It was just another ordinary Sunday, with about two ongoing research projects in my hands—one related to the company’s research report and the other concerning AI. In my quest for more inspiration and material, I spent roughly two hours browsing the Netease Treasure Trove (Cangbaoge）. I gathered some insights that I’d like to share, and as I kept writing, it evolved into an article, so here it is.
This article will be divided into two parts:
An overview of the differences between web3 game marketplaces and the Treasure Trove, along with recommendations for what a game asset marketplace should be like.
Exploring the concept of “meta-economy interaction” and extending it into the realm of AI and gaming.
1.Inspiration from Treasure Trove for Game Marketplaces
I have a particular interest in transaction taxation, and even our company’s previous deck focused on “IAT” (In-app Taxation) as a new monetization model. However, I had never written about game-specific marketplaces before, and I was somewhat skeptical about the possibility of such a third-party-created and simple game asset aggregation platform thriving in the past, present, and even the near future.
My research on Treasure Trove once again validated my previous thoughts: a centralized gaming marketplaces fundamentally conflicts with a high-frequency, zero-fee pursuit of decentralized transactions.
1)As a first-party self-operated platform, the value of all in-game assets within the Treasure Trove is supported by the player base’s Daily Active Users (DAU) and transaction volume.
It’s evident that the first step in creating a marketplace with demand is to ensure that these assets have a broad and willing audience willing to pay a premium.
2)Customizing the platform doesn’t work well in the fast-changing world of web3.
Customization requires collecting data about what players like and need, which works best for games with a stable and long-lasting player base. However, in web3 games focused on making money, customization might not keep up with the latest trends.
3)Both the platform’s income and user experience improve when things go from inconvenient to convenient, and from low turnover to high turnover.
Having zero transaction fees and high liquidity doesn’t benefit the platform’s revenue or the stability of asset prices in any way. People familiar with NFTs might remember how the emergence of Sudo Swap and Blur deeply affected the NFT market in the past year. Zero transaction fees and bulk trading undermined NFTs as luxury items or symbols of belief.
In our two previous Folius’ GameFi decks, we discussed how increasing friction in transactions can boost revenue. The Treasure Trove also implements several similar designs. For instance, the “Express Transaction Service” can reduce the review time from 4 days to 0 days and the secondary transaction time from 8 days to 0 days. This service charges sellers an additional 4% fee. In the Treasure Trove, both sellers listing items and buyers making purchases incur fees. Another service that benefits buyers is the “Pre-order during the Publicity Period,” allowing them to reserve game assets in advance, but it comes with a 5% reservation fee based on the total price.
4)Incorporating game-like mechanisms that cater to human nature is a part of what a marketplace delivers.
Many web3 NFT marketplaces focus on transaction forms and matchmaking efficiency, and gaming is no exception. However, applying a crypto asset mindset to game asset marketplaces overlooks the strong demand for gamified sales activities inherent in gaming. In the Treasure Trove, some games feature bidding systems where a higher bid increases your chances of winning, essentially turning it into a game of managing expectations. This is likely because of regulatory constraints, making it difficult to implement such mechanisms within the game itself. Moving them to the trading marketplace aligns more naturally and not only increases user transaction frequency but also boosts the platform’s profit margins.
In fact, there have been suggestions in the NFT trading market, similar to those seen in Chinese e-commerce platforms like Taobao Fruit Garden or lottery systems. However, as of now, not many major trading markets have adopted these suggestions. Only Rollbit has successfully incorporated NFT lootbox as a unique gameplay element within its crypto asset-based gambling platform. This represents a significant opportunity for emerging gamified marketplaces.
5)If there’s no existing demand, you can artificially create it.
The difference between the fifth point and the third point is that while the third point relies solely on the platform’s transaction rules, the fifth point emphasizes human intervention. You can gather a group of people to support a new trading scenario.
For example, within the Treasure Trove, there’s an “appraisal"event. The platform invites a group of players who are considered authoritative within the game’s ecosystem to act as appraisers. Sellers pay these appraisers to expedite the sale of items they have appraised. These appraisers can be thought of as experts in cost calculation, helping sellers determine the value of their assets and providing certification.
It’s likely that the platform will also provide more exposure and prominent placement for items appraised by these experts. The platform takes a percentage of the fees paid to the appraisers, which is akin to advertising fees. By enlisting a group of individuals known as appraisers, the platform softens the perception of this fee, making it seem more reasonable and user-friendly.
2.Before we delve into the concept of “meta-economic interaction,” we need to establish a new “dimension of value measurement” for in-game assets.
This thought process was triggered by the transactions within the Treasure Trove:
The items traded within the Treasure Trove are often more than just fragmented assets; they represent a holistic experience. In games with rich and layered experiences, a single weapon or piece of equipment might enhance your power, but it doesn’t fundamentally change the game experience. So why do players choose different characters and replay the same game in some instances? It’s because owning a whole account means experiencing the progress within that account, much like gaining a new life. Therefore, the most valuable assets are those that can fundamentally provide new experiences, whether it’s through unique characters or complete accounts.
