Intro

Prediction markets allow participants to bet on the outcomes of various events. In 2014, Truthcoin (later renamed to Hivemind) developed a first crypto-oriented prediction market on the Bitcoin network, where BTC holders could use their coins to vote on outcomes of events, with rewards or penalties based on the majority decision. Despite many new features and solutions introduced in this space over time, prediction markets still struggle with widespread adoption. Let’s dig deeper.

Players in the Prediction Markets System

Each market operates under its own system, with numerous connections linking various players (stakeholders) to one another. Every agent seeks to maximize their gains and typically aims to behave efficiently. To determine all the relations and try to find a leak there is a need to name mentioned players. Overall, all of them can be split into two separate fronts: demand and supply side.

---

e.g. The movie industry has got 4 main players who contributing each other: movie-goers and exhibitors (theaters / cinema chains) from demand side and actors, film studios from supply side.

---

The predictions markets might be considered with a similar point of view. Overall, there are 7 main stakeholders ( players ) both from demand and supply side.

In prediction markets, the demand side consists of several key participants. Speculators play a major role, participating by betting on the outcomes of future events with the goal of profiting from accurate predictions. Their behavior is often driven by insights, analysis, or intuition, as they seek financial gain by being on the correct side of a bet. Alongside them are hedgers, who participate not to speculate, but to mitigate risk. These participants use the market to protect themselves from unfavorable outcomes, for instance, a company hedging against a potential political decision that could negatively impact its operations. Their primary objective is not profit but risk reduction, and they tend to approach the market more conservatively than speculators. Another important group are forecasters, experts who bring deep knowledge of specific fields and leverage data and models to make informed predictions. Their involvement adds credibility and often drives more accurate pricing within the market. Meanwhile, public sentiment participants, including everyday users, join markets either to express opinions, follow trends, or simply for entertainment. Their participation may be casual and less data-driven but adds a layer of diversity to market perspectives.

On the supply side, platform operators or market creators are the backbone of the system. These are the entities, either centralized or decentralized, that establish and maintain the infrastructure for the prediction markets to operate. Their goal is to ensure a seamless user experience, manage liquidity, and facilitate efficient trades, all while earning fees from the platform’s operations. Liquidity providers, or market makers, play a critical role in maintaining market fluidity, making it easier for participants to enter and exit positions. By continuously buying and selling contracts, they minimize price slippage and help maintain market stability, profiting from transaction spreads or rewards. Another vital agent is the oracle, which acts as the bridge between the real world and the market. Oracles provide accurate, verified data on event outcomes (like election results or sports scores), ensuring the correct resolution of contracts. Without reliable oracles, prediction markets risk failing due to inaccurate or delayed event resolutions.

The success of prediction markets depends on the interplay between these players. The demand side creates liquidity and engagement by participating in the market and placing bets. Meanwhile, the supply side ensures that the infrastructure is in place to facilitate smooth market operations, accurate outcomes, and sufficient liquidity. By naming and understanding these players, we can identify potential inefficiencies or "leaks" in the system.

---

In addition to already identified players, according to the Nick Whitaker & J. Zachary Mazlish, some other ones may be pointed out:

• “Savers: who enter markets to build wealth.

• Gamblers: who enter markets for thrills.

• Sharps: who enter markets to profit from superior analysis.”

---

The behavioral roles described above might be compared with those mentioned earlier, e.g.

1. savers = hedgers;

2. gamblers = speculators + public sentiment participants ;

3. sharps = forecasters

Prediction Markets’ Liquidity Models

In the realm of prediction markets, Automated Market Makers (AMMs) and Order Book models are two core mechanisms that facilitate trading, each with distinct advantages and challenges.

AMM-based prediction markets use liquidity pools, where asset ratios are adjusted algorithmically to set prices. Instead of needing buyers and sellers to match directly, users trade against the pool, ensuring continuous liquidity.

A leading example of AMM use in prediction markets is Azuro, which employs a Concentrated Product Market Maker (CPMM) model in a singleton liquidity pool. This shared pool enables all prediction markets on the platform to scale according to demand, while Azuro’s Liquidity Tree system optimizes fund management and distribution across events. Azuro also prevents impermanent loss (common in AMM pools) for liquidity providers, making it appealing for passive investors.

