Analyzing On-Chain Games: The Future of Complex Applications on the Chain

TL;DR#

Why should we pay attention to on-chain games? (1) In the context of blockchain infrastructure innovation, on-chain games may be the prototype of the next generation of killer applications. (2) On-chain user growth faces bottlenecks, and it is hoped that on-chain games can sustainably bring more active users to the blockchain. (3) With the acceleration of blockchain expansion, the 4G era of blockchain will drive innovation in complex on-chain applications. New public chains, zkEVM/VM, and modular features such as account abstraction, asset management, parallel processing, and ZK characteristics lay the foundation for low unit yields and high TPS in complex on-chain applications.
What is the current state of on-chain games? (1) Currently, on-chain games are in the early stages of exploring product forms and validating the market. So far, there has not been a mature on-chain game, and the playability and completeness of most on-chain games are unsatisfactory. (2) On-chain games still have innovations compared to traditional games and GameFi, such as Dark Forest, Isaac, and Treaty, which feature decentralization, composability, on-chain collaboration, and contractual interactions.
On-chain games can be divided into two categories: ported and native. (1) Ported on-chain games unlock the possibilities of traditional games; simulation and management games are more suitable for being ported to on-chain games due to their composability, asynchronous turn times, and financial game attributes compared to other types of games. (2) Native on-chain games are more likely to bring innovation to complex on-chain applications. Related concepts include hyperstructures and autonomous worlds, which expand the imaginative space of on-chain games.
What is the current state of on-chain game engines? (1) In the early stages of on-chain games, on-chain game engine teams often bear the responsibility of pioneering game products, and the number of engine teams is far fewer than that of game teams, with examples being MUD on Ethereum and Dojo on Starknet. (2) Beyond on-chain game engines, there are other on-chain game infrastructures, such as the gaming guild Guildly and the game store Cartridge.gg.
Why is Starknet a treasure trove for on-chain games? (1) In terms of performance, Starknet can theoretically achieve 9000 TPS, and its STARK proof system can merge 60 million L2 transactions into a single transaction on Ethereum, with Gas potentially as low as 0.001 U. Moreover, STARK has network effects, where transaction costs decrease as transaction volume increases, making Starknet suitable for large-scale on-chain games. (2) Starknet's native account abstraction and contract wallets are better suited for on-chain game interaction scenarios compared to traditional wallets. (3) Starknet emphasizes native application incubation. Its development team StarkWare has held multiple hackathons with partners like MatchBox, Only Dust, Volt Capital, and Mask Network, fostering the development of numerous on-chain game ecosystems.
We believe Starknet is the most important ecosystem for on-chain games, and we will delve into Starknet's technical features, on-chain game ecosystem, and the design and construction of on-chain games in the next research piece.

What are On-Chain Games#

On-chain games (Fully On-chain Game, also known as completely on-chain games) refer to games that exist on the blockchain in the form of contracts. Unlike GameFi, which merely mints in-game assets as tokens, on-chain games store state and execute logic on-chain, thus possessing characteristics of decentralization, permissionlessness, and composability. They tend to pioneer entirely new game mechanisms rather than merely making incremental improvements to existing games or putting items on-chain.

Most incomplete on-chain games simply mint key items in the game as tokens and circulate them on-chain, while game data storage and computation occur on centralized servers off-chain, such as Axie Infinity and StepN. Although they have achieved significant success, these games have some obvious issues:

Limited gameplay and content.
Centralized operation, where project parties can change game rules.
Known for their Ponzi models, often regarded as a continuation of DeFi rather than a new "game category."
Merely a port of off-chain games, not a native on-chain application.

Can on-chain games solve these problems and promise a better future? The answer is: no.

Blockchain is a brand-new technology; instead of using it to replicate, modify, or graft old products, it is better to create a native, entirely new product with it. The problem of horse manure in the era of horse-drawn carriages was never solved but was eliminated with the advent of the automobile era. On-chain games are not born to solve the old problems of Web2 games and incomplete on-chain games; they are committed to complete innovation.

