《GateLive Roundtable》Episode 10: The "Super Cycle" of Prediction Markets: From Wild Growth to Smart Arbitrage

“Gate Live Roundtable Discussion” is a Chinese-language crypto roundtable interview series created by Gate Live, airing promptly every Wednesday at 8:00 PM, focusing on the most discussed industry topics of the moment. The series periodically invites core practitioners and frontline observers from blockchain, Web3, DeFi, Ethereum ecosystem, stablecoins, compliance, and policy fields for in-depth exchanges in the live broadcast room.

The roundtable emphasizes a relaxed, open, and authentic dialogue atmosphere, exploring market trends, industry disagreements, and key variables from multiple perspectives, helping viewers form clearer, more rational judgments amid complex market narratives.

This episode’s theme: Predicting the Market’s “Super Cycle”: From Wild Growth to Smart Arbitrage

Guests: Well-known KOL in the Chinese crypto community — Chloe, Mr. Misi, Crypto Da Sima

This program’s content is for informational exchange and opinion discussion only, and does not constitute any investment advice.

(This content is compiled from the live replay, with text assisted and appropriately edited by AI. For the full content, please copy the link: https://www.gate.com/zh/live/video/56d31e77ccc127f7f8379919700c9d9e)

Host Jesse:

Hello everyone. Welcome to tonight’s Gate Live roundtable. I am your host, Jesse!

In this cycle of the crypto industry, there’s one unavoidable topic: predicting the market. Because it’s extremely hot—so much so that even Robinhood’s CEO has proclaimed the arrival of a “super cycle of prediction markets.”

Why has a sector once considered “marginal gambling” suddenly become a focal point at the intersection of traditional finance and decentralized finance? Some say 2026 will be a pivotal year for prediction markets transitioning from “event-level” to “state-level,” evolving from simple win-lose games into a new decision-making infrastructure that integrates information, capital, and judgment in a closed loop.

For most of us watching on screen, there are really only two main questions: First, how do you actually play this? Second, how can I reliably make money from it?

Today, we’ve invited three practitioners and observers deeply involved in this field. Chloe, Mr. Misi, Crypto Da Sima—welcome!

Whether you’re an industry trend observer or a trader with capital ready to explore, this conversation might help you clear the fog around prediction markets and see the truth behind this “cognitive bias” game.

Let’s start with brief self-introductions from each guest.

Chloe:

Good evening everyone, I’m Chloe. I entered this industry in 2024, starting with a Web3 fund. We’ve invested in some prediction market projects and observed their development cycles. I’m personally quite interested in this area. Later, I also worked in exchange marketing, and now mainly focus on my own projects, managing my account, and some research and investment in projects. Thanks everyone.

Mr. Misi:

Hello everyone, I’m Misi. I’m a blogger who enjoys “farming alpha” and researching strategies. If you also like farming alpha and researching projects, follow me—I often share project insights on my Twitter. Glad to be here discussing today’s topic with all the teachers and viewers. Thank you.

Crypto Da Sima:

Hello everyone! I am the ten-year king of League of Legends, top three in PUBG Asia servers, and a direct disciple of Wuhu Da Sima—Crypto Da Sima! Entered the crypto space in 2017, started gambling in 2018, went bankrupt on futures in 2019, did DeFi in 2020, farmed NFTs in 2021, went bankrupt on DeFi lending in 2022, rebounded in 2023, and in 2024, with narratives around Solana and AI, returned to the peak. In 2025, I started doing stable arbitrage. I’m happy to meet everyone tonight in the Gate live room. I usually follow trades on Gate, and my research strategies and trading live streams are also shared on Twitter, Bilibili, and YouTube. Welcome everyone to follow.

Host Jesse:

Thanks again to all the teachers for joining. Let’s not waste time and jump straight into today’s topic.

Many people’s understanding of prediction markets still sees them as just “casinos in disguise,” but now insiders are trying to define them as “a new type of information infrastructure.” Why has this qualitative shift happened? What changes did the entry of traditional giants like Robinhood and ICE bring to the sector?

