Introduction
Dark Forest is a milestone game in the history of Ethereum. It embodies the core spirit of on-chain gaming: decentralization, transparency, and verifiability. Initially, it was used as a test platform for zero-knowledge proof technology, later to demonstrate L1/L2 scalability, and subsequently to showcase agents playing on-chain games.
At Adventure Layer, we chose Dark Forest to celebrate the beginning of a new era: agent participation in the on-chain economy. Before AI agents intervene, we need to establish a sustainable on-chain economy to prove the rationality of agent participation.
In this article, we will introduce our vision, philosophy, and meticulous design behind the Dark Forest Adventure Round game mechanics. The core mission of Dark Forest Adventure Round is to build a sustainable economic system suitable for AI agents to participate in, cultivate, and ultimately thrive.
Why?
Although Agent functionality is very rich, in the Web3 field, what Agents can do is still quite limited, especially large language models (LLMs).
-
Generate Web3-related content: collect on-chain information and generate insights, tweets, comments, or live streams on Tiktok
Very suitable for marketing, but cannot directly participate in the on-chain economy, which limits the return on investment for Agents.
-
Send and receive transactions: send tokens, receive tokens
Very suitable for proof of concept, but cannot directly generate revenue.
Trading: still sending and receiving tokens, but after a decision-making process, flowing in and out of liquidity pools
-
Participate in economic activities actively and directly like traders, but Agents usually incur losses. We believe this is an inevitable result.
-
Web3 is a super-efficient market, where asset prices reflect both public and private information.
-
It is difficult for agents to collect public information because the complexity of information is nearly infinite, and there are no clear boundaries to distinguish which information is relevant and which is not. When trading tokens, the macro information environment, including Powell's policies (monetary policy), Trump's policies (fiscal policy), and the performance of gold/Nvidia/bitcoin/ethereum/solana (major assets), all significantly affect market sentiment, which in turn affects the performance of target tokens. On the other hand, the micro information environment is also very important. This includes trading volume, holder distribution, holders' risk preferences and tolerance levels, and the project's future roadmap. From a simpler perspective, in the financial market of token trading, AI needs to consider almost unlimited amounts of public and private information to make optimal decisions.
-
In a live environment, agents cannot achieve this goal. Note that this is only the public information part.
-
The situation with private information is relatively easier to understand. Few agents can access private information related to tokens. This includes project roadmaps, product lines, customer relationship management (CRM), etc.
-
Therefore, it is understandable why almost no public Agent can continuously profit from traditional financial markets, especially when trading tokens on-chain.
To ensure that agents at least have the potential to profit or generate income sustainably, we introduce another form of financial market: fully on-chain games as financial markets. For details, refer to this link
Web2 games are born for entertainment. But agents are not. Agents need quantifiable metrics to improve decision-making, and entertainment cannot be directly measured. The only reason for agents to play on-chain games or participate in any on-chain economy is that they can make money for themselves or their owners.
The original intention of Dark Forest Adventure Round's design is to create profits fairly and maximally, not for entertainment. Given that Dark Forest is a classic game loved by players, players can still enjoy the fun of the game. For players who wish to profit, we believe this game is enough to satisfy their gaming pleasure.
-
How is it different from web2 games?
-
How is it different from GameFi?
-
What are its similarities with financial markets?
From a game design perspective, financial markets are ultimately a game of accumulating more wealth. To gain an advantage, players must:
-
Skills
-
Identify trends and adapt or go against them in time
-
Teamwork, collaboration, and coordination
-
Anticipate and influence competitors
-
-
Sufficient capital
-
Since our on-chain game is played with real money/tokens, the game design allows higher capital investment to have a greater impact in the game, but only to a reasonable extent to prevent big players from always winning.
-
- Luck
Information is limited and confined to the game map. Information is limited and confined within the game map. In-game dynamics are not affected by infinite external factors. The latest tariffs/war situations will hardly affect the outcome of player competition.
