Written by: Nick Ruzicka

Translated by: Yangz, Techub News

The prediction market is entering a golden moment. Polymarket's coverage of the presidential election has made headlines, and Kalshi's regulatory victories have opened new avenues. Suddenly, everyone wants to talk about this "world truth machine." But beneath the hype lies a more intriguing question: if prediction markets are so good at forecasting the future, why are they not ubiquitous?

The reason is quite simple; at its core, it is an infrastructure issue, and in the U.S., it also involves regulatory barriers. While these regulatory barriers are gradually loosening (for example, Kalshi has received approval from the U.S. Commodity Futures Trading Commission, and Polymarket has opted for an offshore operating model), the flaws in infrastructure remain widespread. Even in regions where prediction markets are legally allowed to operate, fundamental developmental bottlenecks are still hard to overcome.

Platforms that dominate in 2024 are throwing money at these problems. According to an analysis by researcher Neel Daftary for Delphi Digital, Polymarket has burned through about $10 million on market maker incentives, at one point paying over $50,000 a day to maintain liquidity in its order book. Today, these incentives have plummeted to just $0.025 per $100 traded. Similarly, Kalshi has spent over $9 million. These are not sustainable solutions; they are merely band-aids on structural wounds.

Interestingly, the challenges hindering the development of prediction markets are not mysterious. They are clearly defined, interrelated, and— for the right builders—completely solvable. After communicating with builders in the field and analyzing the overall situation, five recurring questions emerged. We might consider them as a framework, a common language to understand why prediction markets, despite their theoretical potential, remain in beta mode.

These questions are not just problems; they are a roadmap.

### Challenge 1: The Liquidity Paradox

The most fundamental issue is liquidity. Or rather, it is the chicken-and-egg problem that causes most prediction markets to become "ghost towns." The mechanism behind this problem is simple: first, when a new market launches, it lacks liquidity, leading to a poor trading experience for traders (high slippage and price impact make trading unprofitable), causing them to leave; next, low trading volume scares away professional market makers who need stable fees to hedge risks; without market makers, liquidity continues to thin out, creating a vicious cycle.

This can be confirmed by data: on Polymarket and Kalshi, most markets have trading volumes of less than $10,000. Even larger markets lack the depth necessary for institutional participation—any significant position can cause double-digit price fluctuations. The root cause lies in structural issues. In ordinary crypto liquidity pools (like ETH/USDC), you deposit two assets and earn fees through traders' exchange actions— even if the price is against you, both assets still retain value. But prediction markets are different: you hold contracts that could go to zero. There is no rebalancing, no residual value, only binary outcomes.

Worse still, you may be "precision hunted." As the market approaches settlement and the outcome becomes clearer, more informed traders know more than you do. They will buy winning contracts from you at favorable prices while you are still pricing based on outdated probabilities. This "toxic order flow" continuously bleeds market makers. To address this, Polymarket shifted from AMM to a central limit order book in 2024: the order book allows market makers to cancel orders immediately upon realizing they are being exploited. But this does not solve the fundamental problem; it merely provides market makers with a defensive tool to slow their losses.

Platforms can circumvent this issue by paying market makers directly, but the subsidy model cannot scale. For flagship markets—presidential elections, major sporting events, headline crypto events—this model is effective. Polymarket's election markets have deep liquidity, and Kalshi's NFL market can compete with traditional sports betting. The real challenge lies in all other markets: those that could make prediction markets useful but whose trading volumes do not justify millions of dollars in subsidies.

The current economic model is unsustainable. Market makers do not profit from spreads—rather, they are compensated by the platform. Even protected liquidity providers with limited losses (4-5% per market) need ecological subsidies to break even. The question is: how can providing liquidity become profitable without burning cash?

