Liquidity is constrained, and there is insufficient fluctuation; the apples of the prediction market are not so easy to pick.

Author: David

Compiled by: Shenchao TechFlow

This summer, I have been quietly trading while thinking about and building tools related to prediction markets. Since I joined Polymarket in mid-2024 to participate in election cycle trading, I have been intermittently involved in this field for over a year.

In June of this year, the conflict between Israel and Iran reignited my passion for prediction markets. At that time, I was not only trading real-time events on Polymarket for entertainment but also using it as an important source of information to guide my actual portfolio trading. In the following months, I conducted in-depth research on prediction markets, from their origins and multiple iterations to the conceptualization of future possibilities, as if I had entered an endless maze of knowledge.

Understanding a niche market with huge potential that few people discuss or take seriously excites me immensely.

But then, John Wang appeared. At the beginning of August this year, I noticed that John started frequently mentioning his in-depth research on prediction markets on Twitter, so I sent him a private message suggesting we chat. Although I can't disclose the details of the call, shortly after that, he became completely immersed in it and, through a series of intensive tweets, almost single-handedly brought prediction markets into the public eye.

Nevertheless, while I am also excited about the initial development stage of prediction markets, it is still in its infancy. Although there is much discussion about its positive implications, many challenges and limitations in its current form still need to be addressed if prediction markets are really to become a new trading mainstream.

Liquidity Restriction

The first major flaw of prediction markets is the liquidity dilemma. For most professional traders, liquidity in the market is already insufficient, let alone allowing funds to execute large-scale trades. Furthermore, due to the difficulties in market making for binary prediction markets, there are very few willing market makers; at the same time, the trading volume of these markets is inherently low, leading to limited profit opportunities for market makers, which diminishes their motivation to participate.

There are several reasons why binary markets find it difficult to establish market pricing. First, there is a high inventory risk and it is difficult to hedge. Since these markets are event-driven, their characteristics lead to little or almost no possibility of mean reversion after major news headlines emerge. For example, if a market's trading price reflects an 80% probability of a certain event occurring (“yes”), but then a piece of news significantly lowers that probability to 30%. If market makers are positioned incorrectly in this situation, they may be forced to hold large losing positions that are often difficult to exit. This risk can be mitigated through hedging, but it is not always easy or efficient to find capital solutions to achieve that hedge.

Why are market makers afraid of being “trapped”?

Another issue is “toxic flow” and the lack of diversity in demand. Market makers typically profit from the bid-ask spread. For example, buying shares of X at $1 and then selling them at $1.01, repeatedly, while having no directional view on the underlying asset. Whether market makers can be profitable largely depends on the higher the proportion of “low-information demand” in the market, and the lower the proportion of “high-information demand.”

Taking the stock market as an example, “low information demand” usually refers to investors trading due to hedging other positions or rebalancing their portfolios. They are not buying because they believe the stock will rise, but because it is needed for portfolio construction. This type of demand is usually beneficial for market makers, as buyers are not sensitive to price.

On the other hand, “high information demand” or “toxic flow” is exactly the opposite. These buyers typically possess undisclosed information or some advantage, believing that market pricing is incorrect and attempting to profit through trading.

The proportion of these two types of buyers needs to remain healthy in order for market makers to provide sufficient liquidity in a profitable manner. However, the current demand in the prediction market lacks diversity, with almost no other types of participants besides speculators, and it is susceptible to the influence of “toxic flow” from insiders. To improve liquidity, this demand structure must change.

Retail Investor Restrictions

From the perspective of retail investors, there are many limitations in predicting the market, and I will briefly describe these issues.

First of all, the market lacks sufficiently attractive opportunities and potential returns. Most markets on platforms like Polymarket and Kalshi typically have low volatility, and the potential returns are not enough to capture the interest of retail investors. Even if a seemingly certain result has a trading price of 70%, if its expiration is two months away, the appeal is still not enough for modern retail investors who seek “dopamine stimulation.” Additionally, due to the challenges faced by the aforementioned market makers, these markets are also unable to offer leverage trading options to enhance potential returns.

Secondly, the yield ceiling characteristic of the binary market reduces the motivation for “early layout,” which is precisely one of the important reasons why stocks and cryptocurrencies are attractive. Currently, there are some new prototypes being tested that introduce reflexivity by removing binary outcomes, but whether they can succeed remains to be seen.

Thirdly, event-based markets reduce reflexivity. This is both an advantage and a disadvantage, as it means prediction markets are less susceptible to manipulation or issues like “cabal” control in cryptocurrencies. However, this also limits potential gains and does not provide the 100-fold returns that retail investors crave. I have some thoughts on this, but I won't discuss them today.

Poor discovery mechanism and user experience

Anyone actively using prediction markets will encounter numerous frustrations in the current user interface (UI) iterations. There are simply too many issues, especially for deep users, and these problems accumulate to become quite bothersome. The worst, in my opinion, is the market discovery mechanism.

Polymarket and Kalshi currently have tens of thousands of markets, and the number is continuously growing, but the vast majority of markets you may have never heard of, and there is no easy way to find them.

The dawn of hope

The good news is that many of these challenges are not unique to prediction markets.

Early decentralized finance (DeFi), perpetual contract exchanges, and short-term options contracts have also faced similar issues. If anything, this indicates that prediction markets contain enormous opportunities. Currently, prediction markets are still very niche.

Taking Polymarket as an example, its 250,000 active users had a trading volume of $1 billion last month. In contrast, the top 100 traders on HyperLiquid almost each reached this trading volume.

We can be excited about new things, but we must also maintain a pragmatic attitude and face their current actual state in order to push them to new heights.