Centralization: What's The Problem?

Traditional prediction markets are not broken – they’re working exactly as designed. As the saying goes, the house always wins. But why is that? “The house” is often a centralized entity that creates the rules, and the odds are stacked in their favor. All value is typically captured by these centralized entities – whether casinos, bookmakers, investment banks, insurance companies, or exchanges. They, and their shareholders, are profit seekers, and operate in accordance with their own tried and tested incentive structures. With promises of quick profit while using every trick in the book to extract maximum value from their customers, they create a framework of incentives to get users in the door - and walk out with less money than when they walked in. In a centralized prediction market, participants take a position on an outcome and, eventually, the centralized entity compares these predictions with the final outcome to decide on the payout. As such, a single entity has a significant influence on the overall activity in the prediction market. Because of this control and opacity, the centralized entity may have vested interest to adjust parameters to maximize their profits.

Traditional prediction markets have already been disrupted by the emergence of online markets and bookmakers - enabling global access, greater convenience, and greater choice of odds. Yet the pricing of risk still remains centralized and inefficient: odds and payouts are pre-calculated based on opaque proprietary logic designed to hide significant fees. Meanwhile, bookmakers make every effort to hold on to players’ funds, exposing users to significant counterparty risk. Ultimately, online bookmakers remain centralized entities driven by a mission to extract as much profit from participants as possible. And while casinos prioritize experience, online bookmakers compete on margins because they’re out of innovative ways to attract users.

Participation in such markets is often unproductive and unprofitable for most: odds are pre-set against the player, fees are high, and revenue is never shared with its participants. In the worst case, the centralized entity may manipulate the final outcome too, resulting in a loss of user funds. Ultimately, participants have to trust the authenticity of the platform in order to ensure privacy, security, fairness, and safety of funds. Payouts are reliant on solvency of a single point of failure. Opaque risk management and risk distribution means sometimes when you need your payout the most is when you will be unable to receive it (insurance companies disputing claims, investment banks unable to uphold over-the-counter derivatives terms).

Beyond risk, participants in centralized prediction markets bear various costs for participation through high platform fees. The opaque, black box pricing mechanisms of centralized prediction markets introduce significant fees so that participants do not receive optimal market-derived pricing for their positions. The justification for these fees originate from the fact that these centralized prediction markets are for-profit exchanges recovering their regulatory and administrative costs. Alongside high trading fees, deposit and withdrawal fees, the prediction market takes a cut from user profits, too. This is an obstacle for reliable participants as the entry barriers and high fees discourage participation, leading to less reliable predictions. Meanwhile, the pricing of risk itself is inefficient because no single market participant (or entity) can accurately predict accurate odds. Market intelligence is the best metric for evaluating the value of an asset, but it is not sufficient to price that asset. Thus, centrally determined odds introduce further costs for participants.

The strength of any prediction market is directly proportional to the number of participants contributing to the crowdsourcing of information needed for an accurate forecast. However, centralized prediction markets are limited by stringent regulations and capital control. The privileges of prediction markets and their significance are only enjoyed by the developed countries that tend to have advanced financial markets, sufficient liquidity, institutional framework, and regulatory favorability. Market supervisors act as gatekeepers that restrict who can participate and which events participants speculate on: users cannot create their own prediction market for a specific event unless they run a regulatory and licensed platform.

A key challenge for any prediction market is confirming the accurate result that the market payout depends on. For example, in a sports event, the market must know the final score of the event, the winner, and any other relevant metrics that the market payouts may depend on. The current market suffers from a lack of a scalable and widely ranging "oracle"—a single source of information or APIs that can be trusted to deliver accurate results. Instead, there is a marketplace of independent data providers and APIs for specific outcomes: perhaps one for basketball games, one for weather, and yet another for congressional votes in a certain district of the United States. The problem with the current market for prediction markets is that it's a fragmented mess of unreliable data.