By Yijia at Curio Research. Co-authored by 0x992 at Smrti Lab.

There has been a lot of excitement as well as skepticism over on-chain games. Such a level of attention is almost always beneficial. Similar to how Axie was born from the depth of the bear market following CryptoKitties fame in 2017, we could see on-chain games rise to unforeseen heights in the next bull market.

Here I wish to examine 5 key paradigms of on-chain games and how each could be over/underrated. Many experiments will be conducted here at Curio Research, but the space is so wide open we cannot possibly attempt them all. I’m hoping this post can be a source of inspiration for fellow builders excited about similar things, and we’d love to be of help in any way, whether it’s ideation or co-building. Sic itur ad astra.

At Curio Research, we build games with novel modes of player interactions by leveraging unique properties of the blockchain. We’re a bunch of optimistic ideologues, fascinated to build an open and user-driven metaverse from first principles. We’re committed to building world-class products and experience for our players, while making sure that the tools we build also serve the greater gaming community. Only then do we truly succeed at popularizing blockchain games.

------

To start, what do I mean by “on-chain games”?

Being on-chain is a spectrum. A game with wallet logins replacing traditional logins and nothing else can call itself an on-chain game, perhaps barely so. On the highest end, you can imagine a game which logs every pixel-level rendering change on-chain in addition to state changes. Most interesting blockchain games we see today, such as Axie Infinity and Dark Forest, lie somewhere in between.

In this post, I will be focusing on properties of games on the higher end of the spectrum, for example with part or all of game logic on-chain. However, these properties often extend to games with bigger off-chain components, only to a lesser extent.

------

Until when this article is written (May 2022), I observe 5 key paradigms in on-chain games:

Interoperability

For on-chain games, players can meaningfully use a wide array of assets and identities in arbitrary games. Instead of the traditional approach where assets and identities are designed around games, games can now be designed around existing assets and identities. This has a couple of advantages.

For one, interoperability greatly expands the design space of gameplay by combining standalone games into a collective MMO experience. Imagine the following scenario. Each Gargantuar killed in the Plants vs. Zombies' infinite version can unlock tech boosts in Civilization, and new cities captured in Civ generate a corresponding level of sun in the PvZ game. This is a silly example, but you can see how very complex dynamics can arise from combining multiple games in interesting ways.

In addition to game design, interoperability aids user acquisition by giving utility for NFT assets. For identity in particular, a Universal Avatar Format (UAF) can be adopted. Many active NFT collections (BAYC, Azuki, etc.) today have their utility limited to display and trading. Games are perfect arenas for NFT owners to gain nontrivial utility for their NFT assets. Pixel versions of NFT collections such as VOX and Meebits can be built in the same format and plugged into any blockchain games. This is very much like the Warcraft 3 Game editor for avatars, so users don't need to buy an avatar to enter every new game.

On a high level, interoperability connects siloed virtual worlds into a metaverse — a universe of universes, literally — by rendering players’ identities and belongings useful in most or all situations. Ready Player One presents a vision of such a virtual reality metaverse OASIS with a guild system (IOI:  Innovative Online Industries) and cooperation between various IPs such as Gundam, Overwatch, and HALO. Players freely choose their appearances from these IPs. They also carry around their sneakers, weapons, and racing cars more conveniently than the physical world.

Composability (aka. UGC)

Players can freely build content and logic on top of the game client and engine. This is in sharp contrast to games today, where most if not all of game development relies on the studio.

Some of the most popular games and genres in history arose from modding. For instance, the prototype of DOTA came from a player-made map in Warcraft 3. The entire MOBA genre of games, including DOTA chess, then evolved from mods in DOTA. Similarly, Counter-Strike was initially a mod for Half-Life. Jailbreak, one of the most popular games created by high-schoolers, is made by using Roblox Studio. The examples here go on a long list, and they correspond to many major shake-ups in the industry.

A thriving UGC ecosystem can directly result from composability, leading to more sustained player interest and/or shorter production cycles. Consider Dark Forest and its vibrant plugin ecosystem. For a game with likely fewer than 1,000 DAU, there are more than 50 client plugins and multiple full clients, all made by the player communities. It can be exciting to think about logic-side mods and how much activity they empower, once they’re enabled, in addition to client mods.

As much as their appeals are understood, UGC ecosystems often require a long timeframe and careful moderation. Incentives seeking to accelerate ecosystem growth often attract a wrong profile of players and end up hurting the existing community. This may not be the best example, but think for a moment about the many alternative Layer 1 blockchains which give out hundreds of millions of grant money to attract builders for “ecosystem development”, and how counterproductive they have been. Incentives are often too easy to cheat; only through continuous adjustment do platforms get more people to build on top, and time is a major dimension there.

At the core of the timeframe challenge is UGC platforms’ struggle with curation and discovery. When nearly all content can be created by anyone, how do players discover games, mods, and plugins which fit their interest most? How can the community help with the process? Steam’s attempt to normalize paid mods in 2015 had failed largely due to lack of curation — players were angry when they paid for mods which turned out to be unenjoyable.

