This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.
This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.
Thanks to having read an excellent analysis of Super Mario World, primarily on its level design, on Patrick Holleman’s site Reverse Design, I was prompted to play through World again, including its secret stages. Although reappraising World through the assistance of Holleman’s document hasn’t caused me to like the game -- which I've not been fond of for a long while now -- much more, it has helped in reformulating some criticisms, and was enjoyable to read in its own right.
Coming off of Super Mario Bros. 3, World feels tonally cold. Part of this has to do with how there were conversational, or at least monologic, interactions in SMB3 that are not present in World to nearly the same degree of regularity of volume. Not even the Princess has a word to say to Mario once he’s rescued her. Another part of this has to do with how SMB3 presented itself through a theatrical conceit (e.g., the game’s opener is a curtain being raised), and how the flatness and closeness of stages’ backdrops played into this. World gives every indication that its world is “real” in a way SMB3’s was not, while maintaining SMB3’s flatness and monumentalizing details. Consequently, it is responsible for invoking in me the closest thing to horror vacui that I think a Mario game is capable of. And whereas SMB3’s bonus stages were accessed by mysteriously appearing icons on the world map or set spots where the player met up with Toad, the only bonus stages in World are tic-tac-toe rooms, accessed by collecting 100 bonus stars (gotten from hitting the moving bar on a stage’s goal). For all of its bright colors, happy music, and cute critters, World seems uncomfortably distant.
Most of all, I’m not enthusiastic about World because I don’t think its level design matches the level of quality seen in the NES trio (personally, I’m replacing the North American/European version of Super Mario Bros. 2 with the Japanese version). For me, this lack is mostly a result of what sorts of ideas relative to the newer mechanics it ignores. World's most major and unrealized new mechanic is the cape. Calling the cape “broken” is, in my opinion, fair, but to be more accurate it’s “broken” because the game’s design doesn’t support its tactical advantages or rise to meet its exploratory implications.
The cape, gotten by the cape feather, is an extension of SMB3’s super leaf/raccoon tail power-up with a few crucial developments. As before, the cape functions primarily as a way to cover large spaces aerially and lessen the toughness of jumps by a gliding mechanic. Now, however, the player can take off using the new spin jump, making them immune to practically any threat they may land on when descending; they can glide by holding, instead of rapidly tapping, the run button; they can divebomb and slam into the ground, or swoop back up; and they can fly through the air indefinitely by rhythmic, backward (relative to Mario’s orientation) presses on the D-pad. Competent usage of this last feature is fairly easy to learn and satisfying to control. It’s an extraordinary technical accomplishment for the SNES’ launch game. It should also be noted that flight via the cape, even when the perpetual floating feature is not used, covers an enormous horizontal and vertical range, often allowing the player to disappear beyond a stage’s ceiling and bypass large stretches.
Maybe the most disappointing thing about the cape is that there is so little to find in World’s skies. The vast majority of the time, the only thing aside from emptiness that the player will find are flat walkways with some coins on them. More rarely, there will be a small, plain pocket containing a secret exit (e.g., the upper end of Donut Plains 1 or Vanilla Dome 1) or an item. World had the opportunity to, at least in its more vertically spacious stages, provide lower and upper routes in a given stage, yet it doesn’t compare even to SMB3 -- a game that, in terms of instances, actually has less upper involvement. World's Vanilla Dome 3 is an exemplary exception, housing a couple of alternative upper routes; it’s also disappointing, in that these alternatives hold no level design of note, and the majority of its upper limits are strangely enormous and empty (perhaps an odd attempt at creating mood).
What is also disappointing is that the amount of performative output required of the player has been lessened and that the amount of secondary routes to elevated portions have been practically eliminated. In stage 3 of SMB3’s second world, for example, there are several elevated platforms that are seemingly out of reach if the player doesn’t have the raccoon tail; but players might curiously or accidentally stumble upon a series of initially invisible blocks that provide a fun little spatial/platforming event and allow the flightless player an alternative. In World, there either isn’t any way to reach its vertical extremes without a cape, or the prevalent ways to do so -- such as the climbable vines that sproute out of some blocks -- are obvious and lack interactive friction. This comparison is also one of the reasons why SMB3’s stages feel so much more mysterious and alive than World’s.
