Why the CD-i Zelda games failed
by on May 9, 2018

Spoiler alert: It’s because the CD-i Zelda games aren’t very good.

In theory, I could just end this article there and be done with it. After all, in the words of the great Shigeru Miyamoto, “A bad game is bad forever.” On the flip side of the argument, however, Spanish philosopher George Santayana told us the popular adage that “those who cannot remember the past are condemned to repeat it.” To push the thought further, it is one thing to know whether something is good or bad, but it is another thing entirely to know why something is good or bad. Game design in the ‘80s and early ‘90s was highly experimental, and good practices hadn’t yet been established. One can see the same experimentalism in the modern era with VR: What makes VR work well, and what deserves to be burned in an alley dumpster fire? Let’s face it: There are a lot of bad games out there. But why they’re bad is an interesting question.

So it merits a closer look at the Unholy Triforce to find out what specifically caused it to fail. Where did the Unholy Triforce go wrong?

CD-i Month is a month-long “celebration” of (or an excuse to poke fun at) the “unholy Triforce” of Zelda games: Link: The Faces of Evil, Zelda: The Wand of Gamelon, and Zelda’s Adventure. These games were released for the Philips CD-i console in the mid-‘90s with Nintendo’s blessing… much to their regret, and now they pretend that they don’t exist. Throughout the month we’ll explore how they came to be, what they did right, what they did wrong (a lot), and their impact on the Zelda community. In short, we want to laugh with — or maybe laugh at — these relics of the past.

Time and money: The developers had neither

If I asked you what the first thing you think of when you hear The Faces of Evil or The Wand of Gamelon, the chances are that you’re immediately thinking of the rather, ahem, “interesting” cutscenes found in either of the titles and just how atrocious they are. Don’t be ashamed; we don’t judge here. However, while we could easily lay the blame at the cutscenes’ feet, that’s really just symptomatic of a more sinister flaw: a lack of budget and external time pressure. In short, Philips had set up the first two games of the Unholy Triforce to fail before development had even begun.

Akin to movies, video games trying to break into the mainstream audience are relatively expensive and time consuming to make. Sure, there are exceptions to the rule. Flappy Bird was a one-hit wonder with just a one-man development team and a short turn around, and one could easily find dozens of similar stories like that one. However, large, epic works require significantly more time and money to craft. From a fiscal perspective, Animation Magic, the developers of the first two Zelda CD-i games, was given a strict budget of $600,000 (US) in 1991, which would equate to about $1.1 million in today’s currency. Sure, the two games weren’t anywhere near as ambitious as the epic games of Halo or Final Fantasy VII, but… how does that stack up with other similar titles?

Even E.T. the Extra-terrestrial, one of the games that caused the Video Game Crash of 1983, cost 57 times the cost of the first two CD-i Zelda games.

Let’s start with one of the worst video games in history — one of the games that single-handedly caused the Video Game Crash of 1983: E.T. the Extra-Terrestrial for the Atari 2600. That game cost an exorbitant $22 million to make ($56.9 million in 2018 dollars), costing almost 57 times the cost it took to make the two Zelda CD-i games. However, if we want to compare it to something more apropos (and good), we can look at Final Fantasy VII for the PS1. The development cost of that game was somewhere north of $40 million ($63.6 million today), which still is many times more expensive than the cost of producing the CD-i games. Granted, Final Fantasy VII was one of the more expensive games to make back in the day. However, tell any game team that you’re giving them just one million to make a game today for a mainstream console game, and you would be laughed at. Depending upon your exact staffing requirements, you’d be hard-pressed to be able to hire many more than 25 full-time employees for a single year and stick within the budgets.

And considering that a little over a year is all the time Philips gave for the development of these two games — and yes, they mandated that both games be developed in a year’s time — the team would no doubt be required to make shortcuts. Add in the fact that this game required features that the Interactive CD format could uniquely provide, such as voice acting and animated cutscenes, and the constraints were sharpened even further. It’s well-known that four Russian animators were flown to the US in order to develop the crazy cutscenes — one of the first Russian outsourcing endeavors now made possible with the fall of the Berlin Wall — in order to make the games fit within the budgetary restrictions.