Kalshi's strategy points to a path: in April 2024, they brought in Wall Street's mainstream market maker Susquehanna International Group as their first institutional provider. What was the result? Liquidity grew 30-fold, with contract depth reaching 100,000, and spreads below 3 cents. But this requires something that retail market makers cannot provide, such as dedicated trading desks, customized infrastructure, and institutional-level capital commitments. The breakthrough is not better rebates but getting the first serious institutional participant to view prediction markets as a legitimate asset class. Once someone takes the lead, others will follow: risk is reduced, benchmark pricing emerges, and trading volume follows.

But there is another key point: institutional market makers require specific conditions. For Kalshi, this means needing CFTC approval and a clear regulatory framework. For native crypto and decentralized platforms—those without regulatory moats or large-scale long-tail builders—this path is not feasible. These platforms face a different challenge: how to cold-start liquidity when you cannot provide regulatory legitimacy or trading volume guarantees? For all players outside of Kalshi or Polymarket, the infrastructure issue remains unresolved.

Builders' Attempts to Break the Deadlock

Quality-based market maker rebate mechanisms are beginning to reward those who genuinely enhance the trading experience—such as order book dwell time, quote size, and spread tightness. This approach is pragmatic but does not address the fundamental issue: someone has to pay for these rebates.

Protocol tokens offer another idea—subsidizing liquidity providers through token releases, replacing the model of consuming venture capital funds. This approach has successfully helped Uniswap and Compound achieve cold starts. However, whether prediction market tokens can capture enough value to sustain long-term releases remains uncertain.

Cross-market reward tier mechanisms incentivize market makers to provide liquidity across multiple markets, allowing them to spread risk while enhancing participation sustainability.

Instant liquidity models inject funds only when needed: bots monitor large trades in the memory pool, instantly injecting concentrated liquidity, capturing fees, and then withdrawing immediately. This model is highly capital efficient but requires complex infrastructure and does not solve the fundamental problem—someone has to bear the treasury risk. Although instant liquidity strategies have generated over $750 billion in trading volume on Uniswap V3, this arena has been dominated by well-capitalized players, and yields are becoming increasingly thin.

Continuous outcome markets directly challenge the binary structure itself. Traders no longer choose simple "yes/no" options but express opinions within continuous ranges. This can aggregate liquidity that would otherwise be scattered across related markets (for example, "Will BTC rise to $60,000, $65,000, or $70,000?"). Projects like functionSPACE are building this type of infrastructure, but it has yet to undergo large-scale practical testing.

The boldest experiments completely abandon the order book model. Melee Markets introduces bonding curves into prediction markets—each outcome has an independent price curve, early participants receive better prices, and believers are rewarded. This system does not require professional market makers at all. XO Market requires market creators to use LS-LMSR automated market makers to inject initial liquidity, with market depth automatically improving as funds enter. Creators can earn fees, which ties interests to market quality. Both solutions address the cold start problem without professional market makers. Melee sacrifices exit flexibility (positions must be locked until outcomes are revealed), while XO Market allows continuous trading but requires creators to invest upfront capital.

### Challenge 2: Market Discovery and User Experience

Even if liquidity issues are resolved, there is a more pervasive problem: most people cannot find the markets they care about, and even if they do, the user experience is quite unfriendly. But this is not just a "user experience issue"; it is a structural problem. The market discovery challenge directly exacerbates the liquidity dilemma.

Polymarket always has thousands of markets running, but trading volume is concentrated in a very few areas: election markets, major sporting events, and hot cryptocurrency topics. All other markets go unnoticed. Even if a niche market has decent depth, if users cannot naturally discover it, trading volume will remain low, and market makers will eventually withdraw. This creates a vicious cycle: without a discovery mechanism, there is no trading volume, and thus no liquidity can be maintained.

During the 2024 election cycle, Polymarket's top markets accounted for the vast majority of trading activity. After the elections, the platform's monthly trading volume remained between $650 million and $800 million, but these trades were scattered across sports, cryptocurrencies, and trending topic markets. The thousands of other markets—hyper-local issues, niche communities, quirky events—have almost no trading records.