In a presentation released by Folius in February, we also emphasized the importance of personal progression. In April, I came across a team that allowed players to sell NFTs containing the progression of their game characters. These NFTs included various levels of unlocked skins and skills. Unfortunately, such interesting designs have been rare to come by lately.
People fixated on fragmented asset units (such as 1155, 6551, etc.) may not realize the key to this. It’s not about how many fractions you can divide an asset into or whether it can be bundled, or whether a tradeable sword is more valuable than a grain of sand. These factors are irrelevant. What’s crucial is identifying the “datasets” that players perceive as valuable and placing these datasets on the blockchain as assets.
Moreover, why are “whole” assets more important than “fragmented” assets? It’s because they encompass multiple dimensions of value.
In-game asset trading can indeed be considered an experience. Many web3 games currently emphasize the freedom of in-game asset trading, similar to the trading of minerals or armies on certain planets. However, building a trading system in this direction, regardless of how large the Daily Active Users (DAU) are, might risk losing something significant.
During the subscription era, we used time (or “proof of work”) as the standard for measuring value. We were concerned with the stability of in-game soft currencies. For instance, when we argued in the past that “Fantasy Westward Journey Online” had a more stable economic system compared to “World of Warcraft,” it was because “Fantasy Westward Journey Online” employed hourly subscription, while “World of Warcraft” relied on a monthly subscription model. As a result, “Fantasy Westward Journey Online” could better stabilize the value of various in-game assets and control the inflation rate of soft currency.
However, today, there are fewer open economic games, and fewer players focus on the prices of individual items. Those who used to pay attention to the stability of item prices differ greatly from players who now spend a lot of money on gacha games. Free-to-Play (F2P) and gacha have become the mainstream, and thus, the measurement of value has evolved from “proof of work” to “proof of capital + proof of luck + proof of work + proof of skill.” In other words, it’s about having the money to spend on gacha, winning, and efficiently cultivating the account with multidimensional value.
In this complex multidimensional value, players can roughly calculate the cost of cultivating such an account. Bringing this account back into the game results basic resources and a unique game experience. Even if you want to sell it, the price is essentially the original price plus the value charged from the new owner. What players gain is time saved and a completely new experience, not direct economic returns. It’s an excellent commercialization.
Furthermore, this dimension of value measurement emerges only after “F2P” era. When games became “unfair” and players can spend money to buy numerical advantages, these values, which contain economic value, are ultimately applied to accounts through randomness and strategy. This creats an entirely new asset class that’s more suitable for trading among F2P players. However, F2P gacha games generate substantial revenue, and everyone shifts their focus to art and industrial capacity. There is little consideration for secondary commercialization in the secondary market, which provides an outlet for the accumulated value of characters.
3.What is “Meta-economic interaction” then?
First, let’s consider a scenario: Which one is more suitable as an experimental ground for AI gaming, “Immortal Conquest” or “Civilization 6”? (Without considering API interfaces, implementation difficulties, etc.)
To start with the conclusion: “Immortal Conquest” is better.
In “Civilization 6,” resources are obtained through a player’s time and technology. Multiplayer games are often played among friends in a local network, with entertainment and passing time being the main objectives. This leads to many informal negotiations and resource exchanges driven by emotions, such as “Can I borrow this?” or “Can you give me that?” Additionally, the diplomatic relations that can be claimed are limited and somewhat distant. Nations can form friendly alliances, but they cannot engage in more complex and deep-seated hierarchical relationships, such as submission and paying tribute.
On the other hand, “Immortal Conquest” has two distinctive features:
- Hero Card Gocha System - In “Immortal Conquest,” a significant portion of military strength and skills depends on heroes, which are obtained through paid Gocha. This dynamic creates a strong economic interaction within the game. For example, players automatically calculate the “cost of drawing a hero to initiate a war” as equal to “the cost of paying salaries to five counties.” Therefore, economic interactions occur more frequently, including actions like sending out-of-game red envelopes or paying salaries.
- Alliance Relationships - As a massively multiplayer online mobile game, “Immortal Conquest” has more intricate social design than a game like “Civilization,” primarily focused on single-player and local network play. The seamless world map of “Immortal Conquest” requires players to connect regions for military campaigns, and cooperation between alliances is crucial. Alliances receive resource allocation buffs, and siege warfare requires the cooperation of players with varying economic power and social status. As a result, strong interdependencies are formed around economics and territories. It’s evident that players’ willingness to spend money significantly impacts the dynamics and overall atmosphere of a multiplayer game, whether for better or worse.
We can see that monetization (paying money) establishes a bridge between the game and reality, and a player’s financial strength and luck in the real world have a substantial influence on the game.
Moreover, bypassing in-game confrontations, there may also be numerous uncountable economic interactions and human interactions because assets are bought by players. Thus, in-game values can reflect a player’s real-world economic power, which can be input into the game as numerical values, used as bargaining chips in negotiations with other players, transferred to another player’s real life, or re-input into the game as numerical values. For example, there’re country manager players who, through WeChat (a social platform), monthly transfer money to county players as salaries.
