The first volley is always “It’s aged poorly!” You can envision the scene: probably five or six “|” separators in the bio, perhaps with a “19 | minors DNI,” pfp drawn from the seasonal anime in vogue, 10,000 favs, and a replies section littered with “never understood the hype for this one” and “the later ones are so much better.” I get it; it’s annoying. If you are someone who gives old games their due (and perhaps even prefers them to newer games), then seeing someone write off an old classic because of an unwillingness to engage with a game that lacks the conventions of modernity can feel like a rejection of the retro era as a whole. It’s disheartening to see small roadblocks keep a younger generation from experiencing what were once life-changing works.

However, keep digging and you’ll find a particular counter-volley. A different scene: quote tweet dunk, a bio that advertises a 2,000 follower YouTube channel, custom-drawn header of a Genesis or Neo Geo game, 2,500 favs, and the sneering response “You imbecile! Games don’t age!”. It would be easy to imagine a strict thesis-antithesis argument here: “game aged poorly” versus “game aged well.” Instead, the crowd defending these older titles has contorted themselves into this objectivist, formalist stance in order to make this point. Is this true, that games don’t age?

Let’s Take It Literally

This copyright notice from the Super Mario 64 source code indicates that it was initially completed on May 20th, 1996. [src]

Super Mario 64: a watershed moment for 3D game design, and an equally vexing game for a younger audience raised in the dual-analog tradition. A simple Wikipedia check tells us that the game was initially released on June 23rd, 1996 in Japan only; checking the source on this leads to this VGC article, which seems to be unsourced. A brief search around regarding this date didn’t clarify where it came from, although contemporary publications would likely shed light on this. Of course, this age is a rough estimate, given that there were likely multiple manufacturing runs and each cartridge would have been produced (or “born”) on one of many days leading up to the actual release. If we care more about contents of the cartridge, we could look at the compile date; TCRF contains a screenshot of a copyright notice in the source code from the “Gigaleak” that lists the date as May 20th, as seen above. Regardless, this puts Super Mario 64 at 28 years old, having been kicking for over 10,000 days. This puts it comfortably between “mostly has gotten shit together, young enough to still piss off his parents occasionally but old enough to keep misplacing his lighter” and “been married for a few years, getting career in order before having a first kid” by my estimation, but perhaps this is personal taste. Regardless, Super Mario 64 is literally getting older by the day.

This is pedantry; of course the game is literally growing older, but its contents are affixed in place, a rigid digital signal both locked into read-only memory and disseminated across the web. To start this view, we must first admit that the game’s contents did evolve over time, through a development history that we can say ended sometime in May. From assets to code, the game grew from an embryonic state (initial engine tests, jumping around test rooms and what not) into demos feature-complete enough to show the public, and eventually the team looked at their work and said “it’s done.” From these baby steps all the way up to the finished product, the game morphed and expanded; it gained more lines of code and the assets transformed from sketches to floor-ready art. During development, we must admit that the game aged. It grew up!

In this 1995 prerelease screenshot, Mario walks towards Peach’s Castle, with different texture work and HUD sprites than we would see in the final release. [src]

Unfortunately, it might be clear to the discerning eye that the game actually continue to evolve past the point of that initial build date as well. TCRF’s page for version differences immediately reveals that there were actually revisions past this initial “final” version, including localizations for America, Europe, and China, as well as an updated “Shindou” version for Japan that included Rumble Pak support. Between versions, the primary changes are aesthetic, with most actual bug fixes that would affect mechanics occurring between the Japanese and American releases (with the exception of the infamous Backwards Long Jump, which was fixed in Shindou). This Shindou version would later be updated (via hot-patching instead of actually rebuilding the game) in the Super Mario 3D All-Stars collection, combining its polished engine with a new translation. There’s also Super Mario 64 DS, although I think most would consider that a different version completely. I’ve already written about it on a couple occasions.