Benefits of AMMs in prediction markets include continuous liquidity (users can trade anytime) and capital efficiency, as the single pool can dynamically allocate liquidity across markets. However, AMMs often struggle with slippage on large trades, as prices adjust based on pool composition, and they are better suited to simple binary outcomes rather than multi-outcome or conditional bets.

By contrast, Order book models rely on matching buy and sell orders to execute trades, with prices set by users through bids and asks. Central Limit Order Books (CLOBs) are the most popular form of order books in prediction markets. In a CLOB, all open orders are visible in a centralized list, and trades occur when matching prices are found. CLOBs provide transparent, user-defined pricing but typically require high participation to maintain liquidity and avoid high volatility in thinly traded markets.

Polymarket is a prominent example of a CLOB-based prediction market. It uses a hybrid off-chain/on-chain model where an operator matches orders off-chain before executing trades on-chain. This model offers efficient pricing for binary outcomes (such as YES/NO on specific events), allowing users to place limit orders for more precise pricing. The transparency of CLOBs enables price discovery and market depth, making them ideal for high-liquidity markets. Another example, Opinion labs uses an order book model for trading.

Benefits of order book models include transparency in pricing, allowing users to see and act on real-time bids and asks. They also handle complex or multi-outcome markets more effectively than AMMs. However, they require steady user activity to ensure liquidity and generally demand active order management from liquidity providers to stay competitive.

Prediction Markets’ leaks

Demand side

The demand side of prediction markets faces leaks rooted in the individual motivations and personal challenges of different participant types: Hedgers, Speculators, and Forecasters. These groups have distinct goals, yet all contribute to a demand-side ecosystem that struggles with interconnected structural issues. The absence of consistent demand from these groups undercuts liquidity, hinders price accuracy, and ultimately limits the utility of prediction markets as a robust financial instrument.

Hedgers are generally conservative investors seeking low-risk opportunities to mitigate potential losses from adverse events. In prediction markets, however, the zero-sum (or even negative-sum after fees) nature poses a fundamental issue. Unlike traditional securities or bonds, which offer long-term positive returns and support capital growth, prediction markets do not accumulate value over time. This discourages hedgers, who typically prefer investments that can appreciate and protect savings, and view prediction markets as unreliable for risk management. Additionally, the absence of an interest-earning component or collateralization benefit limits the attractiveness of prediction markets for hedgers who might otherwise use the platform to offset risks in traditional markets. Thus, prediction markets are viewed by hedgers as too speculative and ineffective for long-term financial goals.

Speculators are thrill-seekers and opportunists who seek high returns, often with a higher tolerance for risk. They view prediction markets primarily as an entertainment or gambling vehicle rather than as a financial instrument for wealth-building. However, prediction markets face a challenge in meeting their demands. Unlike sports betting, where the fast-paced nature and quick market resolution create continuous excitement, most prediction markets offer long-term events with delayed outcomes. Markets based on election outcomes, regulatory decisions, or scientific breakthroughs can take weeks, months, or even years to resolve, which dulls the excitement for speculators.

---

I’d place myself in the speculators' camp, and naturally, after profiting from Donald Trump’s presidential win (Happy MAGA), I was keen to identify the next lucrative opportunity. I turned my attention to the potential of Young Thug’s pardon within 100 days of Trump’s inauguration. I counted 71 days until the inauguration, added another 100 days for the market to settle, and decided to close the book on it.

---

Furthermore, prediction markets lack the leverage opportunities that many speculators seek to amplify small investments into significant profits. This limitation stems from insufficient liquidity; thin order books make it risky to offer leverage, as extreme price movements could quickly lead to insolvency for exchanges. In contrast, highly volatile crypto assets like memecoins provide speculators with the potential for exponential returns, capturing their interest with the possibility of outsized gains. Without comparable upside potential or safe mechanisms for leveraging positions, prediction markets cannot compete with other speculative vehicles that offer higher stakes and faster returns.

A comparison of potential returns between the $MAGA memecoin and the presidential election prediction market reveals distinct profit dynamics, each catering to different investor motivations. By calculating the return on investment (ROI) for each, we can highlight how these two types of markets appeal to speculators with different preferences for speed and profit potential.

For $MAGA, known for high volatility, ROI calculations suggest significant profit potential under ideal conditions. Buying at $1.70 and selling at $5.45 yields an approximate 220% ROI. This high gain highlights why memecoins appeal to speculators: rapid price swings offer substantial returns within short timeframes, though such outcomes rely on precise timing and are not guaranteed.