While Animation Magic did develop two games in a year under budget, some things had to be sacrificed to get it there.

By contrast, the development of A Link to the Past, which came out in 1991, had its roots established three years earlier in 1988. Granted, the first concept of A Link to the Past was originally for the NES instead of the SNES, so there was a hardware change that affected the team, naturally extending their overall timeline. Nevertheless, it still took Nintendo three full years to produce a single game whereas Philips required its team to create two in less than half the time.

A mismanaged, underpowered console

While cartridge-based consoles were king in the United States and Europe throughout much of the ‘90s, compact disc-based consoles were taking off in a fairly big way in Japan. Many companies had started coming out with CD add-ons for consoles even in the late ‘80s. However, those add-ons wouldn’t make their way to North America until 1990, starting with the TurboGrafx-CD, the Commodore CDTV, Sega CD, among others. The Philips CD-i dropped in 1991, right in the middle of the pack, just three months prior to the more popular Sega initiative of the Sega CD add-on.

The first major reason why the CD-i failed to take off admist all that competition is due to its absolutely ludicrous price point. The three most popular consoles (at least in North America) in the fourth generation of consoles – the Super Nintendo, the Sega Genesis, and the TurboGrafx-16 – all started out with a price point at US$200 (roughly $360-400 by today’s standards). Also around this time was the SNK Neo Geo, and that was already excessively stretching the boundaries of consumer wallets at a price point of $650 ($1170 today), and that was the main reason it never made significant inroads with players.

As for the CD-i? You needed to fork over US$1,000 (about $1830 today) just for the console itself. Granted, both the Sega CD and the TurboGrafx-CD add-ons were expensive, both costing US$300 each, raising the overall expenditure — at least at launch — to $500. This means that the CD-i is still twice the price of its major CD-based competitors. Furthermore, unlike the popular practice in the day of including a game with the console, the CD-i didn’t — including only a mere demo disc, meaning you had to add a game on top of your purchase. Sure, you got a controller with it (just one though, when the standard practice was still two), but the earliest versions of the console only supported one controller.

The CD-i would cost about $1830 if sold today, twice the price of its major competitors, the Sega CD and Turbografx-CD.

At that price point, it’s no wonder that it didn’t sell well. However, there was one thing going for the CD-i that other consoles didn’t have, and that was the promise of educational and edutainment software. The CD-i as early as 1992 offered consumers to buy Compton’s Interactive Encyclopedia along with a variety of learning titles, one even featuring the Berenstain Bears and Richard Scary’s Busytown. It was this reason specifically my best friend in elementary school got a CD-i (and thus the reason I have any experience with it at all). It was like investing in a family computer… save that it wasn’t a computer nor even nearly as versatile as one.

Philips CD-i also had two other harsh negatives going for it in the hardware department. The first was that the default controller that came with the console was wireless. This might not sound bad at first blush considering the popularity of wireless controllers now, but things like Bluetooth and RF (the technology behind the Nintendo WaveBird controller) didn’t exist in 1991, so Philips’ default controller had to suffer by using mere infrared communication, the same communication your TV remote uses. Infrared is exceptionally slow and also suffers greatly from line-of-sight issues; you more or less have to point your controller within some narrow cone of the IR receiver or else your inputs will be lost into the ether. While there were wired controllers in some versions of the console (not to mention purchasable for an added price), they weren’t a universal pack-in.

The CD-i came bundled with some of the worst controllers for actually playing games with.

Furthermore, the Genesis and SNES had already begun experimenting with additional buttons on their controllers. The Genesis started out with a three-button controller (with an added Start button for four) and would eventually create a six-button variant. The SNES came out with a six-button controller (plus Start and Select) from the very beginning. While the CD-i wired controller had four buttons, it only had two distinct buttons. Two of the buttons were identical inputs (the “One” button), the third was the other (the “Two” button), and the fourth simply was a shortcut for pressing both buttons at the same time. So not only were you extremely limited in the games you could play because you had to support the default infrared controller, but you also couldn’t showcase a wide variety of actions at the same time.