Moreover, user experience barriers only add to the problem. The interfaces of Polymarket and Kalshi are designed for users who already understand prediction markets. Ordinary users face a steep learning curve: unfamiliar terminology, the conversion of odds to probabilities, and the meaning of "buying YES shares." These frictions may be manageable for crypto-native users, but for everyone else, they are enough to deter them.

Better algorithms can indeed help, but the core issue lies in distribution: how to match thousands of markets to the right users at the right time without causing choice paralysis.

Builders' Attempts to Break the Deadlock

The most promising strategy is to "go where the users are," rather than asking them to learn a new platform.

Flipr allows X users to directly @ a bot to trade prediction markets on Polymarket or Kalshi within their information streams—seeing a market discussed in a tweet, they can @Flipr to place an order without leaving their current application. This effectively embeds prediction markets into the conversational layer of the internet, turning social information streams into trading interfaces. Additionally, Flipr offers up to 10x leverage and is developing features like copy trading and AI analysis—essentially aiming to become a complete trading terminal that just happens to reside within Twitter.

The deeper insight here is that for startups, distribution is more important than infrastructure. Instead of spending millions to guide liquidity like Polymarket, it is better to aggregate existing liquidity and compete on distribution channels.

Platforms like TradeFox, Stand, and Verso Trading are building unified interfaces that aggregate odds from multiple platforms, route orders to the best trading venues, and integrate real-time news feeds. If you are a professional trader, why operate on multiple platforms when you can achieve better execution prices through a single interface?

Moreover, the boldest experimental solutions view discovery mechanisms as social problems rather than algorithmic ones. Fireplace, associated with Polymarket, emphasizes "investing with friends," recreating the atmosphere of collective betting rather than solitary guessing. AllianceDAO's Poll.fun goes even further: it establishes P2P markets among small groups of friends, allowing them to create markets on any topic, directly bet against each other, and settle through creator or group voting.

This hyper-localized, hyper-social model completely bypasses the long-tail problem by focusing on community rather than scale. These are not just user experience optimizations; they are distribution strategies. The platforms that ultimately succeed may not have the best liquidity or the most markets, but they will provide the best answers to the question of "how to present prediction markets to target users at the right time."

### Challenge 3: The Limited Expression Dilemma

Data shows that 85% of traders on Polymarket are in a losing position, enough to raise alarms for anyone optimistic about prediction markets. Part of this data is unavoidable, as predicting is inherently difficult. However, another part is a structural issue: platforms force traders into suboptimal positions because they cannot effectively express their views.

Do you have a more nuanced prediction? Unfortunately, you can only make binary bets: buy "yes," buy "no," and then choose the position size. There is no leverage to amplify beliefs, nor can multiple views be combined into a single position, and there are no conditional outcomes.

When traders cannot effectively express their beliefs, they either lock up too much capital or have insufficient position sizes. In either case, the flow of funds that the platform can capture will decrease.

This challenge can be broken down into two different needs: one type of trader wants to amplify a single bet through leverage, while another wants to combine multiple views for betting.

Leverage: A New Approach to Continuous Settlement

Traditional leverage does not work in binary prediction markets due to the brutal path dependency problem: even if your directional judgment is correct, price fluctuations may liquidate you before settlement. For example, a leveraged position on "Trump winning" could be wiped out during a poor poll, even though Trump ultimately wins in November.

Currently, better solutions have emerged, namely continuous settlement perpetual contracts based on real-time data streams. Seda is building true perpetual contract functionality based on data from Polymarket and Kalshi, allowing positions to settle continuously without waiting for the final results of discrete events. In September of this year, Seda launched perpetual contracts for the real-time odds of the Canelo vs. Crawford boxing match on its testnet (initial leverage of 1x), proving the feasibility of this model in the sports betting domain.