The horse-drawn train of the 19th century may seem absurd today, but it was a compromise of the public psyche, policy laws, road conditions, and product forms at the time. Today's alliance chains and GameFi are also compromises of new technologies against old ecosystems, akin to the horse-drawn train of a new era, but we can all feel that they cannot become the future. New technologies will gradually reshape the entire ecosystem, and on-chain games may be an immature beginning.

Therefore, a more important question is: why do we need to pay attention to and explore on-chain games? What development stage are on-chain games in?

Why Do We Need On-Chain Games?#

【I don't know what this thing will be worth in the future, but it's new, so let's do it first.】

——Max, founder of Astro Aerospace Industries, said

The Inevitability of Technological Evolution#

On-chain games are an inevitable path of technological evolution.

Every technological revolution in history often only erupts into some super products several years later. The secondary innovations and inventions that drive the large-scale popularization of technology are often overlooked, but it is precisely these innovations and inventions that push each other, colliding to produce truly innovative products. Behind every so-called "from 0 to 1" innovative product in history, there are often several underlying technologies.

Behind tanks are oil extraction and refining, steel processing, engines, and explosives; behind smartphones are operating systems, touch screens, processors, sensors; behind AI are computer science, neuroscience, statistics, psychology, and philosophy. However, it is nearly impossible to identify which specific N innovations and inventions ultimately lead to large-scale adoption and change the world, as they are more likely to be a branch on a technological development path rather than the "final fruit."

Therefore, currently immature on-chain games may mature with technological improvements or become a transfer station on another product development path. Regardless, they are worth paying attention to and exploring; as the forefront of on-chain applications, even if they are not the ultimate killer application, they may be in a field closely related to killer applications—just as every innovative product in history has shown us.

Blockchain Needs Active Users#

Blockchain needs application innovation to bring in more active users or to activate existing users.

According to the above chart, we can see that the number of daily active users on Ethereum is facing a bottleneck. We have witnessed the explosion of the DeFi ecosystem, but from an observer's perspective, NFTs and GameFi are continuations of DeFi. Currently, the blockchain lacks applications that can ensure users remain continuously active on-chain and achieve large-scale adoption, making games the target area for further expansion in the blockchain industry. The experiences of Axie Infinity and StepN indicate that GameFi is unsustainable, so more enduring and healthier on-chain games will replace GameFi and become representatives of the next generation of on-chain application innovation. Beyond industry demand, some tech enthusiasts also want to prove that blockchain can achieve more interesting, complex, and immersive applications, not just digital currencies and DeFi.

Complex On-Chain Applications#

On one hand, for blockchain, on-chain games are an inevitable path in the development of the blockchain technology tree, and the blockchain also needs a larger scale of active users.

On the other hand, the latest developments in new public chains, zkEVM/VM, and modularity provide a network environment with higher TPS and lower Gas for blockchain applications, and their superior features such as account abstraction, asset management, parallel processing, and ZK characteristics lay the foundation for application innovation.

Looking back at the development history of Ethereum with low TPS and high Gas, applications with low TPS requirements and high unit yields (where each transaction generates high income, thus being insensitive to high Gas) emerged first, such as DeFi and NFTs. In the early days, they often simply issued tokens, designed Ponzi models, and added "simple" operations to reap substantial wealth.

However, on a network with higher TPS and lower Gas, applications that allow for lower unit yields (where the income per transaction is low, thus tolerating high Gas) and higher TPS requirements—referred to as complex on-chain applications—will emerge. We can establish a coordinate system where the horizontal axis represents TPS and the vertical axis represents unit yield, plotting many applications to understand which applications exist at different TPS/unit yield levels, thereby inferring which applications will emerge on networks with higher performance.

