An Analysis of Platform Games, Part 1

Some context

This piece (and its follow-up) were originally written as a class project; I’ve decided to present my results in hypertext because I need to include images and video. Also, the blog-post format allows for a slightly nicer tone than a formal paper.

The class, as a whole, was about the concept of originality, and the varied ways in which creative works can draw on earlier, preexisting works. The project could be basically anything, as long as I was able to plausibly claim relevancy.

The Data-Driven Approach

At first, the most interesting topic I could think of was to perform a data-driven analysis of… something. I decided fairly quickly that games – specifically, platformers – would be a good group to work with: something I had a decent amount of experience with, and constrained enough in gameplay that I could come up with a slew of questions to ask about any given game without needing to put N/A too many times. Additionally, the genre is old as such things go (and still fairly popular) which gives a nice, wide variety of titles to pick from. This also allows me to include and several very long-running series, and thus explore any design patterns that might show up in them.

To choose the categories on which I would evaluate games, I had a few different goals.

1. Each had to be a valid question for most or all platformers – this ruled out more granular ideas, such as going into detail on character progression systems that would only really appear in a few subgenres.
2. Each had to relate directly to the gameplay – visuals and music, while important, are often limited by technical constraints, and one game can be given any number of artistic styles. For most (all?) platformers, the actions available to the player are much more important, and more objectively comparable, than the aesthetic decisions.

In the end, I settled on twenty different fields. Each is named below, along with its defining question, a summary of what I expected to find, and any elaboration that felt necessary.

Series:

• What series is the game in?
  • My process was to keep adding games as they occurred to me until I ran out of time, and seriality wasn’t the only thing I was trying to cover. As such, there are games with no series here, as well as series where I didn’t include all of the games.
  • I expected to find similarities within a series in excess of similarities between games not in the same series.

Subgenre:

• Does the game fit into any smaller category within “platformers”?
  • These are unavoidably subjective; I came up with four.
  • Action refers to a platformer where the actual platforms are more often a setting for combat than an agility or navigation challenge. Combat is generally some mix of easily-dispatched enemies and significant boss fights, with few or none in a middle-ground.
  • Challenge platformers, in contrast, have little or no combat and a very strong focus on difficult movement-based challenges, each challenge usually being fairly small and contained. High precision and low reaction times are a must.
  • Collectathon is… slightly negative, as a term, I’ll be honest. These games are about hunting through a large world for hidden things. New content is gated behind semi-arbitrary counts of how many such things you’ve already found. Usually, there’s several types of things to collect, and the things you’re finding are more often scattered or hidden, rather than being placed explicitly as the targets of distinct levels or challenges.
  • Metroidvania is, contrary to how it sounds, a fairly widespread name for the category. These are games where the player explores an interconnected world, continually gaining new abilities and/or learning skills and mechanics that allow them to access an ever-increasing portion of that world.
  • I expected the same thing as with Series: More similarities within a subgenre than between separate ones.

Secondary genre:

• Does the game fit into any genres besides “platformer”?
  • The only genre I found among the games I included was RPG, short for “role-playing game” and referring to games with some form of progression of capabilities for the player-character, as well as (usually) a significant narrative aspect.
  • I expected the same thing as with Series: More similarities within a secondary genre than between separate ones.

Year of release:

• What year was the game initially released in?
  • Remakes don’t count, and in cases where the game was released to different markets at different times, I used the earliest. This doesn’t completely pin down the environment at the time of development, since games might take more or less time to make, but still gives enough context for the potential of some sort of progression.
  • Here, I hoped to see general trends over time – perhaps something else would be rare before a given year, and common afterwards, or vice versa. Evidence of developers seeing that something works, and running with it.

Primarily vertical or horizontal:

• In what direction does the player traverse the world?
  • Valid values are Horizontal, Vertical, and Mixed.
  • This, along with Backtracking and Physical Structure, was my attempt to compress all of level-design into something simpler to compare between games.

Backtracking:

• Is the player often required to cross the same areas multiple times?
  • Valid values are Yes and No.
  • This, along with Primarily vertical or horizontal and Physical Structure, was my attempt to compress all of level-design into something simpler to compare between games.

Physical structure:

• Does the player commonly have multiple valid routes by which to progress?
  • Valid values are Open (indicating many choices), Linear (indicating no choices), and Mixed (indicating long alternating periods of one or the other).
  • If there’s only one correct route but also one or more incorrect routes (dead-ends, optional areas, places to return to later), and exploration is required to know which is which, that counts as multiple routes.
  • This, along with Primarily vertical or horizontal and Backtracking, was my attempt to compress all of level-design into something simpler to compare between games.

Jump combat:

• Is the standard “jumping” action also used to dispatch enemies?
  • This is distinct from an action with the primary purpose of attacking, which incidentally causes vertical movement. Valid values are Yes and No.
  • My hope with the three Combat options was to find commonalities in tone, pacing, or focus. I expected Jump combat specifically to indicate a game focused on the running and jumping, enemies as obstacles and platforms in equal measure.

Melee combat:

• Is there an action dedicated to attacking enemies near the player character?
  • For example, a sword-swing, punch, or a projectile with extremely limited range. This does not require that the game visually displays some form of physical contact, only that the attack has a small maximum reach. Valid values are Yes and No.
  • My hope with the three Combat options was to find commonalities in tone, pacing, or focus. I expected Melee combat specifically to indicate a game light on complex player-controlled acrobatics, heavy on flashy automatic stuff, a very hack-and-slash heedless-charge style, perhaps with a combat system approaching that of a traditional fighting game.