Chloe:

Many think prediction markets are just “legal gambling.” Why are they now called “a new type of information infrastructure”? What qualitative changes did the entry of giants like Robinhood and ICE bring?

The difference between prediction markets and gambling:

1. No house: The platform itself doesn’t set odds or betting lines; all trades follow a “pairing” model. Each bullish bet must be matched with a bearish one to execute. The platform acts purely as an intermediary, a place where bullish and bearish views collide directly.
2. Trading process, not locking in results: Unlike traditional betting, users can buy and sell at any time before the outcome. Their profits can come from changes in event probabilities, not just the final result—similar to stock markets.
3. LP rewards: To incentivize liquidity provision, liquidity rewards are given to users, similar to DeFi protocols.

From the U.S. presidential election and the popularity of Polymarket, prediction markets have challenged traditional polling and stirred Wall Street. Bloomberg has integrated Polymarket’s forecast data, and many quant funds are starting to value prediction data.

More and more people see prediction markets as an effective financial information market. Previously, when an event occurred, various politicians and experts would comment. Now, prediction markets can truly reflect the public’s voice.

Prediction markets turn “anything can be bet on” into reality—everything can be priced, traded, and forecasted. From geopolitical conflicts, tech giants’ futures, to controversial criminal cases, billions of dollars in bets are placed.

• October 2025: ICE (NYSE parent company) invests $2 billion in Polymarket. With ICE’s backing, hedge funds start using prediction markets as official tools for macro risk hedging (e.g., Fed rate fluctuations, M&A success rates).
• January 2026: Robinhood acquires MIAXdx, a compliant entity with DCM (Designated Contract Market) and DCO (Derivatives Clearing Organization) licenses. This means Robinhood is no longer just selling products for others but fully linking prediction contracts with sports betting, macro indicators, and entertainment events. By early 2026, Robinhood’s prediction contract trading volume exceeds $9 billion.

Qualitative significance:

• Data assetization: ICE considers prediction market data a more accurate “real-time sentiment indicator” than research reports.
• Institutionalization: With ICE’s endorsement, hedge funds begin to use prediction markets as formal tools for macro risk hedging.

Previously, prediction markets were called “gambling” because only retail traders played; now, with Robinhood and ICE involved, they become “labs” and “price setters.” When big money starts buying “truth” probabilities, this sector ceases to be a game and becomes infrastructure.

Mr. Misi:

Actually, I was initially quite skeptical about prediction markets—thought they were just a rebranded gambling platform.

But over time, I’ve changed my view. The biggest difference from traditional info dissemination is—some things aren’t just talk; you have to put money where your judgment is.

It’s like betting with friends before: “If you do this, I’ll pay you X.” Right? But to actually realize that, you’d need witnesses or proof. Now, you can create a prediction event directly on the platform, and both sides deposit funds into a pool beforehand.

In this way, prediction markets act like a guarantee platform. When two parties predict an event, the platform guarantees the bet. With Web3’s decentralization, you can see where the funds go, making the betting fairer.

In the past, we saw opinions from media, KOLs, or reports, which are costless and can be wrong without penalty. But in prediction markets, every judgment involves a bet—you must bear the risk of your statement “this event will happen.” Over time, it becomes a more “constrained” way of aggregating information. So, the information here tends to be more “truthful.”

Plus, infrastructure has improved in recent years: on-chain settlement is transparent, rules are written in smart contracts, and anyone can participate. This is the first time these hot topics can run globally—anywhere, anyone can join.

That’s why some now see it as an “information tool,” not just a place for speculation. The reason for its recent popularity is simple—people are more anxious about making judgments, considering more factors. Market changes are rapid; policies, hot topics, and chain activity evolve so fast that by the time traditional info comes out, it’s often too late. Is there a way to see the true market expectation faster? Prediction markets fill this gap. For example, I saw someone say that based on hotel bookings, Trump might or might not come to China in March, and there was already a betting market on Polymarket about it.

