigger mechs, usually piloted by himself. There's a thematic link between them and badniks, and an opposition between them and Sonic. That much is clear! But there are no exact rules to what they can be and, most importantly, to how they work. If you squint really hard, you can see that a boss like Labyrinth Zone's is mechanically closer to Hidrocity Act 2's first set piece, in which a wall moves underwater intending to crush you, than to any other boss in Sonic 1: the idea of an unavoidable section where you have to move quickly to reach a specific point, but hindered by water and threats, ties them together. The quetion of "what is a Sonic boss", specifically, must then be prefaced by the caveat that the feeling of completing a boss-like challenge can appear basically anywhere. A boss is just a rhythm group that insists upon itself. But that's fine! Bossess being a device used so abragently, we just need to understand what it is about them that will kill you.

The Blueprint is in Green Hill

​

The most distinctive feature of a boss fight in classic Sonic is not Eggman! It's that it takes place in an arena, which, especially in the context of Sonic gameplay, has interesting implications. For almost any other challenge in Sonic, coming at it running and with space to maneuver increases your chances of overcoming or avoiding it. But as the edges of the screen become walls and Sonic generally has to start the fight with limited movement options, Sonic is at his weakest when it starts and so he will stay for the remainder of the challenge. That's powerful, because it inverts the logic of the rest of the stage: The player will necessarily have to look for ways to do whatever is required from their most vulnerable position and avoid, at all costs, returning to it. So it's safe to say designing an interesting boss starts with designing the arena, not Robotnik's mech.

Egg Wrecker, the very first boss a player will face in Sonic 1, shows this very elegantly: standing anywhere except the lower corners of the screen is not safe at some point during the wrecking ball's movement period; not only that, but the platforms on which Sonic has to jump to attack are the areas on which the ball will stay for the most time! This is very simple, but the idea that Sonic will have to expose himself to the most danger to attack is what makes this elegant. It follows then that the safest time to attack is when the ball is on the other side. However, doing so means the player will usually not have the time to hide under the platform for the next hit.

It follows then that Sonic must move like a pendulum, much like the wrecking ball, to be effective. While some may argue that this makes the bos a puzzle, we much prefer Rodgers definition of it as a dance. There are many right ways to dance, but they all rely on finding a rhythm for hitting and hiding. The best players will be able to never hide, simply swinging from platform to platform and basically being one step ahead of the ball. But this is the least safe path: each additional hit disrupts the rhythm ever so slightly, increasing the pressure to move faster so as not to get hit.

There being several possible dances for the same rhythm is crucial to Sonic bosses — we would argue here that this is the most important part of them. Being able to hit them constantly being more dangerous, and demanding that the player keeps it up, makes it so that taking risks might be worth it, but the skill required is actually much higher to perform. So the best player will able to completely erase the line between an "attack phase" and a "hiding phase", but that requires constant movement — and the boss' job is to disrupt it, more than anything.

A common criticism of Sonic bosses, in and out of the fan discussion circles, is that you're able to just ignore the patterns altgether and "spend" rings to hit Eggman more times, trivializing them. We don't think this is the whole picture here: losing rings, and having to retrieve them, disrupts the player's rhythm more than anything, leaving them more vulnerable to the next hit, so this strategy has a very hard limit because it leads to a slippery slope.

That's not to mention another aspect of boss fights: the player's performance right before them matters, because it dictates how many rings and lives they will have. Essentially, how many rhythm disruptions they can take. This is one reason why the checkpoint right before the boss is so important! It limits how many rings the player will have if they die (generally to three), unless they go ut of their way to backtrack an get more rings, exposing themselves to the dangers we already covered and builing up the skill that they will need to overcome this obstacle.

TL;DR

• Think of failure and frustration as a step towards fun, not as its enemy!
• Failure comes desire and desire is manipulated by failure. If, by killing Sonic, you make the player see their skills or the stage differently, you did it!
• Tie the thematic progression of your game to an increase and to a mix of the ways a player can die, as well as in a variety of mechanical themes;
• The mechanical theme is the particuar way each stage intends kills Sonic, making the player use Sonic’s speed differently;
• In order to achieve that, you have to be varied in the elements you use, dangerous or not, and in the way they behave, according to a single theme as much as possible;
• The roadmap to failure goes like this: player is interested in reaching a certain point → they assess how they're supposed to do it with the limited time they have → the player's attention has go to back and forth between patterns and individual elements on the screen → this messes up their execution somehow → they fail;
• It’s okay, important even, for the player to hate some stages. Trust them;
• Badniks give life to a stage. The way they’ll pose a threat is also dependent on the stage geometry and structure around them, a well as the reason why the player is engaging with them;
• Instead of asking “what is this element doing here?”, try to ask “what is Sonic doing here?” and try to change that. Focus on the player’s possible desires and make them work to achieve what they want;
• Bosses are dances, and rhythm disruptin is your most essential tool in them;

 

Thanks! I completely missed CNZ2! I deliberately put off including Mania as an example for now though. The main reason being that since it was done in a completely different time with the original classics in mind, this can lead to distortions such as being "hyper-classic" in some areas while being a total outlier in others. I need to investigate it further!