Openness (aka. Permissionlessness)

All player actions — playing, trading, forking, modding, collaboration in the form of guilds — are available to anyone in the community. Openness, also known as permissionlessness, very much echoes the core web3 ethos.

As a direct result of openness, nontraditional roles and experiences will emerge for players. Imagine an on-chain virtual world. Alice mainly plays instances curated by the community; Bob builds plugins and game mods; Cindy participates in curation and governance, possible without “playing” the main game; Daniel implements an in-game data platform and speculates over its assets. All of these processes can be fun in unique ways, and all can be viewed as “play”. This is wildly different from most traditional games, where players navigate deliberately-crafted worlds and acquire near-identical experiences.  As the complexity of on-chain worlds increases, the diversity of roles in them follows. One day, virtual societies may operate in such vibrant diversity that you may call them societally “Turing-complete”. Some like to call this “non-skeuomorphism”.

Even aside from its utility, the very perception of openness can be a significant attraction for players. As a game without state-of-the-art graphics or gameplay, Minecraft attracts many of its players with precisely its lack of rules and barriers. Today, it’s one of the biggest games in the world with over 130M MAU and many mini-worlds inside it.

At the same time, more roles inherently add more complexity in the balancing of the ecosystem, allowing malicious actors to emerge. Traditional game studios are able to rebalance game elements by looking at how players play. By partly relinquishing the control on game logic, on-chain games have more parts of the game to balance, including gameplay, economy, and governance process. IOI in Ready Player One is a good example of a malicious guild. The interest of Nolan Sorrento, the CEO of Innovative Online Industries (IOI), seeks to control the OASIS himself by inserting intrusive online advertising. IOI uses an army of indentured servants and employees called "Sixers" to find the egg, which is most likely not intended, if even expected, by James Halliday the creator of OASIS.

Incentive Alignment

Similar to their play-to-earn predecessors, on-chain games have an arsenal of financial tools to incentivize constructive player behaviors and to align individual interests with the game ecosystem. This is a native characteristic brought by the inherent financial quality of the blockchain.

At the root of play-to-earn (or play-and-earn and the other of its million euphemisms) is the “reverse freemium” model — Instead of the ad revenue model, which often feels intrusive to players, players can directly obtain NFTs and other utility tokens by proof of work or play. This is already manifested by many play-to-earn games. By postponing the addition of incentives, on-chain games can bootstrap communities around gameplay and reward constructive behaviors more precisely, as opposed to time spent in pure-p2e games. For example, authors of the top community-curated game mods can claim a percentage of trading revenue within those instances.

Compared to existing p2e games, however, more streams of in-game data for on-chain games also bring more opportunities for micro-modeling and exploitation. In “A Theory of Fun”, famous game designer Raph Koster points out that games are fundamentally formal mathematical systems. Therefore, they often rely on a lack of rigorous modeling in order to engage players to the maximum. Visor.gg, a once popular in-game analytical tool for Overwatch, was forcibly shut down by Blizzard because it helped players make decisions in a way that circumvents gameplay challenges. On-chain games have less control in this regard. As soon as on-chain virtual worlds reach fruition, problems like MEV which plague open blockchain systems will likely follow.

Persistence

For the first time in history, games can provide verifiable guarantees for supporting assets which operate under a fixed set of physical laws, forever. This is done with smart contracts which place constraints on how certain game logic can be updated, enabled by standards like Diamond (EIP-2535). These verifiable guarantees give rise to persistent worlds.

Persistent worlds, if successful, are capable of hosting much greater value than transitory worlds. Imagine a million-dollar item in a game instance. It could be a piece of land, a house, a car, or anything important in the context of the game. What does this game need to be to sustain the value of such an item? Centralized games like Fortnite or The Sandbox are unable to supply verifiable guarantees against adverse game rule changes, or worse, game termination. On-chain game instances, on the other hand, protect players against both. As long as the instance stays relevant, the game contract provides a guarantee that the valuable item not only remains in its current state in-game, but operates under consistent rules. This can unlock tremendous total value for on-chain virtual worlds, tantamount to property rights at the founding of the United States. The increased value hosted results from psychological departure for players from the traditional gaming world, which faces many constraints (closed-sourced engines, licenses) and risks (censorship).

Not to say it is an easy task, though. In fact, designing a sustainable persistent world is perhaps one of the most difficult tasks in gaming. Currencies of smaller economies are more prone to failures due to ultra-liquid events. Think about Soros and the 1997 Asian financial crisis. An on-chain virtual world too starts off as a small economy. How do we balance existing versus new player interests? How do we welcome whales to participate while minimizing their negative impacts on economic stability?

------

On-chain games will continue to evolve, and in a very short time we may realize some of these paradigms invalidated, or new relevant ones we have not anticipated. One thing is certain, and that is — the future will only be as imagined if we continue building it.

Special thanks to Aiko Dai, Kydo, and Nico Vereecke for reviewing. Feel free to join our Discord here to discuss ideas and be an early part of our mission.