In contrast, the presidential election prediction market offers a steadier, more moderate return. Buying at $0.44 and selling at $0.95 yields a 116% ROI, reflecting the slower growth typical in prediction markets as event dates approach.

In summary, the $MAGA memecoin’s 220% gain outpaces the 116% gain of the prediction market, demonstrating why speculators are drawn to high-volatility assets like memecoins. While memecoins deliver the fast, high returns that speculators seek, prediction markets offer more stable but slower gains aligned with real-world events.

Forecasters are sophisticated, informed participants who invest significant time analyzing and understanding probabilities, hoping to profit from precise predictions. However, prediction markets fall short for this group as well. With minimal participation from hedgers and speculators, prediction markets lack the liquidity and demand necessary to make forecasting profitable. Prediction markets that are limited in size and scope lack the depth that would otherwise attract large, informed bets, resulting in insufficient profits to justify the time and expertise required for accurate forecasting. This creates a no-trade problem, where forecasters hesitate to enter a market because they are aware that other participants are also highly informed. Without less-informed participants or “noise” to take advantage of, forecasters face the uncomfortable reality of trading against other sharp players, where the risk of loss is elevated, and potential returns are diminished.

Moreover, the limited market size means that even if forecasters hold an informational advantage, their potential profit is capped by the small volume and narrow spreads typical in thinly traded prediction markets. In contrast, large financial markets offer enough volume that even a slight informational edge can yield significant profits. Without such size, forecasters are deterred by the realization that even highly accurate predictions will yield relatively small rewards.

Together, these challenges create a demand-side bottleneck in prediction markets. The absence of hedgers, who typically provide stable liquidity and low-risk capital, weakens market stability. This, in turn, deters speculators who seek fast-paced, leveraged, high-stakes betting opportunities that simply don’t exist in undercapitalized markets. The lack of hedgers and speculators then discourages forecasters, who recognize that without a broad pool of participants—including those with differing levels of expertise—the market remains low in liquidity and high in informational efficiency. The result is a market where few participants trade, and those who do are often high-knowledge forecasters. This leads to a “forecaster vs. forecaster” scenario, where sophisticated participants are forced to trade against each other, dampening their confidence in any informational edge they may hold.

Ultimately, these demand-side leaks form a self-reinforcing cycle of low participation and limited liquidity. Without hedgers to stabilize markets, speculators to drive volume, and forecasters to refine pricing, prediction markets struggle to attract the diversity and scale necessary to become functional and liquid financial instruments.

The main supply-side leak in prediction markets is the lack of liquidity, driven by two primary factors: volatility and information asymmetry.

Volatility in niche or novel markets, such as predicting outcomes with minimal historical precedent, discourages market makers from providing deep liquidity, as the risk of mispricing is high. This leads to wide spreads and reduced participation from large bettors. That’s going to be more prominently when it comes to the self-made markets proposed by platform users - feature mentioned in partially every relevant whitepaper.

For instance, let’s take a look at the prominent $336k-volume prediction market at Polymarket - ‘Sundar Pichai out as Google CEO in 2024?‘. It’s obvious that this prediction wouldn’t be based on historical data because it has never happened. To figure out how much inefficient liquidity is let’s model a potential bet counted by 500$.

To generalize the calculation of slippage for a large buy order in a thinly traded market, we can define a cumulative cost formula that iterates over each price level in the order book until the desired investment amount $T$ (target dollar amount) is met.

Let:

• $Pi$ represent the price at each level $i$ in cents.

• $Si$ represent the number of shares available at each level $i$.

• $Ci$ represent the cumulative cost after buying all shares up to level $i$.

• $T$ represent the target dollar amount for the total buy order (e.g., $500).

where:

• $i$ iterates over each level in the order book, and the calculation stops once $Ci$ $≥$ $T$.

Here’s a structured table using the given order book data and applying the cumulative cost formula up to a target of $500. I’ll calculate the cumulative cost $Ci$ at each level, and highlight the final slippage based on the initial ask price versus the highest price reached.

To fulfill the $500 buy order, you encounter 218.18% slippage from the initial ask price of 4.4¢, ending at 14.0¢. This highlights the impact of limited liquidity, as larger buy orders drive prices significantly higher, resulting in substantial slippage.