 

(I do like Slicers in Metropolis, but) you raise a good point about them being counterintuitive but that not being totally accounted for. You could say if you simply don't hesitate to roll, you won't get hit -- but absolutely everything else in the stage forces you to hesitate and look around carefully, even the clunky corkscrew that will always throw you directly on them. And they typically sit in tight tunnels that won't let you jump over them even if you react quickly and you know exactly what they're going to do, so in practice they demand a specific movement and the margin of error is really small. I don't really enjoy it when the game demands specific movements rather than specific skills, so this should be accounted for in the text.

 

I've just added something extremely important: the idea of a "cycle of fun" isn't mine at all. I took it from the man himself, Hirokazu Yasuhara, who gave a lecture in 2017 about how to make a game fun. I'm just giving more insight into the negative part of it, and refining the idea that the "stage of practice" isn't antithetical to fun.

In the same year, he posted a more complete version of his presentation in Japanese, and said he'd post the English version later on, but never did. I'm looking into translating that¹ because it'd be extremely useful for us.

¹ hiring a translator to do it, that is;

 

I'm not of the mind that cheap difficulty is always off-limits, or that memorization isn't a valid skill to have -- especially in a game in which you can sometimes completely change course and avoid what you memorized altogether. I can see a stage in which a player that ignores warnings and environmental cues could "deserve" being hit by a cheap challenge at some point, especially if the visual theme conveys that too. But even then, it'd be an extreme case, and I understand I'm in the minority.

I should mention, though, RNG definitely has a place in building difficulty, possibly even i Sonic. I'd definitely reject the idea that good gameplay should be easily reproducible, at least because my idea of good gameplay has much more to do with self-expression than with mechanical excellence. After all, risk management is a skill one can have too, and randomness instantiates challenge in a way easily reproducible good gameplay can't, and it's pretty unique as a method.

Though I haven't written anything about bosses yet, the one thing I'm certain of is that Death Egg Act 2's boss (the first one, at least) is like one of the best classic bosses. Prbably one of the better boses in the series. It uses the stage's gimmick to hit you, but you can use that same gimmick to hit it. There's a lot of danger, but there's a lot of space for a player to figure their own way out of it too. I've taken a newfound appreciation for the entirety of Death Egg too while replaying S3&K too. I think it does the whole "final stage vibe but not completely stifling s you can still go pretty fast if you knw how" masterfully. Truly the climax of the classic series!

 

Detailed the relationship between failure and rhythm, improving the section that leads up the idea of each stage killing Sonic differently. By the way, changed the term "challenge mode" for "mechanical theme", which describes the idea much more accurately.

Also brought the idea of "rhythm groups", as proposed by Smith, Cha and Whitehead (not Christian), being a better framework for analyzing and making platformer levels than the idea of "paths". As an example (also in OP), see how Marble Zone Act 3 looks like when divided in rhythm groups:

​

EDIT: Forgot to point out that some rhythm groups are structurally similar to each other, and that's what will give a level a sense of theme cohesion! Silly me.

 

For my 1000th post, an update: I've added a section about bosses! I think it's especially important given all the debate about Superstars bosses. It's incomplete, and there's much more to write, especially about the relationship between bosses and the stages they're in. But for now, this will have to do.

At first, I was sort of disappointed when I noticed most Sonic bosses don't use the stage's mechanical themes, nor do you have to master them in any significant way to beat them. But then, after studying them more, I realized that might be a good thing: the most most most important thing about them is them having a "ready or not" quality to them, and needing to observe something while it tries to kill you is its own kind of fun.

 

Update: I've moved (almost) everything regarding badniks to a single section just about them, that is greatly because we love to hate them. They are super cute! In the future, there should be some text about the parameters I only mentioned, but the most crucial parts are covered, either there or in other sections. Added some considerations that relate to the points @Starduster brought too.

(I'm sad that they never made a ring-sucking badnik in a stage where you can use rings to open up paths (like Scrap Brain Act 2). But oh well.)