Before Polymarket became a popular application in the crypto world, Taiwan had already had a prediction platform that combined academic research with market mechanisms. Led by National Chengchi University, the Prediction Market Research Center launched the “Future Events Trading Exchange” in 2006—not only one of the most representative prediction markets in the Chinese-language world, but at one time also viewed as a “wisdom-of-the-crowd” tool closer to real results than polls.
“Predicting the future with prices” had already happened in Taiwan long before the blockchain era.
From academic experiments to a market-based platform
The origins of the Future Events Trading Exchange can be traced back to early experiments with prediction markets in Taiwan’s academic community in 2003. In 2006, Chengchi University established the Prediction Market Research Center and, together with the Institute of Information Science of Academia Sinica and private companies, formally launched this online platform.
Unlike traditional questionnaires, prediction markets assume that when participants need to “place bets” (even if only with virtual assets), they will evaluate information more seriously—thereby generating prices that better reflect true probabilities.
Talking in prices: the core mechanism of prediction markets
The operating logic of the Future Events Trading Exchange is, in essence, close to that of futures markets. The platform breaks “future events” down into tradable contracts—for example: whether a certain candidate will be elected, whether a certain economic indicator will hit a target, or the outcome of a specific sports event. Traders are not trading an asset; they are trading the “probability that an event will occur,” which is just like Polymarket today.
When the market broadly believes the probability of an event is rising, the contract price goes up; otherwise, it goes down. This makes price itself an immediate probability signal—an quantified expression of market consensus. The biggest difference from polls is that this is not just sampling opinions, but judgment backed by an “incentive mechanism.” Because once users make the wrong judgment, they pay a cost.
In its trading and settlement design, the Future Events Trading Exchange used a virtual currency system rather than real money. This creates a key difference from modern on-chain platforms. Participants traded based on publicly available information, market signals, or private information, and the final payout was determined by the outcome of real-world events. But because there is no real-money loss-and-gain pressure, it also brings a structural problem—some trades may be driven by emotion or belief rather than rational prediction.
This is also one of the major sources of later criticism from academia and the market: when the “costs aren’t real enough,” price signals may get distorted.
Contract design is extremely broad: from elections to entertainment, everything can be traded
The Future Events Trading Exchange once offered an extremely diverse range of contract types, covering fields such as politics, finance, Cross-Strait matters, international affairs, sports, and entertainment. For example, after Cross-Strait tourism opened in 2008, the platform designed a range contract for “the number of Mainlanders visiting Taiwan by year-end,” allowing participants to use prices to predict where the data points would land. Other popular underlying topics also included presidential elections, stock market trends, U.S. interest-rate decision-making, and even results from Star Avenue.
By 2010, the platform had accumulated nearly 2,000 sets of contracts, more than 13,000 individual contracts, and trading volume exceeding 200 million contracts, showing that at the time it had indeed formed a certain scale of information market.
Outperforming polls—and also getting it wrong and collapsing
The value of prediction markets comes from whether the “wisdom of the crowd” is superior to traditional methods. The Future Events Trading Exchange performed impressively in some election forecasts. For the 2014 local elections, out of 22 counties and cities it successfully predicted 18, for an accuracy rate of 82%. For the 2022 Nine-in-One elections, the accuracy rate was as high as 95%, and it was once considered better than the polls of the same period.
However, the 2012 presidential election: market prices clearly deviated from the final outcome. Outsiders questioned “deliberate manipulation” and excessive speculative behavior, which damaged its credibility.
Shut down and transition: an unfinished experiment
In 2010, the Future Events Trading Exchange temporarily shut down due to a corporate restructuring transformation and a dispute over cooperation. Although there were later reports about testing and restarting, overall operations gradually faded from public view. In recent years, most mentions treat it as a historical case rather than a market still operating continuously. Its official website could no longer be loaded, and the last update on its official Facebook page was in 2023.
This means that Taiwan’s prediction market experiment—once ahead of the rest of the Chinese-language world—ultimately did not develop into a long-term, stable, institutional tool.
The fundamental differences between the Future Events Trading Exchange and Polymarket
If you see Polymarket as the representative prediction market of today, then the Future Events Trading Exchange is more like an “academic version for the Web2 era.” Both share the same core concept: standardizing future events into tradable contracts, with market prices reflecting probabilities. The difference lies in the underlying infrastructure and incentive mechanisms:
Polymarket is built on the blockchain, settles in USDC, and involves real financial risk plus global liquidity. The Future Events Trading Exchange used virtual currency—more focused on research and public-issue experiments. The former is a financial market; the latter is closer to a social science experiment.
Looking back from today, prediction markets are not new. Long before blockchain emerged, Taiwan had already tried to predict the future using market mechanisms, and even once gained an advantage in election forecasting. Before Polymarket, there were already many blockchain prediction markets too, but it still remains the most successful.
Before Polymarket, this article: Taiwan had prediction markets long ago—Chengchi University’s “Future Events Trading Exchange” existed 20 years ago. First appeared on Chain News ABMedia.