These differences may seem slim, and Super Mario 64 is relatively well-preserved across a host of different platforms, but version differences do fundamentally change the experience for other titles. In Elden Ring and other modern Fromsoft titles, boss patches predicated on user feedback after the first couple weeks of release softened their impact, leading some players to scramble to take on the hardest bosses before patches in order to experience them in their original state. Speedruns for Minecraft are broken up into different categories based on the version range (the speedrun.com leaderboards have five separate bins for Any% Glitchless), as continual updates add and alter primary game mechanics that continually necessitate strategy changes. The general move towards live-service incentivizes these practices, from Monster Hunter games offering bounties of items and game-breaking weapons to let players breeze through the early game and catch up to the content drip, to fighting games using seasonal passes to completely alter the metagame year over year (as if this didn’t already happen in the ’90s!). As these games grow older they fundamentally change; their creators are unwilling to let the works rest.

The lobby of Phantasy Star Online: Blue Burst on Ephinea. [src]

MMOs are some of the hardest hit by this phenomenon. On and off over the last couple years I’ve dabbled in Phantasy Star Online, the seminal Dreamcast Diablo clone. I first started playing on the Blue Burst port to PC, which is functionally the “definitive” version of the game released and supported beyond the death of its predecessors. Of course, the game is unsupported now, leaving me to play using a modified client set to use the Ephinea fan server. I later joined some friends on the Episode I & II+ port to Gamecube through the separate schtserv fan server. I was shocked to find the progression was quite different in the latter: dungeon exploration was free-form, but each successive dungeon was locked behind the prior dungeon’s boss. Makes total sense, but Blue Burst had used a different system: progression was gated behind a “government quest” system, segmenting each dungeon to reduce the length of the crawls and provide more break points. A very sensible change had Sega not elected to make every dungeon available from the jump and make restrictions for each difficulty purely level-based; this was the result of a late-era update aiming to make it easier to help new players jump in. Returning to Ephinea due to the convenience of play, I found that lobbies for the lower difficulties tended towards repetitive runs of “Towards the Future”, a boss rush quest generally considered to be the best pure experience point dump in the game. While I’m grateful for the opportunity to experience an otherwise abandoned game with a wealth of random people, I couldn’t help but feel tired at the way the complete negation of the game’s structure created a social environment purely built around grinding towards the end-game. This is aging as a multi-factor process affecting the way a game functions in the present: stacked rereleases creating fans unwilling to play the old content when restarting from scratch, a developer ceding to demands to forego the built-in progression, and the relative small scope of the game making its userbase primarily long-timers with a smattering of curious retro lovers dipping their toes in.

While the argument could be made that Super Mario 64’s maturation has been gradual and generally authentic to its “original” experience, so quickly using this ammo to fight back in a larger conversation about how games age over time is bound to backfire eventually. On a very literal, arguably physical level games evolve over time; they fundamentally age. This is not to say they can’t be perceivable at points in their past (at least in the model we’re discussing now), but it irrevocably changes the way modern players will experience them.

No, I Meant It Figuratively

You’re still rolling your eyes at the pedantry. “Yes, I get it, it’s literally older day after day, and the game might undergo updates over time. What I meant was that the basic mechanics and visuals don’t change; they’re still identical to how they were back when I played it at launch.” For Super Mario 64? Fair point, perhaps. What this group is really fighting to convey is the idea that the bits of the game that some modern-day gamers dislike are just the same as they were back then. To put it in a way that squares with the above, lets rethink “games don’t age” as “games don’t decay.” For a game like Super Mario 64, the composite mechanics and assets remain steadfast even years later. If you press forward on the joystick in a version of Super Mario 64 just like you did back in the day (accounting for other things that do age, such as N64 controllers and their infamously flimsy joysticks), Mario will still react in the same way he always did. I even put forward a sort of objectivist framework based on state spaces that is built upon this idea: a game is a system, it is (loosely) deterministic, and you can map out each “state” (or variable configuration) of the system such that there are actions that change the state.

