Polymarket vs. Kalshi: Which platform will ultimately stand out?

Author: Payment 201

In this episode of Unchained, Markus, founder of 10x Research, delves deep into the core of predicting market competition. He also discusses: Will more platforms emulate Polymarket and launch their own tokens? He elaborates on a nearly certain trading opportunity hidden within Polymarket and shares 10 strategies for participating in prediction market trading without holding opinions on the events themselves. One key insight is: "What is truly valuable is not the wisdom of the crowd, but the wisdom within the crowd."

Takeaways:

1. The essence of prediction markets is probability pricing markets, structurally closer to one-touch barrier options than traditional gambling.

2. Every prediction market trade has a clear expiration date and certainty of win or loss, unlike long-term hold crypto assets.

3. Current prediction markets are still in their early stages, but trading volume and user scale are rapidly approaching mainstream financial platforms.

4. The real moat is not product functionality, but rather the concentration effect of liquidity and trading volume.

5. Professional traders do not "predict outcomes," but seek structural pricing errors.

6. The most stable returns come from events with probabilities that have not fully converged as the time approaches.

7. Low probability "moonshot" trades contribute to most losses and should be systematically avoided.

8. Endgame sweeps and time decay captures are the most cost-effective strategies in prediction markets.

9. The truly valuable signals in the market come from a few high-quality participants within the crowd, not the average opinion of the crowd.

10. Prediction markets are evolving from entertainment betting to professional probability trading and risk management tools.

Laura Shin:

Hello everyone, welcome to Unchained, a show that focuses on the real situation in the crypto industry without hype. I am your host, Laura Shin, and thank you for joining this live broadcast. Before we begin, a quick reminder: anything you hear on Unchained does not constitute investment advice. This show is for informational and entertainment purposes only, and my guests and I may hold assets discussed in the show. For more disclosure information, please visit unchainedcrypto.com.

Laura Shin:

Today's guest is Markus Thielen, CEO of 10x Research. Welcome, Markus.

Markus Thielen:

Hi, Laura, thanks for having me.

Laura Shin:

I'm looking forward to discussing this topic with you. The prediction market space is really hot right now. The two biggest players are Polymarket and Kalshi, both of which have very high valuations and are in fierce competition with each other. I think most people on X can feel that.

And with Polymarket now starting to enter the U.S. market, this competition will only escalate further. Meanwhile, Gemini, Robinhood, and some other platforms are also entering this space. However, this sector is facing some regulatory hurdles. Nevertheless, none of this has stopped the trading volume in prediction markets from continuing to grow, almost increasing every month. In November alone, the trading volume approached $2 billion, at least for Polymarket and Kalshi.

What do you think about the current stage of prediction markets? Where do you think they are on the adoption curve?

Markus Thielen:

Yes, absolutely. One very interesting point is that trading volume has indeed picked up, and it is now at a relatively stable high level. I believe the current trading volume is around $1 billion per week, which is quite high. It's somewhat similar to when Bitcoin dropped from around $100,000 to about $25,000; it was during that phase that the trading volume and activity in prediction markets began to rise significantly. The number of weekly users has actually grown from about 70,000 to nearly 250,000. These numbers are quite substantial, indicating that many people are getting involved.

Of course, this is also related to the fact that Kalshi and Polymarket have collectively raised about $3 billion in the past few weeks. This means they have very strong market promotion capabilities.

I think as we look towards 2026, we will see many things gradually unfold. I believe more investors and traders will start using these platforms to hedge some economic risks, macro risks, and of course, to speculate.

However, currently, about 90% of the trading volume still comes from sports betting. On the crypto-related prediction market side, it remains a relatively small niche, but it is growing.

From multiple perspectives, this market is very interesting. So if I had to judge, I would say prediction markets are still in a very early stage. Because these platforms barely existed a year or two ago, the trading volume really began to grow significantly after the Trump election; that was the first real push.

If we look at the activity from last month, for example, how many users visited these platforms, as a comparison: about 40 million users visited Robinhood, around 30 to 32 million users visited Coinbase, while Polymarket had nearly 20 million.

