Tales of the Rampant Coyote

Adventures in Indie Gaming!

Would “Sword Art Online” Be a Bad VR Game?

Posted by Rampant Coyote on November 7, 2022

In the fiction, November 6, 2022 was the launch day for the game “Sword Art Online,” chronicled in the books, manga, and anime series of that name. In a launch day disaster that makes the Ford Edsel, New Coke, and the Microsoft Zune look like speed bumps in comparison, the players become trapped in VR and the headset will fry their brain if they die or if the headset is tampered with. Thanks to a crazy head engineer with a god complex, ten thousand players are now playing an ultra-realistic fantasy game in a virtual world with real-world stakes.

Now the real world date has caught up with the science fiction, as it always does. When the stories were first penned twenty years ago, I’m sure 2022 seemed a long way away, and Virtual Reality as we now enjoy it probably seemed almost as distant as the Matrix-like “Full Dive” experience the author described. The closest we had was a short-lived “Virtuality” arcade systems in the mid-to-late 1990s, with something like 5 frames per second with flat-shaded, untextured polygons. (Which gave me VR sickness back then, too…) Oh, and the disappointment of Nintendo’s “Virtual Boy.” But Massively Multiplayer Online RPGs were the hot new thing (EverQuest launched in 1999, and World of Warcraft was only on the horizon). So author Reki Kawahara drew on his familiarity with the games of the day and their systems to create what he thought would be a super-cool virtual reality game.

Once consumer VR hit in 2015, we discovered a lot about VR game design. A lot of what we learned was how the timeworn conventions and standard practices of 2D projection (“pancake”) gaming played with a game controller differed from the optimal VR experience. Now, those who have actually followed this blog for a long time know that I am a big fan of experimentation and don’t believe that just because something has become a convention means it is the “best.” We are a long way off from being able to say anything definitive about what is truly the “best” VR experience, and the technology continues to evolve and change. It could be that full-dive VR, if it ever happens, will have a whole new set of rules that will have little in common with the best we can do now.

Let’s say Sword Art Online did release today using current VR technology. Minus, of course, the “death game” element. Would it be a good VR game? Or a bad one? Let’s take a look at a few factors. Not all of these are VR specific, but I’ll focus on that.

High Stakes

Obviously, killing off your player base is not a plan for long-term success. But what about having high stakes? We have many games now with character permadeath (or as an option, often called “iron man” or “hardcore” mode). We’ve had MMORPGs with some pretty high “death penalties.” Higher stakes certainly raise the thrill of a game. But with the enhanced thrill of winning, so comes the increased frustration with losing. Higher stakes encourages a slower, more methodical, and more “grindy” experience. In general, a subset of gamers really enjoy these higher stakes (I often do), but it doesn’t have broad appeal. Most MMOs have reduced the “death penalty” to a minor inconvenience that doesn’t do much more than prevent players from brute-forcing combats with respawned characters.

Not much else is said about the death mechanic in the other game worlds of the series. In Alfheim Online, players would (originally) respawn back in their home city, which meant onerous travel back to the combat zones if they were not resurrected on time. In Gun Gale Online, it is a massively PvP environment, and the stakes (outside of tournaments) are unknown, but unlikely to be very high. In the spinoff series, one emotionally unstable character really wants to raise these stakes herself, planning suicide if she dies in a tournament.

Are higher stakes good or bad? It really depends. I think character permadeath (my watered-down version of Aincrad’s lethal stakes) is bad for building the community that you want in an MMO, though it can work great in smaller-scale multiplayer games. Below that… it really depends on how narrow and hardcore you want your audience to be. A lower death penalty retains more casual players. I’m leaving this one in the fuzzy neutral territory.

World Scope

Aincrad consisted of 100 hand-crafted levels of tremendous detail, down to the taste of individual foods that were largely unique across the dozens of eateries in each town (of which there were usually several on each floor). Each floor had its own theme and style. The first floor was 10km in diameter, and each subsequent floor was a little smaller. The first floor therefore had a radius of 5 km, which means its playable area was approximately 78.5 square kilometers. We also learn that there are some massive dungeons that appeared below the first level over time, although a certain guild tried to keep those to itself.

By comparison, Skyrim VR (a great if clunky port, especially with mods) has a total map size of about half that, at 37 square kilometers, about a quarter of which is unplayable “border.” Not including dungeons and expansion territory. I believe that Skyrim is a mix of hand-crafted and procedurally (or at least tools-generated) content, and it could be a pretty fair comparison. Anyway, bottom-line, we’re talking about 150-200 Skyrims of content for all of Aincrad. Which is kind of ridiculous to imagine (especially if every level was fully hand-crafted), especially when we assume that there is a larger library of objects with a higher level of interactivity than Skyrim offers (even modded in VR).