All of this would perhaps be a shade less dissatisfying and confusing if World had a total absence of level design that acknowledged the more creative mechanical side of the cape and what sorts of potential that allows for. On the contrary, it holds at least one significant oddity in the form of its Forest Fortress’ rearmost stretch, wherein players can take a route above the boss door and manipulatively fly across a lava lake that’s spitting up fireballs. Furthermore, a successful player is rewarded with nine 1-ups at the end of this alternate route -- a uniquely generous reward, suggesting that World’s developers saw this challenge as being considerably difficult (as someone who does not consider theirself a master of the cape’s maneuverings, I found it surprisingly easy, and did it on my first try).
One is forced to wonder, if only by the presence of this section, what the developmental process for World was like. Was there an understanding of the cape’s mechanical implications, and, if so, was there not enough time to react to it? Was there an underestimation of how often players would have the cape, and how ably they’d be able to use it? Did the developers think that the ability to fly by itself -- the performative “freedom” it allotted -- was reason enough?
Another disappointment of World is the reduced, relative to SMB3, extent of its stage designs’ creativeness and weirdness. Continual deference to a "classic" can sometimes be nothing more than backwards-looking fetishization, but I believe one of the central reasons why SMB3 is so widely enjoyed is because it made frequent experiments and dives into substantial novelty. Examples of this include its stage with the hounding “sun” enemy (which was reused later on in a very unexpected environment), its couple of semi-aquatic stages where the entire level periodically dips into the dangerous waters, its “giant world” stages that enlarges standard enemies and environmental elements, the vertical stage that makes reference to the original Mario Bros. arcade game by letting players pass from one side of the screen to the other, or the “tank brigade” stage which confronts the player with a scrolling bombardment of launched threats.
World, by comparison, puts most of this sort of material in its Star World or Special Zone stages, letting it appear elsewhere only in very rare and small doses, such as in its Sunken Ghost Ship or Chocolate Island 2 stages, both of which remain fairly uninteresting and/or partially problematic. Even once the player has arrived in the Valley of Bowser, aside from several moments of enemy reusage (mostly surprising because of the enemies’ long preceding absence (e.g., Mega Mole, Banzai Bill)) and the two effective fortress stages, the greater part of the stage design remains conservative and strangely bereft of a sense of mounting or even present tension.
It is worth noting that World makes an interesting change to its overworld maps. SMB3 introduced the map feature, and it was sure to detail the paths one would or could take ahead of time by roads and preexisting stage markers. World elaborates on this letting the player’s progress reshape the land’s geography. Sometimes, all of a map’s paths are invisible, either preemptively (e.g., Forest of Illusion) or permanently (e.g., Chocolate Island). It also will hide the markers for secret stages and then reveal them as silhouettes once the player has cleared a number of nearby stages. This allows novices to have some imaginative relationship with the maps by wondering about the possibilities and witnessing topographical evolutions. It seems that all of the side-scrolling Mario titles since World never developed this approach, and have instead adhered to SMB3’s overworld model.
One question coming off of these criticisms is: how do we critique World’s level design if it’s often undermined by the cape? -- that is, do we analyze the stages on some basic mechanical level divorced from the cape? do we mention the cape in instances where its usage has the most noteworthy results? do we have side-by-side analyses? As I have gone ahead and described stages that I feel warrant mentioning (for good, bad, and less qualitative reasons), I’ve found the second approach to be the most workable and least exhaustively impeding. My current overall impression is that, out of its total number, about a fifth of World’s stages are well-developed and exciting, but only some of these come from the Special Zone.
A popular sentiment is that a good Mario stage has an element of challenge. It’s sometimes believed that this challenge is related to how many times a player dies on a stage before getting it right. Although I share the sentiment about challenge, in the context of World and Mario-like platformers, I associate “challenge” with the stage’s ability to provide the player with spaces that demand dynamic problem-solving and also have a clarity to design that allow for conceivable first-time success (the apparent incompatibility of this association with SMB2J may be a subject for another article). This is why, like Holleman, I think Soda Lake is a fairly good level (see page 99 of Holleman's PDF document), and why I think Tubular (see page 196) is a bad level.
N.B.: Each stage's title links to a screenshot map of it for visual reference.