Traditional financial institutions like Robinhood bring industry standards—compliance, risk control, product design—that weren’t there before. Once these rules are established, more users will join, and more capital will stay.

So, this sector is no longer just a small circle playing around; it’s becoming a mainstream product.

If you still see prediction markets as “legal gambling,” I think you’re only scratching the surface. Essentially, it’s about using money to express judgments, and market prices to aggregate everyone’s opinions. It solves the “how to price information” problem, not just “how to bet.”

Crypto Da Sima:

First, let’s clarify a common misconception: prediction markets are not gambling; they are an “information discovery mechanism.” Gambling is zero-sum, based on luck or house advantage; prediction markets’ core is price discovery—market prices are the best collective estimate of future event probabilities. This was validated as early as 1988 by the Iowa Electronic Markets: real dollar trading by students on “who will win the presidential election” contracts outperformed polls. Why? Because polls are “questionnaires,” susceptible to lying, laziness, or media influence; prediction markets involve “betting,” where participants use real money to vote, making falsehood costly, and information more truthful.

Philosophically, this echoes Hayek’s 1945 paper “The Use of Knowledge in Society”: no central planner can gather all local knowledge—farmers know weather, merchants know supply chains, intelligence agencies know secrets. Traditional markets solve resource allocation; prediction markets solve the aggregation of event probabilities. They turn private information into price signals, making society function like a superbrain.

Aristotle’s “The Nicomachean Ethics” mentions “the whole is greater than the sum of its parts,” and modern “collective intelligence.” James S. Soro’s classic example: in 1906, villagers in England guessed a cow’s weight with an average error of just 1 pound! Prediction markets upgrade this from “guessing weight” to “predicting presidents, Fed rates, Bitcoin reaching $100K.” It’s no longer entertainment but a knowledge production machine. Philosophically, it touches on “truth”—Peirce’s pragmaticism sees truth as “what all rational inquiry would eventually agree upon.” Prediction markets accelerate and monetize this process: prices are “temporary truths,” constantly iterated with new info.

Switching to math and probability theory: the basic contract is a binary option:

• “Yes” share price = P(event occurs)
• If event occurs, Yes settles at $1; otherwise, $0.

This isn’t gambling but risk-neutral pricing of probabilities. Under no-arbitrage, the price P equals the risk-neutral probability. Why? If P > true probability, smart money sells Yes, pushing the price down; if lower, they buy. This is an extension of the Efficient Market Hypothesis (EMH) into the event space.

Deeper still, by the Law of Large Numbers: as N → ∞, with independent information fragments from each participant, the market price converges to the true probability with probability approaching 1. In reality, during the 2024 US election, Polymarket’s daily volume exceeded $400 million, with over 2 million addresses, and the probability of Trump winning converged from 60% to within 1% of the final result—while FiveThirtyEight’s polls had a 3-4% error. This is the victory of the Law of Large Numbers—not because “more people means more power,” but because “money votes” filter out noise.

Advancing to Bayesian probability: this is the real core of the “superbrain.” Each trader has a prior:

• A Wall Street analyst’s prior from models;
• An on-chain whale’s prior from wallet flows;
• An ordinary retail trader’s prior from Twitter trends.

When they bet, they’re updating their posterior:

P(H|E) = [P(E|H) × P(H)] / P(E)

where H = “event occurs,” E = “market price observed.” But the magic is, the market price itself is the global posterior! Each new piece of info (e.g., Fed’s dovish speech) updates traders’ beliefs, which are then reflected in buy/sell actions. The market price becomes a real-time Bayesian aggregator.

For example: in November 2025, there’s a contract on Polymarket about “Fed cutting rates by more than 50bp before March 2026.” Initial price: 42%. Suddenly, CPI data comes in below expectations, and the price jumps to 68%. A trader with a $1 million position (I know this address) updates as follows:

• Prior: 35% based on historical data
• Likelihood: CPI data under “rate cut” hypothesis has an 85% chance
• Market jumps: he buys more Yes, raising his own posterior from 52% to 71%
• Result: market stabilizes around 65%, becoming the best collective posterior.