Bas1s Ventures's Wonder believes: a low unit yield high TPS complex on-chain application is similar to a low ARPPU (average revenue per paying user) game/application in Web2, requiring a sufficient number of user transactions to continuously break through the ceiling of the application's business model/token economy—just as China's Meituan and America's EPIC rely on high-frequency transaction volumes to continuously penetrate other low-frequency businesses, ultimately building their own business ecosystems. And a sufficient number of user transactions require a high TPS and low Gas network environment, making it difficult for such business ecosystems to emerge directly on Ethereum, which has high gas fees and limited processing speed, but they can be born in new blockchain infrastructures.

Historically, the transition from 3G to 4G drove the explosion of mobile internet applications. During the deployment of 4G, many questioned the role of 4G technology—what's the use of faster emails and more blogs? What good is more internet speed? However, the subsequent explosion of smartphone applications dispelled these doubts.

Therefore, in the context of blockchain expansion and acceleration, on-chain games may be the beginning of the next generation of complex on-chain applications, and the early exploration of on-chain games may lay the foundation for the emergence of complex on-chain applications. As Jacob describes in hyperstructures, we have the opportunity to build civilizational infrastructure on the blockchain that can outlast our own lives, creating permanently usable, composable, and publicly reusable network applications for all of humanity and its descendants.

Current State of On-Chain Games#

A well-known on-chain game is Dark Forest, a multiplayer online on-chain game with a space conquest theme. Members of dfdao are also building a custom infrastructure called Light Forest to allow anyone to modify and run their own Dark Forest rounds using custom rule sets, creating content like Dark Forest Arena: Grand Prix.

"The game background of Dark Forest is an infinitely generated universe containing various types of planets and space entities. The most popular game mode in Dark Forest is a week-long free-for-all, where thousands of players, bots, AIs, and even smart contracts compete face-to-face for galactic supremacy. Players are born on a tiny home planet in the universe, and they must harvest resources, conquer nearby planets, and interact with neighbors through alliances, trade, negotiation, or even war to expand their empire." ——"Dark Forest: Three Years of Insights from Fully On-Chain Games (Part 1)"

Dark Forest not only runs entirely on-chain but also utilizes ZK-SNARK to help players hide personal information while participating in the game, meaning many decisions in the game are made under conditions of hidden information or information asymmetry.

Dark Forest is a classic case in the rapidly evolving world of on-chain games, while more cutting-edge examples include Isaac, Treaty, and LootRealms.

Isaac is developed by the Topology team on Starknet. It is a multiplayer online physics simulation game based on the backgrounds of "The Three-Body Problem" and "The Wandering Earth," deployed on Starknet. All players need to cooperate to build factory pipelines and power grids, converting natural resources into various devices to drive the planet out of the solar system. Its uniqueness lies in requiring players to collaborate on-chain to achieve a common on-chain goal.

Treaty, developed by the Curio team, is a large-scale multiplayer online sandbox strategy game. Players can act as governors of their own countries in the world, allocating resources, expanding territories, and making strategic decisions about trust and cooperation with other players, similar to a simplified version of "Romance of the Three Kingdoms: Strategy Edition" and "State of Survival" in Web2. Its uniqueness lies in allowing players to draft contracts executed by code with other players, such as embargo contracts and peace treaties. Players are prohibited from attacking other members after joining the contract and paying a membership fee.

LootRealms is developed by Bibliotheca DAO on Starknet, inspired by Loot. It is a sandbox war strategy game where, in the first version, players need to hold Realms NFT to start the game, and in the second version, they can participate by minting adventurer NFTs. Notably, the adventurers' equipment comes from the previously popular Loot. On February 1, 2023, LootRealms raised nearly 4 million USDC (over-subscribed by 6.35 times) in community fundraising. Its uniqueness lies in its high playability and completeness compared to other on-chain games, and it was the first to implement token distribution and on-chain governance. All transactions within the game flow through a network machine called Nexus, with players providing $LORDS as Gas for Nexus operations. Currently, Mask Network's Kaspar is writing an in-depth report for LootRealms, and we can expect his unique insights. Other sandbox simulation type fully on-chain games include Mithraeum and Conquest.