Vanilla Dome 4: Some of my favorites of World's stages are the ones that seem to reach back to the types of stage designs that compelled momentum-based play in the NES games. This is my favorite traditional type of level design in Mario because it exemplifies my perspective on platforming challenges and puts it in an excited physical context. Vanilla Dome 4 is a string of broken up surfaces, and the skies are host to random firings of bullet bills that start to come out in dense clusters at the end. What’s interesting here is that while the design suggest momentous play, many of the platforms complicate this by being springs. This stage would be better if there were not so many walkways that allowed the player to run far enough to fly. One or two would be enough, placed in trickier spots, but there are four or five, and only one of them is patrolled by a ground-based enemy.
Vanilla Fortress: On top of having a smart layout, it’s made more tolerable as an aquatic level by having design that favors constant horizontal movement, meaning that there are less opportunities for the player to awkwardly float about. This is why it’s my opinion that SMB’s aquatic stages are the best; they’re straight-shots to the end, and so are a sort of continuation of the terranean theme of momentum. The first half limits the player’s room for maneuvering and makes novel use of a mobile obstacle (the swinging ball-and-chain, now underwater, and needing to be contended with under those physical conditions); the second half does the same, with Thwomps situated in ceiling nooks. All throughout are inconvenient smatterings of Fishbones that often come as a surprise out of the screen’s right side. A neat feature is that there is a pipe early on that can only be accessed by small Mario, but instead of being a shortcut it lets the player out at the start of the second half (rather than being let out around its middle, should they have gotten to the pipe at the first half’s end).
Butter Bridge 2: Stage layouts don’t get much simpler than this, but apparent simplicity isn’t a problem when there’s a convergence of mechanics and dynamic enemy placement. Butter Bridge 2 reminds me again of SMB in its calls for momentum through its eminently manageable -- thanks to their formations and speed -- fleets of Super Koopas. This is made more complex by occasional surprising interjections from elevated shell-less Koopas that kick shells towards the player. Significantly, Butter Bridge 2 has a very noticeable dearth of power-ups, investing some unusual amount of consequence in getting hit. For all its successes, Butter Bridge 2 does remind me of how vital and enjoyably frictional Mario and Luigi’s slides during players’ corrective maneuvers were, and how the removal of that in World does seem to be a net-negative.
Soda Lake: This is the only other aquatic level I’ve included on the grounds that it best nullifies those “floating” moments that tend to make aquatic levels in video games undesirable. The main idea here is that the player will be trying to get past deployment centers and firings of Torpedo Teds while avoiding advancing schools of Blurps. This is fun on initial and subsequent playthroughs because the ideal approach (as the more time spent around deployment centers, the more one is exposed to threats) is continual forward movement interspersed by minor positional adjustments. Soda Lake also succeeds because instead of just accumulating the torpedoes’ deployment centers, it offers points of multiple access/progress; thus, the player is forced to make more explicit decisions on the fly about an ideal general route while juggling the minor adjustments.
Cookie Mountain: Cookie Mountain is the one stage whose inclusion I don’t have a concrete reason for. I find it mysteriously pleasant to go through. There is nothing particular to note here, aside from the theme of having Monty Moles burst out of the ground and elevations with less time given for the player to react than was allowed in Yoshi’s Island 2. I suppose the idea in Cookie Mountain is that its openness will compel the player to adopt a “preservation of momentum” approach, and that the periodic emergence of the Monty Moles checks this approach by having the player perform corrective actions. While this is sort of interesting, it’s blunted by the rareness of pits and an overabundance of solid ground. An objection to this might be that the moles would fall into pits and that this would merely create a new, worse problem -- but the moles’ real threat is the suddenness and multiplicity of their emergence, and not any behavioral threat once they do land on the ground. In fact, the less space there is for the moles to walk, the more challenging they are to get past.