This is Bayesian magic at the collective level: no one has all info, but the market “has” all info’s expectation. Philosopher De Finetti said “probability is a measure of subjective belief,” and prediction markets force countless subjective beliefs into a monetary consensus, producing an “objective” probability. It surpasses gambling, becoming an information infrastructure—like GPS aggregates real-time locations of drivers, prediction markets aggregate real-time beliefs of all participants.

Why the “super cycle” suddenly erupts now? There are three core drivers:

1. Infrastructure maturity: Blockchain (especially Polygon, Base) reduces contract creation costs from tens of thousands to a few dollars, with automatic settlement and no middlemen. In 2024-2025, Polymarket’s volume surged from $250 million during the US election peak to over $425 million early 2026, driven by on-chain settlement and one-click wallet login. Traditional betting sites require KYC and regional restrictions, but prediction markets are global 24/7.
2. Narrative and macro resonance: The 2024 election pushed prediction markets into headlines: Trump’s win probability on Polymarket outperformed polls, and media like CNN and Fox News started citing market prices as “real public opinion.” This creates positive feedback: media cite → more participants enter → liquidity improves → prices become more accurate → media rely more. Post-2025, Robinhood’s CEO publicly predicts a “trillion-dollar super cycle,” shifting retail attention from meme coins.
3. Entry of traditional finance giants: Bringing legitimacy and capital infusion.

Prediction markets are completing a “democratization of information.” Previously, only Wall Street elites, think tanks, and intelligence agencies grasped “future probabilities”; now, anyone with a phone can participate in global knowledge production with a few hundred dollars. Robinhood and ICE are not legalizing gambling—they’re transforming gambling into a scientific institute—using money to incentivize truth, algorithms to aggregate wisdom, and blockchain to ensure fairness.

Host Jesse:

Now, with the market’s current state, airdrops are hard, secondary markets are tough. Prediction markets seem like a new way out.

I’d like to ask the teachers: for ordinary retail traders wanting to enter prediction markets now, what’s the simplest profitable approach? Is it just “guess the result,” or are there other methods? What are some practical ways for ordinary people to operate from their computers?

Mr. Misi:

I think prediction markets aren’t just “guess and make money.” If you rely solely on intuition or “I think” feelings, you’ll probably lose.

The real winners are not just “guessing the result,” but doing trades.

The simplest, most accessible method for ordinary people is—look for price deviations.

For example, if an event’s probability is at 60%, ask yourself: is this high or low? If you think it’s undervalued, buy; if overvalued, sell. Essentially, you’re trading the “difference” between market price and your own judgment.

Another easier method is to watch emotional swings.

Sometimes, a news event causes the market to spike or drop sharply—often overreacted. You don’t need to judge the final outcome, just whether the sentiment is exaggerated, and then do a contrarian move. This is more like short-term trading.

A third approach is to exploit information gaps.

Some events’ info propagates with delay—overseas news, policy shifts, on-chain changes—some see early, others later. Whoever reacts first gains an advantage. You don’t need to be a pro, just have fast, broad info channels.

Another safer method is diversification.

Don’t bet everything on one outcome; spread across multiple events to share risk. As long as your overall judgment is correct, you can profit.

So, making money in prediction markets isn’t about “guessing right,” but about “understanding prices better than others.” You’re trading probabilities, not just betting on wins or losses. If you approach it as “I need to guess correctly,” you’ll likely get educated by the market. But if you see it as a tradable market, ordinary people can definitely participate from their computers.

Crypto Da Sima:

For ordinary people, how to reliably make money in prediction markets?

Many first think “guess the result,” but I tell you: relying solely on guessing, 70% of retail traders lose money (Dune data 2025: only 30% of 1.7 million addresses on Polymarket are profitable; most small traders lose). The real trick isn’t “guessing,” but three things: 1. Find mispricings due to information advantage; 2. Provide liquidity and earn fees + platform rewards; 3. Buy early and sell during price swings, not hold till settlement.