As an MMO, this would spread out the player base pretty badly – especially with only 10,000 players from the first day. This is true in the fiction, as most of the players remain in the starter town on the first floor, or in headquarters near the front-line levels, with a few finding residence scattered across several levels in the late game almost two years later. The first level was probably pretty packed the first month, but after the first several months a solo player like Kirito would be able to do a lot of free space to play solo.

In a traditional game, player movement speeds are exaggerated, in part to compensate for how much smaller things seem on a flat screen. Realistic speeds seem way too slow. However, in VR, this scale changes, and a kilometer actually feels like a kilometer. Especially if you are walking or running them on a treadmill.

While there are some negatives, I think overall players would love to see a VR world with this much scope. Most VR players, finding themselves with a mix of Half-Life: Alyx detail, BoneLab‘s free-form interactivity, and Skyrim’s scope, would never want to log out even if the button is available. This would be an absolute win (but a budgetary nightmare as a developer).

Level-Based Advancement

SAO is kind of inconsistent on how level impacts play. There are some episodes / stories that suggest it is paramount, and there are others that suggest that the player’s own skills, equipment, and meta-gaming can bridge a significant level-gap. In the later game Alfheim Online, level and skills were all but ignored, emphasizing player skill like a shooter or fighting game. Gun Gale Online seemed to take a hybrid approach, with players building up their stats over time to use better equipment.

In general, level-based gaming is great for small groups that can level together, and can act as a “gate” to higher-level content. For a community, it limits the ability of friends to play together. Many MMOs have taken some interesting approaches to get the best of both worlds. In the better VR games, player skill and direct interaction takes on a pretty big role, as it should. Without the character’s stats and growth, you just have an action game rather than an RPG. Without direct player action, you don’t have much of a game. When both work together, you get the reason why people play RPGs in the first place.

“Player skill” shouldn’t be limited to physical action and how you swing a sword or throw a knife, either. As illustrated in the stories, a deep knowledge of the game systems, world, and lore combined with logic and quick thinking should confer an advantage.

Sword Skills

In Aincrad, players could trigger a combo skill as in a fighting game, launching into an unstoppable sequence with often devastating results. In fighting games as in SAO, this has a downside of other players being able to detect and counter these sequences if they knew what they were doing, and the initiating player being unable to stop themselves. This leads to the downfall of Kirito at the end of the Aincrad story arc, and nearly results in Asuna’s death in the recent film Sword Art Online: Aria of a Starless Night.

In a console game, this is cool, because a skilled player can use a very limited interface of a gamepad and a few buttons to launch into some extremely cool moves that look cool and demonstrate mastery from a third-person perspective. In first-person VR, however, you ARE the avatar, and taking control of it away from the player is generally a no-no. At best, it breaks immersion and frustrates the player. Worse, it can cause VR sickness. This is likely a “Twinkie Denial Condition” for a VR MMO.

Cutscenes in an MMO?

In the Sword Art Online: Progressive novels, certain plotlines are punctuated by forced events which are effectively cutscene events. Regardless of the character’s stats or the player actions, triggering one of these events forces certain things to happen to the players and NPCs. This normally only happens in “safe” zones, but one of the books has a transition between locations. Since everything is supposed to happen in real-time, there’s no instantly teleporting to the target destination, and the characters are forced to wait through a long transition… (and, let’s just say Bad Things can happen during this transition when other players aren’t trapped by this unskippable scene…)

Even in the books, this is presented as a pretty terrible idea even without the deadliness of the game. It’s a bad idea outside of VR, too.

Procedural Plot Generator

In the stories, the game system uses AI to generate storylines, NPCs, and rewards procedurally based on real-world literature, folklore, and mythology. Given what we’re seeing with AI-generated art, this technology doesn’t seem as far-fetched as it did five years ago. The results are often kind of bizarre, and can trap players or ruin the game geography (I’d call that a bug…). This meant the entire game could run on autopilot without designer input for a couple of years and still provide new and intriguing (if occasionally bizarre) content.

In theory, the tech is there to do that now. How would it play? As a supplement to hand-created plots and as an alternative to randomly generated fetch-quests, I think that would be a win in VR or pancake games. (The image to the left is AI procedurally generated art based on the Wizard of Oz…). Trapping players or destroying entire levels is probably a no-no and there would have to be failsafes in place to prevent that. But this is kind of a holy grail of game design for me, so I’d love to see it work.

Worldwide Gated Content

Until the “level boss” is defeated, nobody can proceed to the next level. It requires the entire community to work together to allow anyone at all access to more content. I like the concept in theory. It is great for a story. In practice, I think it would be a source of frustration. This is even hinted at by the books, especially when it took an entire month to find and defeat the first floor boss. Those late to the party are forever denied the chance to take on the “real” boss. While I don’t think this is a full-on “Twinkie denial condition,” I think it’s generally a bad idea to have player experiences so heavily influenced by the actions of others.

