In this paper we introduce our new prediction market protocol, SynStation, which addresses critical challenges in today's prediction markets, presenting a robust, innovative alternative. We offer detailed insights into community-owned market deployment, autopilot staking, a CLMM-based (concentrated liquidity market maker) multi-pool system, and incentives crafted to align participants' interests effectively. Through these structural elements, our protocol not only resolves existing issues but also establishes a self-reinforcing positive feedback loop—a powerful "flywheel" effect that drives sustainable growth and engagement.
With the success of Polymarket, prediction markets are gaining momentum, and new platforms are emerging rapidly. However, most of these projects are merely forks of Polymarket or Gnosis, inheriting significant limitations.
Foremost among these challenges is lack of community engagement in market deployments, which limits authentic demand and the diversity of available markets. In nearly all existing protocols, teams exclusively decide which markets to open. Given the vast array of potential topics—from sports betting to entertainment, gaming, and elections—it's impractical for any team to consistently predict which topics will attract the most engagement. Conversely, leaving market creation entirely to participants introduces risks, such as inappropriate topics like assassination or markets with deliberately vague rules set by malicious actors. Effective market selection remains a critical challenge, yet no project has provided a robust solution.
The second challenge lies in the exchange's matching mechanisms. Currently, prediction markets rely on either Gnosis's CPMM model or a central limit order book like Polymarket, both of which have significant drawbacks.
The CPMM model, for instance, poses profitability issues for liquidity providers (LPs), who must offer liquidity for all possible outcomes and risk losing their entire stake if they fail to withdraw liquidity before market closure. Incentives like fee rebates fall short of addressing this issue—what’s needed is a fundamental overhaul of the market structure. From a user experience perspective, in multi-outcome markets, users—whether trading or providing liquidity—must buy or sell each outcome token separately, which is cumbersome.
The order book model, while not inherently flawed, faces issues due to niche markets, low liquidity, and an over-reliance on centralized players for all but the most popular markets. This structure challenges the user experience. Moreover it fails to adequately reward LPs due to the tactic called “ghost liquidity,” sometimes referred as “spoofing” and “flickering orders,” where participants game the system for rewards without adding real liquidity. Additionally, as the order book model is optimized only for binary markets, multi-outcome markets require setting up separate yes/no markets for each outcome, resulting in fragmented liquidity, inefficient pricing, and unnecessary profit leakage to arbitrageurs.
In the following sections, we will explore our protocol’s structural solutions and demonstrate how it effectively overcomes these challenges.
In this section, we introduce the first cornerstone of our next-generation prediction market: the community-owned market deployment model.
As noted earlier, deciding who controls the market deployment is a critical challenge. Relying solely on governance or the public brings either inefficiencies or significant risks. The ideal solution is what leverages the strengths of both. The public can rapidly identify trending topics that promise engagement, volume, and liquidity, while governance ensures alignment with the protocol’s mission and long-term sustainability. This hybrid, community-owned model combines both approaches, avoiding exploitation and streamlining topic selection, ultimately enhancing user engagement and retention.
Here, we examine the operational mechanics of a community-owned system. This approach allows the market to propose topics freely, with governance maintaining oversight. Effective governance should be able to prevent spam while incentivizing high-value topics that promise liquidity and trading volume.
Hyperliquid addressed a similar challenge by introducing dutch auctions for deploying new spot markets—a successful approach we adapt here. We conduct a dutch auction for the right to propose new topics using either a Gradual Dutch Auction (GDA) or a Variable Rate Gradual Dutch Auction (VRGDA), both of which support bulk purchasing and target rate setting. To manage demand fluctuations, multiple auctions can run concurrently.
This model ensures that each topic is proposed by a party who values it highly, with proceeds allocated to provide initial market liquidity. The GDA (or VRGDA) framework captures value effectively, allowing parameter flexibility. The target price for each auction round starts at twice the final price of the previous round, with a decay over a 24-hour half-life.
While auction winners are unlikely to propose vague or controversial topics, governance review remains essential. The auction winner must submit a proposal for the prediction market’s topic to the governance forum within 24 hours. A 3 to 7 days grace period follows, allowing proposers to refine submissions with feedback from DAO members. This ensures that the topic, market type (binary, 1-out-of-n, or spectral), outcomes, and evaluation criteria are well-defined. A final vote, lasting another 3 to 7 days, determines approval.