• First major method: information advantage betting (easy entry, suitable for beginners with $100 capital). Core logic: market price = collective posterior probability, but if you have private info, you can spot mispricings. It’s not luck-based, but expected value (EV) calculation. Simple formula: EV = (your true probability P × profit) - ((1-P) × loss). As long as EV > 0, consider participating for long-term positive returns.
• Second method: early buy and sell (momentum scalping, not holding till settlement). Prediction markets aren’t stocks—you can trade during price swings before the event. When news hits, prices jump 5-15%. You buy low, sell high within 1-2 days. Steps:
  • Set up news alerts (Google Alerts + Twitter lists).
  • When market overreacts, do contrarian trades: sell when rumors push prices up, buy when data is good but market hasn’t fully reacted.
  • Target: 5-10% profit and exit.

Chloe:

For ordinary retail traders, what’s the simplest way to profit now? Is it just “guess the result,” or are there other methods? What practical ways can ordinary people operate from their computers? Not just theory, but actionable.

1. Focus on areas you’re familiar with—like predicting World Cup, basketball games—and participate accordingly. There are also more grounded ways, not just Polymarket. During Spring Festival, I saw Probable (which got YZI Labs investment) run special prediction events on small topics like Spring Festival sketches, CZ’s New Year wishes, etc. Now, some projects predict TGE dates or exchange listings—these are also options. You can position yourself early, monitor info closely.
2. Farming airdrops: prediction markets are a hot sector with high airdrop potential—Polymarket, Probable haven’t issued tokens yet.
3. Arbitrage in single markets—least reliant on “guessing.” During sharp market moves, due to imbalance of buy/sell pressure, P(Yes) + P(No) can be less than 1, creating a mathematical anomaly.
   1. Scenario: after a sudden news, everyone dumps Yes, price drops to $0.55, while No remains at $0.40.
   2. Action: buy one Yes and one No simultaneously. Total cost: $0.95.
   3. Result: regardless of the final outcome, you hold a contract worth $1, earning a 5% spread.
4. “Rebate farming”: become a liquidity provider (LP). If you don’t want to judge outcomes, you can “sit and earn.” Platforms like Polymarket need liquidity on both sides.
   1. How: use official or third-party market-making bots (e.g., Polymarket’s reward programs).
   2. How: place buy orders at $0.49 and sell at $0.51. When trades happen, you earn the $0.02 spread.
   3. Income: besides the spread, many platforms will give trading rewards to active market makers—these are pure profit.

Three iron rules for retail traders:

1. Don’t gamble on “50/50”: unless you have insider info, avoid pure 50/50 guesses—they’re gambling.
2. Don’t trade in “small pools”: low-volume markets are easily manipulated by big players, risking being countered.
3. Don’t forget fees: calculate withdrawal and trading costs beforehand. If the arbitrage margin is only 0.5%, it might not cover network fees.

I’d like to hear from other guests as well.

Host Jesse:

Next, let’s discuss whether there are truly stable arbitrage strategies.

There’s a widely circulated “Polymarket arbitrage bible” claiming that arbitrage has become a mathematical arms race. For example, are multi-event (composite) markets offering more arbitrage opportunities than single markets? How to identify logical dependencies? How can ordinary people find high-probability, high-win-rate strategies?

Crypto Da Sima:

Are there truly stable arbitrage strategies? That “arbitrage bible” suggests it’s a “mathematical arms race.” It states: simple buy YES + NO < 1 is outdated; now it’s about edge polyhedra + Bregman projections + integer programming. Are composite markets (betting on multiple related events simultaneously) offering more opportunities? How to find logical dependencies? Do ordinary people have high-win-rate strategies?

Conclusion first: pure riskless stable arbitrage for retail is nearly impossible—window times are only 2-3 seconds, eaten up by high-frequency bots and quant teams. But high-probability strategies (statistically EV > 0, win rate 55-70%) are possible, relying on AI, information asymmetry, and simple logical checks. Not “guaranteed profit,” but “smart people using math and AI to exploit info advantages.”

