I'm am currently trying to beat the last world in Super Mario Bros.: The Lost Levels (SMB:LL). This game is arguable the hardest 2D Mario platformer ever. It's easy to see that Nintendo increased many of the JUMP heights and distances compared to the original Super Mario Bros (SMB). With smaller platforms to land on too, most of the platforming challenges are more difficult. The more subtle reason SMB:LL is hard is because the levels were designed beyond what the game engine could handle. New elements were implemented that betray the trust/heuristics Nintendo established in Super Mario Bros. But most of all, this game is hard because the level design is crafted to test a high level of nuanced knowledge skills. I'll explain each.
The game engine of SMB.LL is a slightly modified SMB engine. Because the level design pushes the scenarios with more enemies and new effects like wind, in certain areas there are obvious performance issues e.g. slowdown. Aside from this drawback, the flying Parakoopa was modified with a wider hitbox to make it easier for Mario to JUMP off of it in mid air. I'm grateful for the assist, but this change isn't consistent with all Parakoopa in SMB:LL. I'm not sure working against the form-fits-functional design of the Parakoopa was worth it just to use them as mid air platforms. Furthermore, the new forward walking Hammer Bros. occasionally have collision issues causing them to fall right through the solid ground (see World 8-1). Consistency is very important. The more SMB:LL pushes the enemy count and unique elements, the more I notice little out of sync details.
Some of the hardest lessons we'll learn in life involve destroying well known lessons and building up new ones in their place. We're creatures of habit, schedules, and association. This quality allows us to cope with the many demands of everyday life while making us stubbornly resistant to change. It can seem somewhat cruel to have to re-learn a video game lesson that used to be in our favor. So I found Nintendo's implementation of backwards warp zones and Poison Mushrooms in SMB.LL to be especially difficult to adjust to. It seems like the smallest changes like the Piranha Flower motion gave me more trouble than Bowser.
More than anything else the nuances involved with overcoming the mandatory gameplay challenges make SMB:LL far more difficult than other 2D Mario platformers. The SMB Mario engine is already very emergent in that the interactions between the player, enemy, and level elements all affect each other to create a wide range of emergent discoveries and challenges (counterpoint). While SMB's level design is fairly simple in that all of the core (mandatory) challenges are passable without using many nuanced techniques, SMB:LL builds on this foundation with core gameplay challenges that foreground nuance. Instead of occasionally stumbling across an optional emergent trap, every level is designed to increase the likelihood that you'll encounter such nuanced situations. If you're wondering exactly what I mean by simple vs nuance, check out this post I made that defines the power-nuance scale of SMB very clearly.
This is not how we calibrate nuance, but it looks cool.
The best way to calibrate your nuance scales (to greatly reduce the subjectivity involved) is to consider the likelihood of exposure a player can get of a particular detail/property. To begin thinking along these lines you have to understand how a game works from the mechanical level. Remember, nuances are simply particular details or properties of gameplay interactions no matter how minute. Sometimes the slightest details can be the key to developing some of the most powerful strategies. Determining how nuanced a property is can easily be swayed by an individual's personal experiences and biases. So the following is a method of measuring nuance objectively.
It's easy to find or conceive a level challenge where the player is forced to use a specific move/technique to progress. Many games design locks and keys like this to ensure that the player understands key concepts before moving on. Such examples have 100% guaranteed exposure placing the mechanic/technique required to progress very close to if not in the power category. Remember, power is the opposite of nuance. An simple way to think of power is what actions/properties are needed to win.
It gets much harder to determine the likelihood that a player will realize the property of a move/interaction in a non mandatory, emergent scenario. After you limit the scope of your investigation (let's say all of Super Mario Bros levels), and you have a move/technique/property in mind to test (let's say DUCK+JUMP as Super/Fire Mario) you must...
- 1. scan every core challenge to determine if the move is mandatory.
- 2. scan for optional challenges where the move is mandatory.
- 3. scan for emergent cases where the move, or something close to the move, would be highly effective. (This is the most difficult and most subjective step)
- 4. consider execution and replication difficulty (see below).
In SMB, the DUCK+JUMP is a nuanced technique because it's never mandatory for any core or optional challenge. I can only think of a few emergent cases in the game where it would be the useful. Without a need for it, I haven't used the move much. Just recently I realized that if you DUCK+JUMP you cannot unDUCK in mid air. This also means, you cannot SHOOT fire balls are Fire Mario in the air if you're in the DUCK+JUMP state. Who knew!
Beyond these steps, you have to consider how difficult the move/interaction is to execute or replicate. If the interaction you're investigating is only between two enemies that are almost never placed in the same level together, it would be more difficult for the player to test. On the other hand, if the interaction is between a player mechanic and a common level element but is very hard to pull off, the chances are still low. For example, Mario's wall kick glitch can technically be done on many vertical surfaces throughout SMB. However, the precise frame perfect timing you need to execute the technique makes it very difficult to pull off. So even if you technically have many opportunities to stumble across this advanced technique, the chances of you randomly hitting the button precisely on frame and being able to duplicate the feat is slim to none.
After you collect this data, you can more accurately determine and compare the nuance levels between two mechanics, techniques, or properties. This kind of nuance evaluation is centered around a simple educational idea; trial and error. Trial and error is perhaps the most widely use and default learning method we use. An important quality of the trial and error method is that it can be used blindly. In other words, you don't have to know anything about the solution to start trying things, failing, and tweaking the next attempt. It's common to let the game state/scenario dictate the type of trials we'll engage in. This makes sense. When you come up to a new element in a game that would be a great opportunity to test it out. The more often you encounter a challenge that tests certain skills, moves, or concepts, the more likely it is for a "blind learner" to stumble across the correct lesson. That, or discover something closely related.
In part 2, we'll look closely at the World 8 levels in SMB:LL, what we lose when designing core gameplay challenges that focus on nuance, and a Mario game that does it right.