Additionally, short-term binary options are another direction that is gradually gaining attention. The cryptocurrency price volatility binary options launched by Limitless in September of this year surpassed $10 million in trading volume—these markets provide implicit leverage through their payout structure while shielding traders from liquidation risk during the contract's duration. Unlike perpetual contracts, they settle at fixed times, but the quick settlement cycles (hours or days, rather than weeks) provide the instant feedback loop that retail traders crave.

Currently, the relevant infrastructure is rapidly maturing. In September, Polymarket partnered with Chainlink to launch cryptocurrency price markets with 15-minute cycles. Perp.city and Narrative are experimenting with continuous information flow perpetual contracts based on polling averages and social sentiment—these are true perpetual contracts that will never settle as binary outcomes. Breakthrough progress also includes Hyperliquid's HIP-4 "event perpetual contracts"—you trade changing probabilities, not just final outcomes. For instance, if Trump's winning probability jumps from 50% to 65% after a debate, you can realize profits without waiting until election day. This addresses the biggest pain point of leverage in prediction markets: even if your final judgment is correct, you may be liquidated due to fluctuations along the way. Limitless and Seda are gaining market attention with similar models, indicating that the market needs continuous trading rather than binary betting.

Combination Betting: An Unresolved Challenge

Combination betting concerns how to express complex multiple views, such as "Trump wins and Bitcoin rises to $100,000 and the Federal Reserve cuts rates twice." Combination betting is relatively simple in traditional betting because traditional bookmakers, as centralized entities, manage diversified risks. Conflicting positions offset each other, so they only need to provide collateral for the maximum net loss, rather than for each individual payout.

Prediction markets cannot do this because they act as custodians, requiring full collateral for each position at the time of the trade. Even moderate combination betting would require market makers to lock up several orders of magnitude more capital than traditional bookmakers would need for equivalent risks.

The theoretical solution is a net margin system, which only requires collateral for the maximum net loss. However, this requires a complex risk engine, real-time correlation modeling across unrelated events, and likely a centralized counterparty. Researcher Neel Daftary suggests starting with professional market makers underwriting limited combination bets and then gradually expanding. Kalshi is adopting this approach—initially offering combination bets only for the same event, allowing the platform to model correlations and manage risks more easily within a single event context. This cautious approach acknowledges that without a centralized mechanism, a true combination market—one that offers the experience of "choosing any betting options"—is unlikely to be realized.

Most builders are betting on creative constraints: limited leverage for short-term markets, pre-approved combination bets, or simplified "enhanced odds" that the platform can hedge. The expression dilemma may be partially resolved (such as with continuous settlement perpetual contracts), but other aspects (like arbitrary combination markets) remain out of reach for decentralized platforms.

### Challenge 4: Permissionless Market Creation

Solving the expression problem is one thing, but a deeper structural issue is: who is qualified to create markets?

Everyone agrees that prediction markets need diversity—hyper-local issues, niche communities, and bizarre one-off events that traditional platforms would never touch. However, the permissionless market creation mechanism has been a disaster to date.

The core issue is that the half-life of market heat is extremely short. The most explosive trading opportunities often arise from breaking news and cultural hotspots. Markets like "Will the Oscars be rescinded due to Will Smith slapping Chris Rock?" would generate huge trading volumes if launched immediately after the event occurs. But by the time centralized platforms review and list the market, the heat has long since faded.

However, a completely permissionless creation mechanism faces three major challenges: semantic fragmentation (the same question appearing in ten versions, splitting liquidity into useless pools), liquidity cold starts (zero initial liquidity exacerbating the chicken-and-egg problem), and quality control (platforms flooded with low-quality markets, and worse, the emergence of assassination markets and other legal nightmares).

Polymarket and Kalshi have chosen a platform curation model, where teams review all markets to ensure quality and clear settlement standards. This builds trust but sacrifices speed—the platform itself becomes a bottleneck.