Categories of On-Chain Games#

In this section, we will showcase the subtle differences in on-chain games. From the current state of blockchain computing power and costs, on-chain games should exhibit characteristics of low concurrency to adapt to the inefficiencies of the blockchain. However, from the perspective of complete innovation, this is not the case. There are two potentially overlooked contexts here: porting, which refers to moving previous games to on-chain, and native, which refers to developing a game that can only develop on-chain, if we still call it a "game."

Porting#

Porting refers to a game that can develop off-chain, but we develop a version with on-chain characteristics. As mentioned earlier, an on-chain game will be more composable, reusable, and decentralized than an off-chain game, maximizing its gaming characteristics in certain aspects. Ported on-chain games intuitively only transfer the storage and computation of the game to the chain.

Inspired by Will Robinson's “Unblocking On-Chain Games: Part Two — The 18xx Genre”, I will use the 18xx game as an example to describe ported on-chain games. 18xx is a turn-based simulation game set in 1846, where 3-5 railroad tycoons compete to make money and build the best stock portfolio by investing in and operating railroad companies in the Midwest from 1846 to 1935, which can be understood as a more complex version of Monopoly. Each company has its own treasury for laying tracks or upgrading tracks and purchasing trains, completely independent of the player's personal stock portfolio and bank. The key to the game lies in balancing when to inject funds into your company to take action or fully pay your own and other shareholders' expenses, both of which affect the company's stock price.

18xx games are very suitable for being ported to on-chain games because:

They have sufficient turn time. In each turn, each player needs to make a series of decisions and then submit them at a unified time point, so the TPS requirement for the blockchain is not too high, and transactions can be aggregated or packaged to reduce Gas.
They have financial game attributes. In 18xx, players need to compete with other players in a financial market, such as asset auctions, so it can integrate well with DeFi.

After porting the 18xx game to on-chain, the stocks in the game could be tokens like Uni, Sushi, DAI, etc., in the "real world," and the trading market in the game could be driven by Uni V3; we could also write past game states into contracts—just like what Treaty has done. Theoretically, we could replace 18xx with any simulation management game—such as the previously mentioned Treaty, Mithraeum, Conquest, and Lootrealms.

Native#

Native refers to those games that are "created for the first time" on-chain—if we still call it a "game."

The founder of Dark Forest, gubsheep, mentioned in “The Strongest Crypto Gaming Thesis”:

Cryptonative games are games that maximize the architecture and spirit of blockchain application development:

The true source of game data is the blockchain. The blockchain is not just used as auxiliary storage for data or a "mirror" of data stored on proprietary servers. All meaningful data is stored on the blockchain—not just asset ownership. This allows the game to fully leverage the advantages of programmable blockchains: a transparent data storage that is interoperable without permission.
Game logic and rules are implemented through smart contracts. For example, battles in the game, not just ownership, are on-chain.
The game is developed according to the principles of open ecosystems. Game contracts and accessible game clients are open-source. Third-party developers have the right to customize or even fork their own game experiences through plugins, third-party clients, interoperable smart contracts, or complete redeployments. This, in turn, allows game developers to leverage the creative output of the entire (incentive-aligned) community.
The game is independent of the client. This is closely related to the above three points. A litmus test for whether a game is cryptonative is: "If the client provided by the core developers disappears tomorrow, can the game still be played?" The answer is only affirmative if the game data storage is permissionless, if the game logic can be executed permissionlessly, and if the community can interact with the core smart contracts without relying on the interfaces provided by the core team.
The game contains digital assets with real-world value. The blockchain provides a native API for the concept of value itself, allowing digital assets to interoperate with cryptocurrencies by default. This enables game developers to build new positive-sum incentive structures for their players and developer communities.