Forest Fortress: Holleman criticizes this stage for having an inappropriate escalation of challenge and having its hardest part at its start, and I share that criticism, but I still think that it has its place among World’s better designed stages. Forest Fortress’ first half brings back the idea behind Iggy’s Castle’s latter half -- an auto-scrolling segment with wide pillars that smash down from the ceiling -- and makes it more involving by increasing the number and threat level of its pits, and by putting route-adhered, circular sawblades along the route. The challenge here is finding harmless spots for rest while waiting to see where the next pillar will come down, and timing jumps over pits with respect paid to the sawblades. Forest Fortress’ second half, while not as engaging, has its moments, with several bits that call for the player to quickly jump out of the way of advancing, ground-based sawblades, and also (as noted before) its alternate path to another boss door, only accessible by steady cape-usage.
Forest of Illusion 4: Another stage that recalls SMB with its theme of a nearly ever-present Lakitu and stage design showing a notable presence of vertical pipes to be leaped over. The catch here is that the Lakitu is a “Fishin’ Lakitu,” and at the end of its fishing rod’s line is a 1-up. Grabbing this 1-up is a trick, because if a player does so the Lakitu will begin throwing down Spinies. Novices are undoubtedly the main target of this trick, but it’s also fairly difficult to avoid the 1-up, since the Lakitu dips down a little while it’s following you. A Lakitu appears later on with no rod and assumes its typical role. This stage is held back by a dulling design decision or two. Ideally, the Lakitu would be unreachable aside from shells kicked upwards, or the pits’ width would be lengthened so that jumps would more dramatically expose the player to the Lakitu's lure.
Roy’s Castle: Roy’s Castle has a touch of simplicity in its start that trends towards dullness, but it’s worth including. The majority of it sees the player alighting on a “caterpillar platform” -- a link of brown blocks -- that’s then activated and follows an unforeseeable path, going through a small variety of configurations to make the player do things besides walking to keep up with its motion. This takes place over a bed of lava and, after, a smaller chamber that’s toothed on the top and bottom by spikes, two of the ones on the top detaching and requiring avoidance. After, there’s a “dash” of sorts, about half as long as the preceding section, where players need to cross a bridge and then contracting platforms (not seen since World’s first castle), all while minding fireballs being shot their way by Bowser statues, in addition to fireballs leaping out of lava and diagonally shuttling through a bottomless room. This description perhaps exaggerates the level of difficulty here, but it’s enough to demand an attentiveness and quickness on the player’s part that is lacking in prior castles.
Chocolate Ghost House: Most of World’s ghost houses are most concerned with having the player figure out some obscure sequence of door transitions, and the level design takes a backseat. I like Chocolate Ghost House, though, because most of it is an obstacle course with the weird, interesting element of pits that shift their location. Aside from the Eeries that appear from the right by themselves or in groups, and exhibit a variety of movements (some come in a straight line; some bob up and down; some travel in a sine wave), there is also a Fishin’ Boo -- a more threatening version of the Fishin' Lakitu. It’s invulnerable to all attacks, and the lure at the end of its line is harmful. This creates a fun triangular dynamic: the player needs to bypass the Eeries, but they need to mind how high they jump to avoid the Boo’s lure while preparing for and reacting to the pits.
Chocolate Fortress: The fortress’ first half, with its pits of lava and constant, unpredictable firings of fireballs, seems to be a reference to the latter parts of SMB’s castles. Holleman describes the appeal of this stage by saying, “Chocolate Fortress is really quite repetitive, and yet still fun, because the iteration of ideas in the challenges are done insightfully.” This iterative quality is best observed in the second half’s neatly evolving Thwomp/Thwimp/spike obstacles, which is one of the rare strings of jumping events in the game that seem to make sense with the spin jump. As with Forest of Illusion 4, I hesitated to include Chocolate Fortress. In my opinion, it should have had the fireballs throughout its entirety, excepting the stretch before the second half’s first Thwomp. I also don’t agree with the decision to put Switch Palace blocks below some of the Thwomps, thus (if the player has completed the relevant Palaces) making them unable to fall down. Rather than this rewarding the player with a less taxing yet still engaging result, it makes these parts boring. At least maintaining the fireballs would’ve made the blocks have some actual relieving effect.
Wendy’s Castle: Holleman describes Wendy's Castle best by saying, “...every challenge builds on the last in a way that prepares the player for what’s coming, while still presenting an increasing challenge.” This stage, in a way, a more successful version of the Forest Fortress, at least in the sense that the spatial stakes have been heightened for the first portion wherein the player contends with ramming pillars and circular sawblades fixed to visible routes. Like Forest Fortress also, the second portion is easier than the first, but it is better contrasted: the greater amount of agentive movement allowed is a welcome expansion after the stop-and-go demands of the first portion.