So from a user perspective, these prediction markets are becoming increasingly attractive, starting to attract significant trading volume across various types of events. I think we are at a starting phase now, and as we enter 2026, many things will accelerate further. As you mentioned earlier, Gemini has just obtained a license for prediction markets. So competition is heating up, and many people want to enter this field and get a piece of the pie.

Laura Shin:

I want to follow up on something you mentioned earlier. You said that when Bitcoin's price started to drop, the trading in prediction markets actually began to rise. Are you suggesting that some of the people who were originally trading Bitcoin have now turned to trading prediction markets? How do you assess this?

Markus Thielen:

I think it's more of a coincidence. If a large number of people suddenly abandoned crypto exchanges to run to prediction markets, I think that would be a bit of a stretch.

This is one of the questions we are trying to study and clarify, but so far, that doesn't seem to be the case. For example, when we ask our subscribers, not many of them are actually involved in prediction market trading. Many are interested, but the actual trading volume is not that large. Many contracts and bets are relatively small in scale; they do not have the liquidity scale we are used to seeing in the crypto market. In the crypto market, you often see $100 billion, $200 billion, or even $300 billion in trading volume, while prediction markets remain a very niche market. Nevertheless, I do believe there are some interesting opportunities here.

From another perspective, this is also a continuation of the "gamification" of financial markets. People want to be attracted to entertaining financial products; it’s just a matter of whether you are participating in the market from an entertainment perspective or from a probability perspective. But in any case, this market is expanding, and it has become a new tool.

If we look at how people historically arbitraged between different price differentials and products, it is actually quite similar to what we are studying in prediction markets: Are there certain structural arbitrage opportunities?

Laura Shin:

It sounds like you are trading in both the crypto market and the prediction market. How would you describe the differences between these two types of markets in terms of trading?

Markus Thielen:

Of course. One thing that many people may not realize is that prediction markets are essentially a form of exotic options trading. It is fundamentally about probability, about understanding probabilities, and it heavily relies on the speed of reaction to news. These are all contracts with deterministic outcomes. Essentially, it’s a "yes" or "no," with no middle ground. In the end, one side wins, and the other side loses. In crypto exchanges, for example, when trading an asset, the situation is completely different. As long as you hold the asset and the price goes up, all holders can potentially make money. So this is a completely different narrative structure.

I would say that prediction markets and crypto markets are structurally very different. But they also have some similarities, especially for mature traders. If we use more "hardcore" options terminology, prediction markets are very much like one-touch barrier options. You need to understand how to price these contracts and understand the factors that affect probabilities. In the crypto market, many times people are buying a narrative, a theme, a story. I think that is the biggest difference between the two.

However, for mature traders, whether in prediction markets or crypto markets, their way of thinking is actually very similar. Of course, everyone knows that many crypto exchanges have their own internal trading teams, fund management teams, or liquidity provision teams. These teams may be proprietary to the exchange or affiliated parties.

In the past two to three years, many things have been revealed, showing how these exchanges have "engineered" liquidity in the early stages. Because without liquidity, retail users cannot match with retail users directly. Without liquidity, there are no trades; without trades, there are no users. We saw this in the early days of BitMEX. I remember when Arthur Hayes was doing a presentation in Hong Kong in 2015, he mentioned that there were a lot of Korean retail investors trading high-leverage futures, with implied volatility and funding rates being very high, and he hoped to attract institutional traders to arbitrage on the other side.

In prediction markets, the situation is quite similar. Many people may not know that there are also very mature professional trading teams on Polymarket and Kalshi. Some are proprietary to the platform, and some are teams that the platform hopes to establish. These teams are trading around the clock and are very professional.

The similarity here is that there are a lot of unique contracts and unique betting targets in prediction markets, but you still need market makers willing to stand on the opposite side; otherwise, you cannot attract trading flow. So the structure remains: patient market makers vs. impatient takers. This is very similar to the crypto market. But the overall structure is still different. However, this structural similarity does allow mature traders to seek arbitrage opportunities across different platforms.

Laura Shin:

Let's talk specifically about the two biggest players right now. We mentioned earlier that other platforms are entering, but let's focus on the current main competitors, Polymarket and Kalshi. They will soon be competing head-to-head in the U.S. market. Polymarket's application is now gradually launching in the U.S., in what can be considered a beta phase. Polymarket has raised $2 billion, with a valuation of $12 billion; Kalshi has raised $1 billion, with a valuation of $11 billion. How would you assess the strengths and weaknesses of these two companies?