The founder of Topology, GuiltyGyoza, quoted Alan Kay and Marshall McLuhan in “On Medium, Validity Rollup, and Digital Physics”: as a brand-new medium, blockchain's most cutting-edge innovators need to create uniquely blockchain-native applications on it, such as "autonomous worlds." More information about autonomous worlds can be found in 0xparc's article on “Autonomous Worlds”.

A simple example of a native on-chain game is the Game of Life GOL2 on Starknet. The Game of Life, also known as Conway's Game of Life, is a cellular automaton invented by British mathematician John Horton Conway in 1970. The rules of the Game of Life can be summarized as follows: cells arranged in a square grid can either be alive or dead. A live cell with 2 or 3 live neighbors can survive; a dead cell with 3 live neighbors can be revived; otherwise, the cell dies.

The game prototype looks something like this:

GoL2 still references the basic rules of the traditional Game of Life, where players create an initial configuration on the grid, and then cells evolve (live or die) according to the game rules, forming various combinations. Players can use GOL2 tokens to restore selected cells or create new games.

GOL2 utilizes blockchain characteristics to generate game content on-chain, but it suffers from overly simplistic game mechanics and insufficient playability—any "achievement" in the game does not provide players with a sense of accomplishment, and perhaps we can look forward to iterations of native on-chain games.

Current State of On-Chain Game Engines#

If the on-chain game market is established, then the game engines that create on-chain games also represent a demand for foundational infrastructure.

However, game engines are directly related to the game production process and user experience, especially when user-generated content (UGC) plays an important role in games. Therefore, in the early stages of on-chain games, which have yet to be validated, on-chain game engines often bear the responsibility of pioneering a game product. There seem to be two investment approaches regarding on-chain game engines that are actually one: whether directly making on-chain games or making on-chain game engines, one must first create a game.

The fact about on-chain game engines is that the teams building them are far fewer than those working on on-chain games.

One example of an on-chain game engine is MUD, which is a development framework for complex applications on Ethereum, but is more commonly used for developing on-chain games.

It adds conventions for organizing data and logic and abstracts away low-level complexities, allowing developers to focus on the functionality of the application. It standardizes the way data is stored on-chain to provide all network code to synchronize contract and client states. This includes synchronizing states directly from RPC nodes or general MUD indexers. MUD is MIT licensed, making it fully open-source and free to use.

On November 22, 2022, the development team behind MUD, Lattice, completed an on-chain MineCraft (Minecraft) called OPCraft using MUD.

Do you remember the previously mentioned autonomous worlds? Lattice considers OPCraft to be an autonomous world: "This is an on-chain 3D voxel world, where every aspect of the world—every river, blade of grass, and patch of snow on a mountaintop—exists on-chain. Every action is executed as an Ethereum transaction in the world." Lattice's idea is to first develop the engine and then develop games through the engine.

On the other hand, the founder of the on-chain game development team Topology on Starknet, GuiltyGyoza, believes that the engine should not be designed first before developing the game, but rather the engine should be designed based on the needs of game development. Topology is a team that integrates knowledge and action; they have already developed on-chain games like Isaac and MuMu, and a new on-chain AI fighting game will undergo internal testing in February. By the way, developers in the Starknet ecosystem are developing an on-chain game toolkit called Dojo, and the development language for Starknet is Cairo, making it similar to the Cairo version of MUD.

Folius Ventures’s Aiko also proposed a physics + chemistry + time game engine in “Sandbox, Simulation Games, and Fully On-Chain Game Engines”, arguing that fully on-chain games based on physical rules are meaningless because they passively let players "discover" logic. In contrast, game physics (the physical rules discovered by players during interactions in the game, such as dropping equipment when hitting the ground) allows for a more interactive experience compared to narrative-driven "passive games," representing a form of "active gaming."