However, as I discussed in my more recent article on state space analysis from the player’s perspective, the way the player actually perceives a game is not in its totality (as the earlier state space framework suggests), but as an ever-growing mapping. As the player continues to play, they continually grow their mental model of the game and the objective states that lie within it. There is a temporal aspect to this, mirroring that which I discussed above - the player grows alongside the game and continues to understand it more fully - but more importantly for this current discussion, even if the work itself could be rationally, objectively represented via my framework above, one can not instantly perceive it in that way, and there is a chance (an almost definite probability for many games) that the player will never reach a total understanding of the system. This is further complicated by the fact that two players may choose to explore the games in different ways, leading to their own personal understood models of how the game is designed. Communication can ease this, but fundamentally when two players discuss a game, they will be attempting to reconcile their two separate internal models of how the game is constructed, all under the guise of discussing a totalizing, objective model of the game.

David Bleich, an English professor known for his staunch subjective approach to modern literary critique (and his eye-rolling insistence on saying the n-word in full when quoting articles in class, which got him suspended from the University of Rochester), takes a different approach to this topic in his article “The Subjective Character of Critical Interpretation.” He spends a good deal of the middle section criticizing Norman N. Holland’s Dynamics of Literary Response, which attempts to present a model of critique such that the text of a piece of literature not only contains its literal, objective symbols but also an inherent meaning, obtained psychoanalytically, that then influences the response from the reader. Bleich points out an extrapolation of his model, which suggests that this “objective” psychoanalytical meaning can be used to predict reader’s responses to the work. This extrapolation is that, if this embedded meaning is objective (and deterministic, my word, not his), then it can be used to explain all previous responses and predict all future ones, given that it is identically perceived by every single reader of the work. He writes (on page 747, feel free to contact me on Twitter for the PDF):

The latter half of his book, in which he discusses how his model should be applied, delicately backs away from the logical conclusion of his model and begins to confront the problem that in spite of the fact that there are certain general and radical limits of response, there is after all no way - psychoanalytic or otherwise - to analyze the work and then predict the reader’s response from the analysis. For if everyone introjected the same process, everyone would have the same response. We already know that everyone does not have the same response, though there are certain boundaries of possible responses. More importantly, however, there is no way to decide exactly what process is embodied in the work. There is no way to prove that one critic’s process is more definitive than another’s. In principle it is reassuring to say that the work of literature embodies a process, and all we have to do is figure it out through successive abstraction. In reality this cannot be done; indeed, it begs the whole question of an individual’s response, since it assumes that the critic is not viewing the work as a respondent but as a purely objective observer.

Bleich’s ultimate point - that we should jettison the idea of objective meaning embedded in the work and focus only on the objective symbols with a subjective interpretation - can be applied to the state space framework. There is no objective “meaning” embedded in the state space of a game, no matter how broad the player’s representation of it is, only objective states. All interpretation we make of a game’s quality, whether as simple as an assertion on how the controls feel or as complex as an assertion on the representative nature of a given cluster of mechanics, are filtered through our thoughts, our experiences, and our preconceptions on how the game should function. If that were not the case, then we would not have such polarized opinions on a game when all working off of the common objective meaning. When a player says that Super Mario 64 has “aged poorly,” they are using the same figurative speech as you to tell you that they have a modern perception on games, that Super Mario 64 is a era of different conventions, and that the conflict there is unresolvable for them to the extent that it hurts their ability to appreciate the game. They might be a younger age, coddled by the games of their era, or they might be contemporaneous with the game, having aged themselves and changed their perception on what they want from a game. Bleich notes a point in Holland’s Dynamics where he identifies a common, negative critical response related to a work (Richard Lovelace’s “The Scrutinie”) and then backs this up by noting its unpopularity among modern critics and the rarity of its reprints in anthologies; Bleich contrasts this argument with another point Holland makes a few pages later, where he ponders the (anecdotal) fact that “The Scrutinie” has actually undergone a recent reappraisal among his students due to what he postulates are particular shifts in dynamics in the desired portrayal of women. Bleich uses this to illustrate the weakness in Holland’s model’s ability to predict responses based on psychoanalytical objectivity, while we can use it to point towards a moment where the critical response for a work changed based on the changing conventions of the times. Age (or really, the procession of time) factors into the critical equation.