Valley Fortress: The only strike that I think can made against Valley Fortress is that it feels a bit redundant and underdeveloped when considering the precedent of Wendy’s Castle, as both of them share the focal theme of ramming pillars. Valley Fortress’ alteration to the theme is that it has removed the circular sawblades in favor of pits or recesses with spikes or lava, and has, in general, increased the rapidity of the pillars’ movement. This is a decent alteration, but the stage needs more in the way of complicating agents that interact with the pillars so that the challenge is more than one’s own timing pitted against the time of another linear object and the stasis of the hazardous pits.
Valley of Bowser 1: That this is the most maze-like stage in World alone grants it a certain distinction. Valley of Bowser 1 gravitates around a few core elements: tight spaces with little room for avoidant jumps, advancing enemies that appear out of the sides of the screen, and a variety of routes that eliminate any clear sense of there being a “correct” path. There’s no way to get rid of the Mega Mole enemies by yourself unless you have a cape and use the spin attack, and the Moles can occasionally be used as mobile elevations to reach parts of the stage. Valley of Bowser 1 probably would have benefited from an additional enemy to create more of a sense of uncertainty that soon evaporates once the player (correctly) realizes that there mostly likely will be nothing more than Mega Moles and Chargin’ Chucks. It probably also would have benefited from greater diversity in surfaces, since practically the entire stage is rock. The inclusion of icy floors in choice spots would add some welcome excitement to navigation.
Larry’s Castle: I almost included Lemmy’s Castle, from Vanilla Dome, but didn’t because its second half is boringly simplistic and the central complicating agent in its first half, the Magikoopa, is a better developed threat here. Larry’s Castle starts out with a big room that is a huge elaboration on the first theme in Roy’s Castle: a caterpillar platform that starts to move once it’s landed on, becoming, in effect, the tactile level design, and traveling through a hazardous space (here, though, there are moments of footing other than the block-string). Larry’s Castle extends the length of the caterpillar platform, allowing it to create more varied shapes, and the room is host to a bunch of ball-and-chains that prompt minor bursts of dynamic spatial negotiation. The interesting use of the Magikoopa in Lemmy’s Castle was relegated to pretty much just its first moments wherein the player needed to descend a level from a contained walkway of yellow blocks. This was only a real issue if Mario/Luigi was small and unable to perform an effective spin-jump (in this case, the Magikoopa’s projectile attacks were needed to create a hole in the walkway from which to drop down). Larry’s Castle adjusts those yellow-block by 90 degrees, so that the only way to bypass them is with a cape-spin or the Magikoopa’s projectiles. It also inserts a hazardous parade of lava pits, spiked pillars, fireballs, and Dry Bones. Paired with the teleporting and resurrecting Magikoopa, this makes for one of the more exciting stretches in the game outside of the Special Zone.
Way Cool: This is worth including because it’s the most mature stage that deals in the theme of automated objects that the player stands on or clings to. Way Cool divides itself into a potion where the player stands atop a platform affixed to a visible course of linework, during which they out they must figure out which On/Off blocks to hit to adjust the course and so reach its end, and also time jumps in between and around Fuzzies affixed to their own rounded or angular courses; and a portion where the player holds onto course-bound ropes and moves up and down along them to avoid further revolving Fuzzies. Both of these ideas first and respectively came into view with Donut Plains 3 and Cheese Bridge Area, but they never progressed past the introductory phase. Way Cool allows them to develop into dynamically arranged elements that will have players juggling methods of avoidance. In the case of the Fuzzies, this is done significantly by pairing Fuzzies on separate or coupled courses that run low enough to make simply ducking impossible. Unfortunately, Way Cool begins with a platform that’s more than long enough to fly off from, and the entire stage can be skipped by perpetually controlling your altitude. If such a platform had to be included, the developers should have at least waited until after the stage’s first portion to insert it.