Ranged combat:

• Can enemies be fought at a significant distance?
  • For example, guns, archery, or some magic. Projectiles are usually but not always involved. Valid values are Yes, No, and Slightly/Special (e.g. ranged combat using thrown environmental objects which are not always present and can’t be hoarded, or ranged weapons only available in certain circumstances).
  • My hope with the three Combat options was to find commonalities in tone, pacing, or focus. I expected Ranged combat specifically to indicate a game based on cautious exploration. Move forward, see what’s there, shoot it from a distance.

2D/3D:

• Does the game features 2D or 3D gameplay?
  • This is about gameplay, not graphics: If the game displays 3D models, but the player can only move within and interact with a 2D plane, that’s a 2D game. Valid values are 2 and 3.
  • I expected this to mark a major split in any number of other metrics.

Collectibles:

• Does the game feature some objects whose purpose is to be sought and found?
  • A game which visualizes level-completed by placing some shiny thing at the end of each level does not count here; that’s just a differently-shaped finish line. Nor does a game which has health or ammo pickups; those have a use.
  • Upgrades do count, if they don’t provide anything mandatory, and need to be found rather than being bought or given as rewards.
  • Maybe they’re hidden, maybe they’re behind a puzzle or you need to be really good to get them, but there’s lots of them.
  • Valid values are Yes and No.
  • I didn’t expect this to be indicative of anything on its own, but thought it might be interesting to see whether it correlated with anything else.

Power-up:

• Are there things to be found which provide a temporary change in gameplay?
  • You get a thing, it gives you some bonus, but eventually (maybe it’s timed, maybe it’s when you get hurt, maybe it’s until-end-of-level) it wears off.
  • Valid values are Yes and No.
  • This seemed like it would be more common in the sort of “traditional” platformer which I expected to be indicated by Jump combat.

Swimming:

• Does the game feature swimming (or other aquatic traversal mechanics)?
  • Typically, games with swimming either repurpose “jump” as “swim up”, or let you move in any direction and use “jump” as “swim forward” (or “swim faster”).
  • Games without will either not feature water, feature water as an instant-death hazard, or have the player sink to the bottom. In the latter case, if the water is not purely cosmetic, this usually means slower movement, altered jumps (either higher or lower), and maybe a timer.
  • Valid values are Yes, No, and Sink (if it’s purely cosmetic, it counts as “No”, not “Sink”).
  • I didn’t expect this to be indicative of anything on its own, but thought it might be interesting to see whether it correlated with anything else.

Air control:

• How much can the player steer while in the air?
  • This is mostly by feel, but as rough guidelines: Strong indicates that you can fall from above a hazard, avoid it, and land directly below it, or that you can jump moving forwards, reverse direction, and land behind your starting location. Weak indicates enough control to aim for a small platform, but not to do the above acrobatics without a very long fall. None indicates that the trajectory of a jump cannot be modified after leaving the ground.
  • I thought this might partially indicate the difficulty of a game, somewhat biased by the fact that the most recent game I’d played with strong air control was a very difficult high-precision “challenge” game (see Subgenres).

Airjump:

• Can you jump while airborne?
  • Valid values are Yes, No and Slightly/Special (indicating that airjumping is available only in limited situations).
  • I could see this indicating depth – an airjump can be a way to make up for a mistake, or the designer can take advantage of it to design more difficult challenges. From my nonmeasured experience, games with airjumps often had some of each (easier sections not requiring the airjump, so it can be used as a save, and harder sections that need you to save it).

Walljump:

• Can the player jump off of the side of a wall (rather than the ground)?
  • Valid values are Yes, No and Slightly/Special (indicating that walljumping is available only in limited situations).
  • This one seemed like a good indicator of complex acrobatics and fast-paced play, like Air control.

Leeway for errors:

• How much room does the game give the player to make mistakes without being kicked out?
  • This was my attempt to partially quantify difficulty, inspired initially by [a blog post][http://www.vigaroe.com/2017/05/difficulty-yoshis-island.html] by the pseudonymous “Ghoul King”, on the idea of separating a difficult game from a punishing one. More on that later.
  • Valid values are Low (one or zero mistakes), Moderate (two or more), and High (basically infinite, such as games where health regenerates so you’re fine if you don’t rush).
  • Having seen the aformenetioned blog post, I wanted to see how the idea would hold up measured across many games.

Multiplayer:

• Does the main game have multiplayer (not separated as a different game)?
  • Multiplayer minigames separate from the main game don’t count; neither do “race” or “battle” modes on a distinct set of rules and levels from the main game.
  • Valid values are No, Co-op, and Competitive, but I didn’t find any of the latter.
  • I wasn’t sure what I would get out of this, really, but it was the third-simplest one to add (after Year of release and Series) because Wikipedia has a line for it in their standard game template, so it didn’t cost me anything to toss it in just in case something interesting fell out.

Story:

• Does the game have a story, and if so, how much does it matter?
  • This is probably the most subjective field here, because just about every game is going to at least have a tiny introductory cutscene or a few paragraphs in the manual to provide context.
  • Yes indicates that the story is a driving force for the game.
  • Cue indicates that the story doesn’t have much impact beyond an in-character way to give the player hints at what they need to do next.
  • No indicates that whatever story might exist isn’t particularly worth dwelling on, and doesn’t even come up often; the fun lies entirely elsewhere.
  • I expected a trend over time here, with newer games being more able to have stories due to relaxing hardware constraints. Was uncertain whether there would be any correlation, especially given how it’s the least objectively measurable thing on the list.

My data is available in CSV form.

Now what?

Well, this was all how I started my analysis. In part two, I move on to how I finished my analysis, and what I did afterwards. Join me next time for And Why it Didn’t Work!