Indeed, in some ways, if one were to look at the phrase “Super Mario 64 aged poorly” and immediately assume that the phrase-sayer was mistaken, that they had found a Super Mario 64 cartridge decaying on the side of the highway like roadkill and looked at it in disgust because of just how rancid it had gotten with age, then I can see why they might respond with an objective assertion to the contrary, that games do not actually decay like meat in the sun, or that “games don’t age.” However, that would imply that this person would have taken the “Super Mario 64 aged poorly” comment literally instead of understanding that the phrase-sayer was figuratively referring to the fact that their critical response to the game has been colored by their inevitable, always growing temporal distance from it (a distinction that the phrase-sayer does not really need to understand in order to get the point across, might I add). If that’s the case, then I’ve already covered that possibility - that the idiom is literal - above. But I will continue on.

On the left here is the Wii Virtual Console version of Super Mario 64, which I first played. On the right is my modern setup, with an EverDrive in a real N64 through a CRT. [left src]

Something my framework doesn’t include is aesthetic elements, and one could argue visual assets equally don’t decay over time. Again, though, if we have to engage in abstractions in order to deal with the perceptive versus objective nature of mechanics, we can more easily understand how there are inherent contextual barriers to objectively processing assets. Look above at the screenshot: it puts the Wii Virtual Console and Nintendo 64 versions of Super Mario 64 in juxtaposition to each other. Given my age (at 25, I fall somewhere between “still losing my lighter” and “getting my career in order” thankfully), I played the Wii version first. Being an emulation of the original, it chooses to upscale the output to 480p (well, I played on a CRT back in the day at 480i, but regardless), and it misses the bar on certain visual effects, such as the dissolve effect when teleporting. It’s easy to criticize Super Mario 64 nowadays for its sharp level edges and lack of proper backgrounds in certain stages, but notice how the lower-res CRT version of this image gives an attractive softening effect that makes the background appear more like mountains rather than polygons (the camera angle helps). Pay close attention to the dirt path and how the blurry texture in the upscale version gives way to the lower sampling of the CRT equivalent, which makes the sediment appear actually crinkly instead of smoothed over. These are subjective measures for sure, and functionally the two look similar to one another, but it clearly indicates that two separate viewing solutions (across two separate platforms) will cause different visual perceptions of a game even when the underlying assets are the same.

At the time this held true as well: composite versus S-video cables, different sizes of CRTs (I was lucky enough to snag this 27" from my granddad recently when he finally switched over to HD televisions for good), different tunings and calibrations (the contrast on mine seems higher than in this Wii screenshot), decay of the tubes and what-not. However, the viewing environment has too been subject to age. While my granddad might have only managed to make the switch away from CRTs in late 2023, most others had completely outfitted their houses with flat-screens by the mid-2010’s. It’s becoming rarer that these properly support 240p in the first place, pushing consumers who would like to use their retro consoles towards console video mods or high-quality upscalers. Modern rereleases of Super Mario 64 upscale the visuals to 720p rather than 480p. We could make an argument on how these visuals were meant to be perceived in the first place (and I find it rather convincing that older titles were not strictly designed for blurry composite signals), but for the purpose of this piece, I’m content merely pointing out that time has continually changed not only what displays are available but what displays are conventional. Further versions of the game follow suit. If we can admit that the abstractions of mechanical analysis are subject to our preconceptions and thus age along with our mental faculties, it should be much easier to admit that the literal vessels of visual output have technologically evolved in the time since Super Mario 64 or any retro game was released.

If we’ve dispensed with our two coarse categories of artifacts from within the game itself (assets and code, or aesthetics and mechanics), then there’s still one avenue we have not explored when it comes to aging. Earlier on I pointed towards a shift in reception for “The Subtletie” as an example of critical reception for a work aging over time. This is only possible due to the shift itself; when one says that a “game has aged poorly,” they are implicitly saying it was “good for the time,” or rather, that the broad reception was once positive, but now seems to be more negative (or at least, it is more negative for the particular person than the broad reception would have led them to believe). A shift is necessary in order to establish that the game has not only aged, but that its critical reception has evolved with the age of the game. For that, let’s quickly take a look at some contemporaneous coverage.