Awesome: It’s odd that this ice-based stage’s name isn’t Way Cool. Awesome is a personal favorite for the way it’s built up of compact, distinct challenges that are all affected to varying degrees by everything having an icy, slippery surface. The first third focuses on iterations of sloped recesses that are watched over by Koopas that either kick their shells or hop into them and speed towards the player; the second third retains the shell-kicking Koopas, inserts some hills to put the player at a lower and less convenient vantage point, and has a dense march of enemies near its end; the last third recalls SMB’s 7-3 with its fragmented walkways and fish that leap up from below the screen in arcing waves. Added to this are a few Rexes and Banzai Bills, and the result might be World’s most challenging segment. I don’t mind the availability of a star power-up near this third’s beginning, but it’s disappointing that it’s preceded by enough running room to perform a flight bypass.
Mondo: An instance of level design weirdness that I wish there were more of. Mondo is basically built as a terranean stage, full of consequential recesses, patrolling Koopa Troopas, and tiered mounds, but it’s covered by a body of water that uniquely rises and falls, and has a current that pushes to the left, against the player. Cheep Cheeps inhabit the water and pose an unusual threat as obstacles due to the water’s current, but are vulnerable at its surface and when it lowers to the extent that they’re dumped onto dry land. A few Flying Hammer Brothers occupy Mondo’s skies, two of them in key spots, and -- like the Cheep Cheeps, and for the same reason -- are harder to deal with than normal. The only problem with Mondo (as Holleman points out as well) is that the water’s rising/lowering behavior is tied to a fifteen-seconds timer, instead of being a dynamic reaction to where the player is in the stage at a given moment. This can lead to situations where the level design simply becomes dull, since the most numerous enemies are the Cheep Cheeps, and they can be defeated on land simply by being walked into. This criticism is not to say that “arbitrary” movement for some body of liquid is not workable, but it is to say that the level design needs to respect this design decision if it’s being made, and Mondo doesn’t do such a hot job in this area.
Outrageous: This is World’s hardest stage overall, and one of its most enjoyable. Part of what makes it so enjoyable is that its enemy selection can’t be gotten rid of so easily: the only consistent way to deal with the Wigglers and leaping flames is to have Yoshi to eat them (however, Yoshi cannot eat the smaller flames the leaping flames leave behind), the Bullet Bills are ejected at surprising times and occasionally in packs, the Piranha plants necessitate an avoidance if the player does not have a cape, fire flower, or Yoshi, and the Flying Hammer Brothers protect their sides and undersides by tossed hammers. Basically every bit of Outrageous that can be divided into a “moment” is an iteration special to that moment. It also makes rare use of an intangible environmental detail (the foreground trees trunks) by using them as strategic, fair (thanks to their minimal width) obfuscators. Besides all this is the fact that flying over the entirety of Outrageous requires a discretion that’s found almost nowhere else. Since the “sky” only goes as far up as the stage is visible within the height of one’s television screen, and since a number of spots in Outrageous have enemies or protrusions that come close to the ceiling, players will only find success if they can maintain their flight somewhere right near the top of the screen, or if they can control Mario/Luigi by an intuited rhythmic comprehension of the perpetual flight mechanic's movement and sound effect.
Cheese Bridge Area: This starts out with a short walkway and then a set of platforms, each of which, once jumped on, follows an automated path designated by a line that is spotted by incoming sawblades. It can be reasonably assumed that most players will activate each platform (each higher than the last), since they are immediately confronted by sawblades on the first two, and also for the sake of diversifying their options. Once the player has landed on the third, last, and highest platform, there’s not any reason for them to relocate. The problem here is that the path this platform takes is the easiest of the three, and it has the most rewards. Ideally, the level design here would tempt the player to switch between routes either with threats or rewards, but it does not. An interesting note: it is possible to bypass this section by taking off from the walkway’s edge while on Yoshi and cape-equipped. The appearance and placement of the sawblades exactly matches the player’s descent and positioning, and this allows Yoshi (whose feet are nearly the equivalent of Mario's spin-jump) to rebound off of the sawblades sequentially. It’s the only spot I’ve found in the game where there is a hidden design to the enemy placement that plays into the course of flight.