Lastly, the hardware ultimately would struggle under its own weight as well. While the Sony PlayStation, the first rather successful CD-based console in North America, had the advantage of three more years of technology to use, its specifications were roughly double that of the CD-i. The CPU’s clock speed more than doubled, the amount of onboard RAM doubled, and the CD-ROM drive was able to read discs twice as fast. While the CD-i hardware itself was significantly better than that of the Genesis or SNES, the main reason why the CD add-ons for Sega and TurboGrafx were so expensive was that those add-ons had to add memory and CPU horsepower to the base unit to process the sheer amount of data that could be stored on a CD-ROM. So while the CD-i wasn’t technically a slouch, CDs are an incredibly slow medium, and the CD-i never felt like an extremely solid, fast-paced game platform, especially when compared to the competition of the day.

Frustrating combat that missed the mark

Thus far, we’ve only really talked about why the hardware of the CD-i failed, not why the CD-i games themselves are bad. And jumping into the conversation about the games themselves is tricky because, if one is relatively objective, the CD-i Zelda games are — something you might be surprised at — the best games that were on offer on the CD-i.

Yes, wallow in that horrific revelation.

The CD-i existed primarily in the era before 3D graphics and the hardware to power them had hit the mainstream. As such, pretty much all of the game’s animation — at least from the gameplay perspective — was sprite-based. This isn’t a condemnation on sprite-based graphics by any means; after all, some of the best games in the world use sprite-based graphics. However, the sprite-based animation that the CD-i was able to effect was, admittedly, rather clunky and often very limited. Just look at games that predate the CD-i and the CD-i Zelda titles, notably Sonic the Hedgehog, Super Mario Bros. 3, and many more; they solidly proved that fast-moving sprites and fast-paced gameplay were actual possibilities. However, most of the animation on the CD-i just didn’t have that speed and smooth movement.

Sure, you might have more detailed scenery, but state-of-the-art graphics these aren’t.

All three CD-i Zelda games don’t even come close to 30 frames per second. In fact, I wouldn’t be surprised if their render speeds were less than half that. Most CD-i games were incapable (or at least felt incapable) of rendering quick movement. Whether this is a flaw of the hardware or a lack of imagination on the part of the game designers, we may never know. Ultimately, with this sort of limitation in place, this would make any sort of compelling combat extremely difficult to create (especially given the infrared default controllers!). As a result, players had to have more precise inputs and predictions to keep from taking damage. Add in the fact that many of the enemies’ hitboxes seemed to be rather off or just plain incorrect and you have a recipe for extremely frustrating combat.

The boss battles could have redeemed it. Even if the moment-to-moment combat wasn’t astonishing, amazing boss battles could have won the day and impressed players. However, as with most of the combat, it wasn’t to be. The games’ developers really couldn’t provide many epic experiences whatsoever. At least with The Wand of Gamelon and The Faces of Evil, most of the bosses ended up being a matter of hitting the boss with the correct item once to win, probably a necessary sacrifice in order to save development time. While the boss battles of Zelda’s Adventure are more complex to be sure, those battles would end up suffering from the general combat malaise of the rest of the game as well as having clunky controls and lack of dramatic animations; it would feel like a strongly inferior Link to the Past at best.

While much of the fault for the combat actually is the fault of the hardware instead of the game itself, that doesn’t excuse the developers from doing what they could to surpass those limitations. And given that one of the fundamental pillars of the Zelda franchise is its combat, for the three games to fail so spectacularly in that department speak to the games’ failure.

The elephant in the room: Those cutscenes

As a result, in order to have any chance at competing against the cartridge-based systems, they had to take the fight elsewhere, namely by adding things beyond the physical storage capacity of cartridges. Since CDs had the ability to store much more data, what else other than superior audio and video. Thus, voice acting and full-motion video came to Hyrule. Depending upon what your worldview is, how you feel about the CD-i Zelda games — especially The Faces of Evil and The Wand of Gamelon — has to do with your perception of the Zelda cartoon of years past. Much of the environments, the characters, and the plot have that same level of quirkiness from the cartoons.