A snippet of Next Generation’s review of Super Mario 64. This is more or less my opinion on the camera system as well, if we’re being honest. [src]

In general, (and I’m just reading off of the Wikipedia page here, for transparency), reviews at the time discussed the visuals and controls (which are one of the most frequent “aged poorly” aspect that I see in modern discussions) in laudatory terms. For instance, GamePro’s review rates the “Graphics” a perfect 5.0, with the included caption “Gorgeous scenery, imaginative (and huge) enemies, and delightful effects add up to the most visually impressive game of all time.” This kind of comparative statement is somewhat destined to not hold up given the constant march of progress, assuming that you look at the aesthetics of games in terms of their technical prowess (which is how I’m choosing to interpret “visually impressive”). Perhaps a more easily disputed statement is in Hyper’s review, which states “The screen resolution is very, very high, giving a sharp, crystal-clear image which is comparable to a [sic, assuming you pronounce it Ess Gee Eye like I do] SGI workstation. The frame-rate is just as high, meaning that the images on screen move more smoothly than any game before.” Again, the Wii rerelease of this game featured a resolution four times larger space-wise than the original, at a time when resolutions went up to 1080p and 480p was somewhat less respectable. This is technological progress - and thus, age, - rearing its head; in fact, 240p was already trumped by most of the 3D arcade content being released at the time, lest we forget that Virtua Fighter 3 released the same year as Super Mario 64, and the latter’s 30 frames-per-second refresh rate was rather consistent but only impressive given the polygon output and the free-roam structure.

There’s one primary aspect that has certainly not aged in the broader critical sense: the camera. Super Mario 64’s infamous semi-automated ratcheting camera system was certainly innovative, but even at the time it received criticism. Just check out the screenshot above, which offers an even-handed assessment: that the camera occasionally fails in both the automated and manual modes, that it often moves in uncomfortable ways while the player is moving, and that it can certainly be learned and grappled with if the player invests the time. Perhaps nicer than the blanket “this camera is unplayably bad” assessments you’ll see on Twitter, but regardless, it was already clear at the time that the camera was a hurdle to enjoying the game. So on this one point I will cede: if someone were to tell you that Super Mario 64’s camera “aged poorly,” I believe you have free license to proclaim that “that’s impossible, everyone knew it was bad at the time already!” Of course, if your real goal behind saying that the game “hasn’t aged” is to defend its honor, then perhaps this isn’t that useful of a concession. But you at least can feel content in knowing that, broadly, both modern and contemporary audiences alike agree/agreed on the issues with the camera.

Ultimately, across all three of these branches of discussion, I am leaning on a subjectivist comprehension of response and critique. You, of course, may disagree, as many have throughout the history of critical theory. However, the more you alternatively lean on objectivism, the more you must contend with the fact that you may be arguing with those who deploy aging through the lens of personal perception. Sure, you can use the opportunity to try to sway them towards your side (assuming that a 17-year-old or whatever who is writing a thread about Super Mario 64 “aging poorly” has any philosophical conception of critique or has an articulated stance on the matter), but when you fight a subjective battle on objective grounds, you’re effectively making a “gotcha” about their lack of precision of terms, and if pedantry is what we’re after, again, please read the above section. If you’re willing to view the debate through a subjective lens, then you’ll be swept into a much murkier world, one where we can grasp at the objective text of our art form from behind the barrier of a limited, perceived state space and a constantly evolving, individualized viewing environment. At least if you subscribe to this view you’ll have one bit of ammo: feel free to tell anyone that reception to the Super Mario 64 camera was mixed to begin with. That’s verifiably true!

Don’t Live in Fear

The visual comparison bit from above is a bit of a detour, but it’s a necessary one from my perspective. A couple years ago, I got into an argument with someone who posted a review of Daytona USA HD due to their insistence on the upscaled graphics ruining the playing experience. His contention was primarily about the crowds in the stands: which, as you can see in the screenshot below, are scarcely more than Seurat-esque dots of color flatly applied to a slope. By upscaling the original 480p image to high definition, what should have been smudged under the lower resolution has now been made garishly obvious. I argued the opposite: the crowds always looked like that, but he just had been able to ignore it better on an old display. Things got heated; he removed the review.