At this point, we should have a brain map to distinguish between “in-game interactions” and “meta-economic interactions.”
Moreover, what are some interesting use cases for meta-interactions?
—AI Agent Meta-Interactions
Imagine, within the value measurement framework we defined earlier, what aspects could AI participate in?
- Capital (Optimizing capital utilization & out-of-game strategies)
- Time (Saving player time)
- Skill (In-game strategies)
Now that AI has bidding and negotiation capabilities, envision a scenario where AI takes over in-game interactions and external interactions with other players. Each player has their own AI agent, and these agents can communicate with other agents, collectively negotiating conditions and bargaining. Doesn’t that sound intriguing?
Consider the following scenarios:
Each player has an initial agent with consistent game strategy and negotiation abilities.
Players need to allocate funds to their agents for actions such as bribing or buying other agents’ services.
Agents provide players with optional strategies every 6 hours, and players select the optimal strategy for their agent to execute.
Agents can learn players’ risk preferences from their strategy choices and optimize future strategies accordingly (e.g., whether to continue bidding on other agents, deploying troops to attack, etc.).
A player can have multiple agents that can be trained independently for different tasks, such as diplomacy and tactics. Players might find ways to manipulate the game without spending a single penny using skilled agents.
At this point, agents become characters with distinct personalities, appearances, and voices. The players who train these agents can earn a share of their commercialization.
This way, games that traditionally required extensive social interactions like Immortal Conquest can transform into more passive experiences with minimal social interaction (similar to single-player games). Agents can also become central points of interaction for players (cute companion anime character) and may even be commercialized as new game content.
Last but not least, why do I find AI games involving meta-economic interactions interesting?
Perhaps it’s because I’ve been studying incentive mechanisms in crypto field for too long. Maybe it’s influenced by zkML, where I believe future AI natives can manage users’ assets for financial activities. Maybe it’s because I’m not interested in generative/passive AI-human interaction after playing Dwarf Fortress. Perhaps it’s my obsession with the idea that games are interactive art. I just feel that the current state of AI+games interaction is far from where it could be.
It’s undeniable that almost all AI+games actions today are merely testing the AI’s ability to mimic human behavior. “Dota experiment” trains AI in micro-operations, “Minecraft experiment” assesses AI’s understanding of the physical world, “Stanford’s AI Town” simulates AI’s understanding of human behavior and emotions. Then why hasn’t there been an AI experiment simulating human economic behavior in games?
Games have always needed to appeal to human nature. Watching AI’s daily activities may satisfy our curiosity, but how long can that satisfaction last? Does it provide as much entertainment value as graphics cards and VR experiences? Does it compare to the rush of adrenaline and dopamine when opening loot boxes? Can it rival the satisfaction of dominating an entire server and fulfilling one’s vanity? If not, how can AI be utilized to create games that are more in line with human nature and stimulate the human pleasure centers? That’s where I find the intrigue.
Additionally, I have been contemplating and researching AI’s role in narrative games. This is my reflection after completing “The Cosmic Wheel Sisterhood” this week. Storytelling is another aspect where I believe AI can significantly enhance the gaming experience and provide differentiation.
Conclusion
The purpose of this article are:
To delve into the intriguing mechanics within the Treasure Trove marketplace. For teams involved in designing marketplaces, there are valuable insights to be drawn from here. It also sheds light on the distinctions between Treasure Trove and Web3 game-related trading markets, highlighting differences in terms of feasibility and limitations.
To elucidate the novel dimension of game assets value. With the evolution of free-to-play (F2P) gaming and shifts in player consumption patterns, it becomes imperative to shift our perspective away from the subscription. Instead, we should explore various asset trading formats, including the notion of trading a segment of the gaming experience, encompassing multiple valuable characters within an account.
If you find the concept of “meta-economic interactions” in AI+ gaming, as mentioned towards the end, intriguing, and if you’re engaged in AI and gaming-related research, don’t hesitate to reach out. There’s a wealth of research yet to be undertaken.
Afterword
It has been seven months since the publication of the article on the “fully on-chain game logic engine.” The original aim was to encourage individuals to break free from the confines of the “physics engine” framework and to pay more attention to “rules,” “contracts,” and other non-physical elements that underpin interactions. Over the past three to four months, 2-3 additional teams have ventured into this direction. While it may not necessarily be attributed to my efforts, it’s gratifying as a researcher. I hope this article has a similar impact.
In my research nowadays, collaboration and the alignment of ideas with others and with practical endeavors have become essential. Sometimes, it’s necessary to hold back on sharing ideas, allowing them to mature and evolve through discussions and interactions with the right people. I wholeheartedly welcome the opportunity to engage in discussions and brainstorming throughout this extended process. Feel free to initiate a conversation anytime; I’m here to chat and explore ideas together. ;)
This article was completed in 7 hours, and it was written hastily. The points made are based on my personal opinions and should not be taken as investment advice.