Non-Private Data

Players have the ability to “peek” over another player’s shoulder at their UI when they have it up. That’s a pretty bad idea. I can see toggling visibility temporarily to validate information (especially when it comes to trades and so forth), but in general these things need to be plainly visible to all, invisible, or selectable by the player. People shouldn’t be able to sneak peeks in their VR headsets while another player isn’t paying attention.

Frauds / Scams / Exploits / Sleep PK / etc.

The SAO stories are at their best when the characters are playing detectives investigating how certain loopholes / scams / exploits in an otherwise robust and literally lethal game system might be taking place. They make for great plot elements in the stories. Obviously, these are intended to be flaws in the game system, and not something a real-world SAO would seek to emulate.

Overuse of Menus / Abstracted Interactions

VR is at its best when the interactions are natural, and feel just like doing a thing in “real life.” In SAO, a lot of the interactions are done by tapping objects and bringing up menus, or simply tapping them. This is an artifact of console games with more limited controls. Even with current technology and the limitations of hand and body tracking, this is unnecessary except with more complex interactions. In general, interactions with objects can and should be a lot more natural, with very few menus. You don’t tap the cream and tap the bread to put cream on it… you pick up the cream and pour it on the bread. While this isn’t a big negative, this would be an unnecessary break in immersion.

Limited Fast-Movement Options

Aside from the one teleport gate in the main city in each level, teleportation or any kind of fast-travel is cost-prohibitive in Aincrad. In general, teleport crystals are so expensive that they are only used in an emergency or to facilitate a major raid. It sounds great for realism, and it can be a lot of fun when you are first exploring, but it quickly becomes tedious. Especially with realistic movement speeds in VR. It’s even worse in an MMO when you want the community to be able to play together, but friends may have to take hours to meet up.


A lot of the enemy actions can toss players in the air. This is even worse than canned avatar actions, and can easily cause VR discomfort. If this is a required gameplay mechanic, a real-world SAO should allow comfort options to prevent them from witnessing their tumbling avatar in the first person. Otherwise, boss fights will be followed by vomit-fests from all but the most jaded VR players.

Aincrad for Modern VR

So where are we? If we were to literally create SAO for modern VR, should it look much like the game from the books, manga, and anime?

Yes and no.  I think the idea of skill activation and canned animations are a bad idea, and getting rid of those–as they are described–might change the feel of the entire system. Maybe. Would diehard fans care? Probably not, especially if they had a reasonable substitute. I think that would be key. Sword skills might allow you to get better numbers if you perform them right, but you never lose control over your avatar.

Interactions should be made more natural than suggested by the books and shows, which would only enhance the experience. Obviously, the flaws and exploits should be avoided.

One of the challenges with VR, as in all MMOs, is maintaining framerate. A ton of PCs and dynamic items filling a scene not only makes it hard to move or interact, it can cause a VR player to become sick. This is a challenge current VR MMOs are tackling. I assume they are doing so successfully, as I’ve never noticed a significant problem in Zenith VR or Orbus VR. Maybe I am just lucky.

And we are getting pretty lucky. We now have a handful of MMOs built from the ground up for VR, with a few more in development. Developers are tackling these kinds of issues, although their budgets are nowhere near what would be required to really build something of the scope and scale of Aincrad.

Ultimately, I believe that Sword Art Online gained its popularity as much for its concept and timing than its execution. VR has long promised the ability to visit other worlds in our lifetime. This concept has been visited many times before (I started with the Dream Park series, by Larry Niven and Steven Barnes), but SAO packaged it up at a time and in a format where we could almost see it happening.  While “full-dive” technology to completely immerse us in VR is still a long way off, we’ve made a giant leap in this direction with modern VR, and I want to take advantage of it.

VR has changed the rules for me for computer RPGs as significantly as CRPGs changed things up from the dice & paper world. Playing an RPG in VR is something else, especially in those few titles that allow cooperative multiplayer. Something like Karnage Chronicles, which is more like a game of Gauntlet than a true RPG, is an absolute blast with friends. I’m still playing Skyrim VR, and it is pretty long-in-the-tooth both for the original version and the VR conversion.

We may never see a “true” VR MMORPG of the scope imagined by these authors. That’s fine. I’ll keep reading. But the next best thing is within our grasp, and I certainly hope to see our VR fantasy worlds grow bigger, better, and more interactive–and in some ways, even better than these science fiction stories promised. At least more fun as a game.

But with a logout button, of course.

Filed Under: Virtual Reality - Comments: Comments are off for this article

Comments are closed.