Clearly the upscaling in Daytona USA HD has made the mottled “crowd” texture more apparent. The reviewers at Push Square seem to agree, noting that “…the high-definition overhaul has left Daytona USA looking a little weird." [src]

This sounds a lot like the “aged poorly”/“can’t age” argument, except I’m on the “not aging” side. On the contrary: I really do agree that the crowds look ugly (especially in a screenshot without the blistering speed helping blur them), and I think they do blend better on a CRT (hence why I try to play on one when I can). However, this is the march of time, and I’m unwilling to hold the game’s age against it. Likewise, this person couldn’t reconcile the thought that the game had aged; he refused to accept any other version than the original, preserved at a moment in its original environment, seen as it was “meant” to be seen. This reaction is fear. It is fear that a game like Daytona USA, a graphical powerhouse like none other at the time and a showcase for the legendary SEGA Model 2 arcade platform, could age to the point of being an early, “rudimentary” effort. Its legacy could only be maintained if it was trapped in amber, right at the moment of first perception, never to age.

I have played Super Mario 64 many times since I was a child, in many contexts, on many platforms, using many different displays. I have aged probably a little under 20 years since I first played it, and in that time played many other 3D platformers, both old and new. I’ll never be able to experience Super Mario 64 as I first did, naive to the possibilities of 3D space and ignorant of the many puzzles and secrets housed in the game. When I play now, I go through the motions, completing trivial tasks that I have explored over and over again through the years. I fiddle with the movement system and try new things in old areas, still sometimes jumping around randomly to kill time. I can see now how plain most of the visual design is, even compared to direct successors like Banjo-Kazooie and its richly decorated environments and backdrops (although the very solid 30 frames-per-second of Mario still trumps the wobbly framerate of its more accomplished descendants). When I play it now I may discover new things, but I will discover them in a different frame of mind than I once did, colored not only by my many replays but by many years of thinking about games as a whole and exploring the vastness of the medium.

Super Mario 64 has aged tremendously. When I was a child, I scarcely experimented with the move-set, preferring instead to stick to tried-and-true backflips and long jumps. Knowing what I know now and having so many similar games under my belt, I now understand just how rich Super Mario 64’s movement system is; just how rare it is to find a game that willingly showers upon its players various tools with little context-dependence or strict movement chains a la Super Mario Odyssey. I now can see how invigorating the open-ended structure of Super Mario 64 is compared to what came after, using collectables as a means to an end rather than the selling point its followers posed it as and avoiding emphasizing its own critical path. When I see modern platformers that lean on prescribed, outlined execution challenges, I realize the appeal of the primitive, sandbox construction of Super Mario 64’s levels, which often leverage simple geometry in concert with the aforementioned movement to yield free-form solution paths. There is a reason that modern games such as Pseudoregalia hearken back to it: by paying respects to one’s mechanical forebears, you contrast yourself against your contemporaries (ignoring the glut of nostalgia-bait indies lol) and recall a context that produced creatively explosive works. The ’90s are so in right now; why are we scared to admit that Super Mario 64 is a product of its time?

When you immediately jump to the defense of an old game with the claim that it is incapable of aging, you are betraying your own fear that it might truly be able to age. You are denying it the right to grow with a new audience, to morph in new contexts, and to provide new insights into its power as a work that may not have evident before. Age is not something to fear. Indeed, Super Mario 64’s age is one of its greatest strengths: it may seem unfamiliar and unconventional to a modern audience, but by that same token it exposes the institutional mechanical consolidation of the present through its boundless embrace of what was once a new horizon in gaming. That means that it will not always appeal to those unwilling to accept it, who want to cling to the present; we can be at peace with this. It is better for Super Mario 64 to continue to age beautifully and be continually interrogated than to be imprisoned in a singular moment in time.

---