Information asymmetry further exacerbates the problem. Market makers are vulnerable to "toxic flow," where informed participants exploit their knowledge to make profitable trades against less-informed liquidity providers. This unpredictability forces market makers to price in the risk of adverse selection, further limiting liquidity and narrowing market activity.

Prediction Markets’ Potential Solutions

To address mentioned leaks, targeted solutions must balance incentives, market design, and risk management. The table below outlines actionable solutions for these issues, their expected outcomes, and potential threats to implementation. By addressing these concerns holistically, prediction markets can become more competitive and scalable.

The success of prediction markets relies on overcoming existing inefficiencies by implementing well-balanced solutions. These solutions align with initiatives already pursued by leading prediction market projects. For instance, Polyquest highlights plans to implement Additional Rewards for Early-Birds through bonus staking programs, incentivizing early participation and bootstrapping liquidity for new markets. Similarly, Azuro employs an aggregated liquidity pool model powered by its Liquidity Tree mechanism, dynamically allocating liquidity across events to reduce fragmentation and ensure scalability.

By adopting such innovations while addressing challenges like sustainability concerns and over-reliance on certain mechanisms, prediction markets can evolve into reliable and widely-used financial tools, attracting diverse participants and achieving long-term growth.

Post Presidential-Election-Rally Circumstances

The aftermath of a presidential election often brings significant shifts in prediction markets, financial behaviour, and speculative activities. The presidential rally acts as a catalyst to pull in seasonally inelastic demand.

In November 2020 ( According to the Richard Chen’s Dune analytics dashboard) , Polymarket saw a monthly trading volume of approximately $20M. At the time, the platform was still in its early stages, attracting a niche audience of political forecasters and speculators. While impressive for a nascent platform, this volume reflects the limited adoption of prediction markets at the time and the sector's relatively small footprint.

Fast forward to the November 2024 presidential election, and the story is dramatically different. Leading up to this event, Polymarket experienced explosive growth. By the peak in November 2024, monthly trading volume had surged to an astonishing $1.095B.

Over the four-year period, Polymarket's trading volume grew by an incredible 5,375% signaling the platform's rapid evolution.

This growth reflects a combination of factors, including improved platform functionality, a broader user base, heightened public awareness of prediction markets, and enhanced liquidity that enabled larger trades. The run-up to the 2024 election also benefited from the hockey stick effect, where user activity and market volumes surged in anticipation of the high-stakes event.

Beyond the significant traction achieved by Polymarket, it’s important to highlight the core vision behind prediction markets: building a sustainable system that aggregates collective opinions to generate truthful outcomes, with participation open to anyone. According to a recent tweet from Polymarket’s CEO, this concept has already proven its validity.

While the peak in November 2024 showcases the potential of prediction markets during major events, it also highlights their cyclical nature.

Open Interest (According to the Richard Chen’s Dune analytics dashboard), which reflects the total capital deployed in unresolved market positions, serves as a key indicator of market liquidity and engagement. It is calculated as:

Throughout early 2024, Open Interest grew steadily as capital flowed into election-related markets. By October 2024, it exceeded $400M, culminating in a peak of $450M on election night, reflecting the intense speculative activity around the event. However, this momentum was short-lived. Immediately after the election, Open Interest dropped sharply to under $50M as markets resolved, payouts were made, and traders exited their positions.

It’s probably related to two distinct substances: constant - market segment’s nature and dynamic - current segment’s leaks.

As for the first substance, prediction markets can be definitely determined as a segment with cyclical demand patterns. According to the economical theory, “Cyclical patterns are characterized by recurring fluctuations in the data that typically span more than a year. These cycles can be linked to factors like interest rates, political climates, consumer confidence, or broader market dynamics.” This post-election plunge highlights the cyclical nature of prediction markets, where activity peaks around high-profile events and drops sharply once uncertainty resolves.

In contrast, the identified leaks in prediction markets highlight their dynamic and evolving nature. To address these challenges and sustain liquidity and user engagement between major events, platforms must prioritize diversifying into short-term or recurring markets, incentivizing liquidity providers, and creating opportunities for continuous participation.

---

Shoutout and thank you to Joshua Yang, Benjamin Sturisky, Alexander Dorokhin, and Aleksandr Litasov for the feedback and conversations on this topic.

Written by Mike Rychko (@mikhryc0x)