Prediction markets have been explored in both traditional finance and the world of cryptocurrencies, with notable platforms including Augur in the crypto space and Betfair in traditional betting markets.
Augur ($REP), at its peak, reached a valuation of $1.35 billion, while Betfair currently holds a market valuation of around €38 billion on the Frankfurt Stock Exchange. Despite this, none of these platforms has fully succeeded in establishing prediction markets as serious financial tools on par with the stock market or other areas of traditional finance. Betfair has come closest, especially with political and financial betting, yet prediction markets overall still largely fall short of being taken as seriously as other financial markets.
Why haven’t prediction markets achieved the same level of legitimacy and adoption as stock or commodity markets, despite being fundamentally economic tools designed to forecast events? Here’s what I believe are some key reasons why prediction markets haven’t worked as well as their traditional financial counterparts:
To explore this further, I invested in a newer prediction market called Polkamarkets and provided feedback to the team regarding their fee structure and market-making approach. At the time, Polkamarkets charged a fee of around 5% per trade, which I believed was prohibitively high and in line with most Sports betting sites like Bet365 or casinos in Las Vegas.
I suggested lowering fees to 0.10% per trade to match Binance’s fees (the largest cryptocurrency exchange at the time). Someone from the team responded via Discord, explaining that they couldn’t reduce the fees because these fees were needed to incentivize liquidity providers. Polkamarkets, which used a liquidity pool system, relied on these providers to match bets against bettors, creating a sort of “automated bettor” with shifting odds similar to Uniswap.
However, I believed that lowering fees would have helped attract more volume, and I also suggested considering an order book model or some way to make an Automated Market Maker (AMM) model function with a much lower fee structure. My reasoning was simple: the high fees were a significant barrier to serious bettors with serious volume like a professional bettor or a hedge fund, really high fees in markets like these are really only paid by consistent losers with low volumes, and if those are all the users you attract, this industry never becomes a serious one, and remains one where people bet $20 for fun while they watch a sports game, and lose, most of the time.
Ultimately, Polkamarkets struggled to attract substantial betting volume and activity with its model. In contrast, Polymarkets — a competitor with a different approach — has seen extraordinary success. Polymarkets may even be on track to become the de facto platform for serious prediction betting, surpassing Betfair:
The success of Polymarkets illustrates a few key points:
- Fee Structure Matters: Prediction markets need low fees to be competitive and attractive to serious users. Excessive fees discourage active participation from experienced players with more volume. Low fees are critical to drawing in regular, serious users. And casual recreative users can also benefit by keeping more of their wagered money and even becoming profitable just by switching to a low-fee platform.
- Market-Making Models: An order book model may be better suited to prediction markets than an AMM model, which Polkamarkets relied on. AMMs, while useful for certain decentralized finance applications like Uniswap pools, can be challenging to implement effectively in prediction markets.
- The right markets: Polymarket has focused on political bets, and mainly on US presidential elections, which is a very smart move since these events attract a really large number of bettors from all over the world and a lot of volume, in fact, US presidential elections have consistently beaten their own records over time as THE events with the most money bet on them in history.
Traditional predictive betting, such as sports betting, has notoriously high fees that make it nearly impossible for most users to consistently profit. As a result, even skilled bettors often end up losing money.
These traditional sites often operate like casinos, offering pure-chance games that can’t be beaten and don’t require any predictive skill — in other words, pure gambling, as well as having other “markets” where predictive skills do matter.
This setup generally fails to attract serious participants who could study variables and make informed predictions about outcomes. With fees set high enough, even events that could potentially be profitable for knowledgeable bettors turn into chance games since it’s difficult for anyone to consistently overcome the house’s edge.
This can be summarized with the analogy that a company might complain that “taxes are too high” in certain countries or states, therefore choosing to be incorporated in different one to retain more of their production, or, in some cases making some business unviable in a territory, due to high taxation.
Recently, however, prediction market sites that are more like financial markets have emerged due to lower fees and the addition of stock-market-like mechanics. I’d recommend checking out Polymarket and Betfair, especially Polymarket, as it is fully cryptocurrency-based and has 0% fees. This approach offers significant cost savings and adds a level of transparency and trust that traditional betting sites often lack thanks to the reliability of smart contracts.
On Polymarket, instead of buying traditional odds, you buy shares in prediction markets, which you can trade at any time before the event concludes. A key benefit of Polymarket is that they currently have 0% fees on all trades and bets, making it as affordable — or even more affordable — than traditional financial markets. This is made possible through smart contract guarantees and decentralized, public arbitration via UMA, a protocol that prevents Polymarket from directly resolving bets. Instead, bets are resolved through an independent protocol that uses incentive mechanisms to prevent conflicts of interest. An analogy to this could be that Polymarket has become a tax-free country with good judges and, generally speaking, political stability.
Another piece of information that shows how Polymarket has succeeded is this comparison between each of the major betting sites, with Polymarket taking the global lead, even above Betfair, the previous market leader in the 2024 US presidential elections, arguably the most important for betting markets in recent times, and possibly the most important event for this matter in history. Polymarket at 84% of the volume, followed by Kalshi, and Betfair.
Prediction markets minus the money-part
Other prediction platforms, like Manifold Markets and Metaculus, also offer valuable insights without real money. Manifold allows users to bet and trade with virtual play-money, while Metaculus removes market dynamics altogether, focusing on tracking users’ prediction accuracy. Metaculus aggregates predictions into a single probability, which still holds predictive value due to the wisdom of crowds phenomenon. However, I believe market-based predictions are superior since they offer the opportunity for people with advantaged information or an edge to sway the market’s probability in a certain direction, while compensating them for their edge in intelligence, while also keeping that edge (without making their information or predictive models public).
The public good from prediction markets
While prediction markets are generally zero-sum or negative-sum games that don’t create goods or services, there is a positive byproduct of their existence: Predictions. Prediction markets can create data that is highly predictive and more accurate than other methodologies, this information can serve people to guide people and help them better prepare for an uncertain future.