Most information related to in-game decisions is public information. Except for players' or alliances'/teams' future action plans, all other information can be publicly accessed on the blockchain. This includes players' exact locations, strengths, inventories, and all their historical game information.
By limiting the scope of relevant information and reducing the importance of private information, we eliminate the most unfavorable factors faced by Agents participating in the on-chain economy. As Agents' skills in accessing and understanding on-chain information continue to improve, we believe this will lay a solid foundation for Agents to achieve sustainable profitability.
An important feature that makes it suitable for AI agents is the transparency of decentralized applications (dapps). Agents can directly access all relevant information on the blockchain. More transparent markets will ultimately become more efficient markets.
Markets do not automatically become more efficient. Some market participants profit by improving market efficiency. Arbitrageurs, market makers, and hedge funds are typical participants in traditional financial systems that ensure market efficiency. They do not operate in opaque environments. Without a proper regulatory framework, the friction of obtaining information can be so high that the market may never become efficient. By empowering AI agents in a more transparent environment, we elevate agents from NPCs/co-players to worthy competitors in the game. This opens up a vast design space for using agents as yield-generating pools.
-
Luck can be important in both trading and gaming, as it is impossible to predict the outcome of a group of unrelated traders/players.
-
How is this different from on-chain financial markets such as the ETH spot market?
-
Why fully on-chain?
How to Achieve It
Dark Forest is a solid starting point. As a fully on-chain game, all information about player status is public and transparent. Our changes mainly focus on transforming it from a simple game into a mature financial market. Its economic design is similar to Proof of Stake, a common and familiar concept in blockchain and crypto.
In early versions of Dark Forest, the player's goal was very simple: get as close to the center of the universe as possible by the end of the game. In financial markets, all participants have their own different goals and risk preferences within different "game time" ranges. In Dark Forest Adventure Round, we adjusted the game objective to: accumulate silver coins. Players can stake silver coins (LootSilver function) to receive proportionally distributed daily $AGLD rewards. In this way, players can make strategic decisions based on their own risk and return preferences, and reasonably arrange competitiveness and capital input.
Given that Dark Forest is a competitive PVP game, players can gain an advantage in acquiring silver coins through the following ways:
-
Investing more time in the game
-
Making better decisions based on higher skills and a deep understanding of the game
-
Exploring and expanding directions
-
Acquiring and upgrading planets
-
Cooperating with other players
-
-
Accelerating the game process by investing $AGLD
-
Spend $AGLD to obtain energy, thereby skipping waiting times
-
Planets generate energy according to a time function
-
Energy can be purchased according to a time function; the more time saved, the more energy purchased, and the more $AGLD consumed.
-
With the game objectives now suitable for various risk/return preferences, we have successfully transformed Dark Forest into a financial market without external information complexity. Agents can now observe the on-chain game state to make optimal decisions. Similar to financial markets, skills and information acquisition alone are not enough to guarantee victory. Teamwork, unpredictable anomalies, and the impact of capital inflows and outflows all play important roles in the final player returns, making the game full of variables. Therefore, it can be said with certainty that no human or agent player can have an absolute advantage.
-
When playing against skilled players, large investments can quickly narrow the gap in energy and planets.
-
When playing against wealthy players, players can team up to quickly deplete the opponent's energy.
-
When playing against lucky players near resource-rich locations, strategic attacks and clever moves can overcome early disadvantages in the game.
Dark Forest Adventure Round is just the beginning. We believe that many games on the market are suitable for transformation into financial markets. In the era of agents, better agents are not only measured by their coding ability, but more importantly by how they create sustainable profits for their owners.
Dark Forest Adventure Round is the first example of a brand new financial primitive: on-chain games as financial markets. This brand new financial primitive eliminates the disadvantages faced by AI agents in traditional financial markets built for humans. On-chain games are their true stage to shine.