On the other hand, Bas1s Ventures's Wonder stated that the current on-chain game engines mainly transmit conditions for triggering certain key events (such as geographical coordinate information) to the chain in the form of data structures, but a considerable portion of the complete logic is executed on centralized servers. For high-concurrency games like MMORPGs and MOBAs, if all logic is conducted in a decentralized environment, there are high demands for network transmission efficiency and costs to avoid significant negative impacts on the gaming experience during state transitions.

Beyond on-chain game engines, there are other on-chain game infrastructures, such as gaming guilds and game stores.

Guildly is an on-chain gaming guild that allows players to share accounts and NFTs and is currently in development.

Cartridge.gg is a game store on Starknet where players can learn about and access multiple on-chain games, allowing users to create and log in to accounts by scanning QR codes without needing a crypto wallet.

Starknet: The Arsenal for On-Chain Games#

Well-informed readers will notice that many examples mentioned in this article are related to Starknet, which is Ethereum's ZKR (Layer 2 built on ZK technology). Its on-chain game ecosystem stands out in the entire blockchain ecosystem.

Starknet is very suitable for deploying on-chain games in terms of performance. On one hand, Starknet can theoretically achieve 9000 TPS, and its unique STARK proof system can merge 60 million L2 transactions into a single transaction on Ethereum. Its ZK-friendly Cairo language has very high compilation efficiency, and developers can generate zero-knowledge proofs for their Cairo code, with Gas potentially as low as 0.001 U. On the other hand, the STARK proof system used has network effects—average proof size and verification time decrease marginally as proof scale increases, meaning transaction costs decrease as transaction volume increases, making Starknet particularly suitable for on-chain games that require large-scale adoption.

Starknet's contract wallet, designed based on native account abstraction, is also better suited for on-chain game interaction scenarios compared to traditional wallets. Most Ethereum and EVM-compatible public chains/L2s adopt a dual-account design, while Starknet uses a single-account design where all accounts are contracts, granting users high programmability. For example, the commonly used Metamask (traditional EOA wallet) on Ethereum cannot customize specific usage scenarios, functions, or transaction limits for an authorization, nor does it support multi-signature or passwordless payments (session keys), making traditional wallets in blockchains with traditional account designs unsuitable for on-chain game interaction scenarios that require high-frequency authorizations and complex interactions, which is precisely the strength of contract wallets.

Moreover, Starknet emphasizes native application incubation. Its development team StarkWare has hosted multiple hackathons with partners like MatchBox, Only Dust, Nethermind, and Mask Network, promoting the development of numerous on-chain game teams.

However, Starknet has some issues. On one hand, currently, Starknet's sequencers and provers operate in a centralized manner, and we need to wait for the design of decentralized protocols and the open-sourcing of provers in the future to ensure network decentralization. On the other hand, due to network speed limits, Starknet's performance has yet to be unlocked, and high Gas and network congestion occasionally occur, leading to poor user experiences. A mainnet update is expected in Q1, and large projects are not recommended to be deployed directly on Starknet's mainnet in the short term.

Overall, we still believe Starknet is the most promising system in the entire on-chain game ecosystem, and we will delve into Starknet's technical features, on-chain game ecosystem, and the design and construction of on-chain games in the next research piece.

Acknowledgments#

Starknet Astro is the most creative media in the Starknet ecosystem, providing cutting-edge and in-depth research. This research is also supported by Bas1s Ventures for some ideas and content, as they are also one of the most cutting-edge and professional investment institutions in Web3.

Bas1s Ventures was established in 2021 and is a new international venture capital institution in Web3, focusing on investment layouts around the large-scale application and technological ecosystem of blockchain, supported by first-line traditional game companies in Asia and worldwide. Team members have previously worked at well-known companies like Binance, and their investment portfolio includes: Delysium, Mirror World, Highstreet, Xterio, Taker Protocol, and more.