Chocolate Secret: This stage has a large vertical section that’s made up of a sequence of six alternating slopes. All but one of the slopes is lined with a small procession of Buzzy Beetles. The only conceivable take-away for intent here was that the player would slide down the slopes on Mario’s bottom, taking out the Buzzy Beetles along the way. There are no elements during this section that make that design work, though. The numbers of Buzzy Beetles in each row is too small to lead to a 1-up (and it’s impossible to kick one of their shells and follow it quickly enough to gain a 1-up that way), and the only threat to Mario during his slide-descent is a shallow pit on the fourth slope toothed by a couple of harmful plants, which Mario slides over anyway. With no rewards and no threats, this section is inexplicably designed, especially when it’s almost as long vertically as the stage’s latter part is horizontally.
Chocolate Island 4: A puzzling stage for its introduction and heavy usage of a stage-unique platform that doesn’t really change the stakes or act of being on that kind of platform. The platforms are tilted at a diagonal angle, seeming to imply that being on them would cause Mario to slide, but this is not the case, even when Mario crouches. Ultimately, the only thing these platforms change is some assumptive temporal approach on the player’s part -- they look weird, and so the player might behave unusually (perhaps timing jumps between platforms more carefully than necessary, or jumping a lot while on the platforms). However, this won’t last long when the player realizes that the platform’s shape is a slope without physical consequence.
Valley of Bowser 2: On top of being World's worst stage, Valley of Bowser 2 is made all the more unenjoyable by it needing to be revisited to find its secondary exit. Its first third is a sort of reformation of a minor section in Donut Plains 2 with rising and lowering impdiments, now a little trickier because the returning bat enemy is initially invisible during its descent from the ceiling. What makes the stage so unfun is its second third: a horizontally inclined network of passages within which is a rising and lowering partition, and a space in between it for player advancement. This is poorly implemented idea because it merges an objective requirement (you will either make it to the next safe spot or you won’t) with the possibility for failure that is not the player’s fault, as there are moments of multiple presented pathways where taking one path will lead to a dead-end and death. On top of this is the portion’s stop-and-go dynamic, combined with the mostly static architecture, which puts a significant limit on player creativity and creates a lot of dead time. Traditionally, Mario levels that impose some reliably meted out limit of progress pair this with some element of dynamism. An example from World itself is Butter Bridge 1, as it complements its automatically scrolling camera with stage design that requires the player to keep moving via thin platforms of varying height that lower when stood upon. In general, I enjoy stages that somehow constrict progressive agency less than stages which do not, but there are ways to make them work better. Valley of Bowser 2 does nothing of the sort.
Star World 5: Star World 5 has a part somewhere around the end of its first third with a P-block and a ?-block. Hitting the latter causes a string of coins to unfold out of its top, providing what is really an as-yet unsolidified walkway. If the P-block is not moved prior to the unfolding of these coins, they will go in the direction opposite to that of the way the player needs to go, meaning that the only way a player can succeed in this situation is to have a cape, hit the P-block, and (first walking away from the required direction, right) take off from the coin-made-block walkway. If the player doesn’t have a cape, they have no apparent recourse aside from dropping into a pit and retrying.
Several things should be clarified now. First, there is no way of getting around that my opinions are informed by having played videogames for more than two decades. As such, my expectations of level design may overestimate players' general abilities, and it may not be unreasonable to critique the list for the absence of stages such as Yoshi's Island 1 or 2 -- stages that are perhaps valuable for the "lessons" they teach through level design to novices. It is likely that many people who played the original Super Mario Bros. were never able to complete it, and found in World a more accessible and enjoyable experience. This is a worthwhile perspective, but one that is not my own. Second, as readers will have noticed, criticisms have been included even for stages that I've deemed successful, such as the lack of fireballs in Chocolate Fortress' second half. I have attempted to keep these critiques, which often contain suggestions for improvement, reasonable. There is always the risk (and this relates to my first addendum) of making hypothetical impositions that tread into the territory of design that is impenetrable to anyone who has not intensely played such-and-such a game for ten years. Third, I have uncharacteristically (for myself) tackled this subject in a list-like, compartmentalized fashion. In my view, speaking about World's level design would ideally take the shape of a unified essay that associates stages' attributes in ways that relate to the analytic premises -- much like what Holleman has done. The risk here is redundancy, so it it my hope that the article can be seen as a shorter alternative to Holleman's intimidatingly, but necessarily, lengthy analysis.