Builders' Attempts to Break the Deadlock

Melee employs a pump.fun-style strategy to tackle the cold start problem: market creators receive 100 shares, while early buyers receive diminishing allocations (3 shares, 2 shares, 1 share…). If the market gains attention, early entrants will capture excess returns—potentially exceeding a thousand times. This creates a meta-market: traders bet on which markets will grow by establishing early positions. The core bet is that only the highest quality markets—those from top creators or truly meeting demand—can attract sufficient trading volume, achieving a natural selection where good money drives out bad.

XO Market requires creators to use LS-LMSR automated market makers to inject initial liquidity. Creators earn fees, deeply tying their interests to market quality. Opinion-based prediction platforms like Fact Machine and Opinions.fun allow influencers to monetize their cultural capital by creating viral markets on subjective topics.

The ideal theoretical model is a hybrid community-driven mechanism: users must stake reputation and inject liquidity when proposing markets, followed by review from community curators. This retains the speed of permissionless creation while incorporating the quality control of a curation model. However, no mainstream platform has successfully implemented this model to date. The fundamental contradiction remains: permissionless creation fosters diversity, while manual curation ensures quality. Whoever can solve this balancing dilemma will unlock the much-needed potential of hyper-local and niche markets in the ecosystem.

### Challenge 5: Oracles and Result Determination

Even if all the other issues—liquidity, discovery mechanisms, expression methods, market creation—are resolved, there remains the most fundamental question: who determines the final outcome?

Centralized platforms have teams making determinations, which is efficient but carries the risk of a single point of failure. Decentralized platforms need oracle systems that can handle any issue without continuous human intervention, but result determination remains the hardest problem to crack. As researcher Neel Daftary explained for Delphi Digital, the emerging solution is a multi-layered architecture that can route issues to the appropriate adjudication mechanisms:

• Objective results use automated data sources. Polymarket integrated Chainlink in September to achieve instant settlement for cryptocurrency price markets, quickly and definitively.

• Complex issues are handled by AI agents. Chainlink Research tested AI oracles on 1,660 Polymarket markets, achieving an accuracy rate of 89% (99.7% for sports events). Supra's Threshold AI oracle employs a multi-agent committee to verify facts and detect manipulation, delivering results with cryptographic signatures.

• Ambiguous situations use Optimistic oracles (like UMA)—once a result is proposed, challengers must stake funds to initiate a challenge. While relying on game theory mechanisms, this is effective in handling clear issues.

• Highly controversial events introduce reputation-based juries—voting rights are linked to on-chain reputation records, not just the amount of funds.

The infrastructure is rapidly maturing, but this remains the trickiest challenge. Errors in result determination will destroy trust, but if resolved properly, it can support the scalable operation of millions of markets.

Why This Matters

These five major challenges—liquidity, discovery mechanisms, expression methods, market creation, and result determination—are interconnected. Solving liquidity makes markets more attractive, thereby improving discovery mechanisms; better discovery mechanisms attract more users, making permissionless creation possible; more markets mean a greater demand for robust oracles. This is a system, but there are currently multiple bottlenecks.

The opportunity lies in the fact that existing giants are locked into their own models. Polymarket and Kalshi built successful businesses based on specific assumptions about how prediction markets operate, optimizing only within established constraints. The new generation of builders can completely ignore these constraints.

Melee can experiment with bonding curves because it does not need to become a second Polymarket; Flipr can embed leverage into information flows because it does not require U.S. regulatory approval; Seda can launch perpetual contracts based on continuous data streams because it is not constrained by binary settlement models.

The real opportunity lies here—not in replicating existing models, but in directly tackling the underlying challenges. These five major challenges are the entry chips. Platforms that can solve liquidity, discovery mechanisms, and adjudication infrastructure will gain not only market share but also unlock the full potential of prediction markets as a coordination mechanism.

2024 will prove that prediction markets can operate at scale, and 2026 will prove that they can be ubiquitous.

Note: The data in this article primarily comes from the prediction market report published by Delphi Digital in October 2025 and the analysis by Neel Daftary.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。