Yet all it takes it to look at any still frame from the games’ FMVs to just see how that comes to pass. For the first two games, the animation from Animation Magic proved to be a super distorted view of the characters, effectively providing an over-the-top, wacky take on the Zelda franchise, something that hadn’t really matched the tone or character of the games released prior to them. On the flip side of the coin, there came a point about a decade or so ago where the community’s attitude about the cartoon went from being “so bad and awful” to “so bad, it’s good!” In that light, I honestly do feel that the FMV animations of Gamelon and Evil deserve a little bit of grace. Are they want I would actually want in a Zelda game? No. But they are bombastic and quirky.

Contrast that with Zelda’s Adventure. Adventure has its share of cutscenes, but that game went the real-life acting route instead of drawn animations. While the cutscenes are certainly higher quality, less jarring, and more on point thematically, personally those cutscenes seem to be ultimately lifeless and devoid of emotional impact. There’s really only one character in those cutscenes that actually speaks, and he doesn’t actually interact with the princess in the story other than to provide a sort of narration. Effectively, that character is breaking the fourth wall and talking to the player, providing a sense of story and direction. Ultimately, it makes the game lifeless and bland, not driving or highlighting the other aspects of the story.

Zelda’s Adventure wasn’t much better, but it trades cringe-worthiness for blandness.

As for the voice acting, in the former games, the voice acting is always paired along with one of the FMVs. As a result, most people tend to heap the voice acting on top of the FMV trash fire, thereby destroying it without even truly considering it. This isn’t me saying that the voice acting is good; heavens no! However, it merely becomes impossible to isolate the voice acting when it’s compared with a shoddy script and shoddy visuals to go along with them. In Zelda’s Adventure, on the other hand, most of the voice acting doesn’t have any sort of real visuals to go along with it. Instead, the voice acting is entirely overacted and awkwardly delivered, providing neither substance nor style. Most of it doesn’t add to the story, either wasting your time by saying unimportant trifles or by simply telling you rather obvious facts you might already know.

Ultimately, neither of the games really allow the CD medium for video games to shine, thus removing the last real possibility of competing with Nintendo’s well-established franchise.

A failed console and a game that didn’t understand Zelda

On paper, a CD-based Zelda should have been able to slam dunk the Nintendo-authored Zelda games to make a fun and exciting adventure, much the same way that Twilight Princess on paper should have been able to easily surpass Ocarina of Time. However, Philips failed drastically to really understand what the Zelda series was about and therefore provided an experience that couldn’t compete with Nintendo’s amazing storytelling and quality control. And, in some ways, the games were doomed to fail right out of the gates thanks to the hardware. Overall, the experience would end up being subpar and lackluster. They’re certainly not the worst games out there in all of video game history. And one has to consider that they were literally some of the best stuff to be found on the CD-i, though if the best game a system can muster barely competes with the worst of another, that doesn’t say much.

Philips failed to understand what the Zelda series was about, and the games were doomed to fail thanks to the hardware.

The only thing that saved these games from being declared utter atrocities was the fact that playing the games was so difficult due to the lack of CD-i systems sold and the general unavailability of the games. Back in the earlier days of the Internet and the Zelda community, the CD-i games would invariably only get a cursory mention on fan sites, largely because screenshots, footage, and information about the games were so difficult to find before the days of Twitch and YouTube. And given that the Zelda series has managed to shine again and again as time has passed on, could there really be any other fate for the Unholy Triforce than being the litany of YouTube poop? No wonder why those games are remembered in that way — or, as some might prefer forgotten altogether.

David Johnson
David Johnson, a.k.a. "The Missing Link," was once the webmaster of both Zelda: The Grand Adventures and ZeldaBlog. He works as a software engineer in the games industry. David also pontificates about Zelda, writes features and guides for ZU, and obsesses about CD-i.