First, define: stable arbitrage = profit regardless of event outcome (risk-free). High-win-rate strategies = long-term EV > 0, win rate > 50%, with fluctuations. In the 2026 supercycle, Robinhood + ICE bring liquidity and legitimacy, but also squeeze out simple opportunities—now it’s AI vs AI, math vs math.

• First layer: basic arbitrage (YES+NO<1)—why can’t ordinary people do it? Classic in-market arbitrage: if YES price + NO price < 1, buy both, settlement guarantees $1. Mathematically, this violates no-arbitrage assumptions: in LMSR (logarithmic market scoring rule), prices must satisfy P(Yes)+P(No)=1, or a Dutch Book exists. Data from 2025-2026 shows single-market arbitrage accounts for 26% of total profit (~$10.58 million), with top addresses making hundreds of thousands per trade. But the window is very short—robots scan via WebSocket, snatch profits in 2-3 seconds. Manual traders? By the time you click, prices are fixed, and slippage + fees eat your margin.
• For ordinary people, high-win-rate strategies (I think possible but not riskless):
  • Use AI to find info advantages: tools like Claude scan Polymarket for hidden dependencies, ask “Are these markets logically linked?” Use Excel to compute joint probabilities, pick price differences >3%.
  • Follow “smart money”: track top AI bots’ small positions.
  • No-loss farming: filter markets with high “No” probability, win about 60% of the time.

Using AI models, ordinary traders can detect small dependencies, and with long-term effort, achieve positive EV.

Chloe:

Everyone, pay attention to “composite arbitrage.” What is “logical dependency” in markets? How can ordinary people find these?

You don’t need calculus—just understand “if A happens, B probably also happens.”

• Example: Market A predicts “a major CEO resigns,” Market B predicts “the stock drops sharply.” If Market A’s probability is at 90%, but Market B’s is only 10%, and they’re logically linked, you can buy the “stock drops” contract, earning from the market’s delayed correction.

Cross-platform “arbitrage tools”:

• Don’t manually monitor. Use tools like AlertPilot (popular in 2026) or similar, which compare Polymarket, Kalshi, and emerging Cboe markets. When price differences exceed 3%, they notify via Telegram, and you can act.

Mr. Misi:

This I can clarify: “stable arbitrage” in prediction markets is basically unavailable for retail traders now.

Many so-called “arbitrage bibles” are actually based on team-developed models and automation. If you rely on manual detection of perfect riskless spreads, you’re already outpaced.

So if you think of “riskless, guaranteed profit,” that stage is over—it’s a matter of math and speed, not ordinary retail capability.

But that doesn’t mean no opportunities exist—just need a different mindset.

For ordinary traders, it’s more about “structural arbitrage,” not pure math.

Simple example: composite markets.

Big events are often broken into sub-questions with logical relations. If the market prices are inconsistent with these relations, you can exploit that.

For instance, if “CEO resigns” is highly priced, but “stock drops” is undervalued, and you know they’re linked, you can buy the undervalued one, earning from the correction.

Tools:

• First, identify strong correlations.
• Second, check if prices are misaligned.
• Third, monitor news and sentiment for overreactions.

This approach is more accessible—no need for complex calculations, just understanding event relationships.

You profit from the market’s slow adjustment, not from system loopholes.

In summary:

• Pure arbitrage is nearly impossible for retail now.
• But exploiting logical mispricings based on event relationships remains feasible.

The key isn’t just calculation but understanding event connections.

Host Jesse:

Great, thank you all for the insightful sharing. Thanks to everyone watching. Today, we clarified that prediction markets are not gambling—they’re a new decision-making tool based on information, capital, and judgment. From wild growth to smart arbitrage, the key is whether we can be more rational and clever than others.

If you found today’s live session helpful, remember to follow our channel and the three teachers.

See you next time! Bye!