Markus Thielen:

The key difference is that Polymarket is a crypto-native platform.

Its on-ramp is very fast, and the account opening process is extremely simple. For non-U.S. users, someone sent me a screenshot showing they were 240,000th on the waiting list, indicating very strong demand. If you are not a U.S. user, opening an account takes basically two minutes, and you can deposit directly with cryptocurrency. This process is similar to many early crypto exchanges: you just need an email, receive a verification code, and then your account is set up. After that, you will have a wallet address where you can deposit different cryptocurrencies or stablecoins, and then you can start betting. The whole process takes less than two minutes. If you are a regulated platform in the U.S., the process will be much slower due to long backlogs. That’s one difference.

Another difference is liquidity. From an institutional perspective, most bets are still concentrated in the sports betting area, accounting for about 90%. Other areas, such as event contracts and crypto-related predictions, are still relatively small.

Of course, the situation will be different during election periods. For example, during the U.S. presidential election, about $3.7 billion was involved. The products on both platforms are highly overlapping in sports betting and political events. Future differentiation will become less and less. I believe they will ultimately converge on wherever there is trading volume.

Additionally, users can also initiate new bets themselves; as long as the proposal is approved, you can bet on both platforms simultaneously, and if the odds are different, you can even arbitrage. Currently, they differ in style and regulation, but they will ultimately converge, especially since the U.S. market is the largest market.

Laura Shin:

Let's talk about these new competitors. For example, Gemini has launched Gemini Titan, Robinhood and Susquehanna have partnership announcements, and a few months ago, Limitless had a rather controversial token issuance. Of course, there are definitely other players as well. What do you think of these newcomers? Which ones do you think might become truly competitive rivals?

Markus Thielen:

In my personal view, the two big platforms are still the most important. They dominate almost all trading activity. Because ultimately, everything in this market depends on liquidity and trading volume, and whether you can attract trading volume very quickly. We have seen this many times in the crypto exchange space: trading volume is key. Unless some significant regulatory event occurs, like those that impacted some exchanges in the past few years, from Mt. Gox to BitMEX, the market tends to concentrate where the trading volume is highest. You can see how Binance grew. It was precisely because of its trading volume that it attracted more users and ultimately established its leadership position. Other platforms that want to get a piece of the pie must come up with very smart strategies to attract trading volume.

In prediction markets, the situation is similar. If you are a smaller prediction market platform, you also need trading volume. And to get trading volume, you need some professional traders.

You mentioned Susquehanna; they are very active in the professional market-making space. I believe they are collaborating with multiple platforms, which is exactly their role as market makers. So the key question remains: How do you "engineer" trading volume? This has always been a challenge faced by crypto exchanges, and it is no different in prediction markets.

Laura Shin:

Among the current competitors, who do you think is leading in trading volume?

Markus Thielen:

That certainly depends on the specific contracts, but overall, Polymarket is slightly ahead. If you look at last month's visitation data, Kalshi had about 5 million monthly visitors, while Polymarket had about 19 million. This is related to Kalshi being more U.S.-focused, while Polymarket is a global platform with a simpler account opening process. Additionally, Polymarket has frequently appeared at crypto conferences and various events over the past year or two, increasing its market exposure, which is also one of the reasons for its lead.

However, from a valuation perspective, the two are very close. Kalshi may have a better revenue model because it charges relatively higher trading fees, while Polymarket currently has lower fees. But because Polymarket is a crypto exchange, it does not have to bear as many traditional regulatory costs and has not prioritized the U.S. market early on, so the compliance pressure is relatively small. However, this situation is changing now as they are entering the U.S. market.

So I believe some of the differences between them will ultimately converge. The final question is: who can attract more users and who can establish partnerships with larger market makers? Because everything comes back to liquidity: no liquidity → no retail → no institutions → no market. This logic has been repeatedly validated in crypto exchanges and is exactly the same in prediction markets.

Laura Shin:

This might also relate to the recent debates on Twitter. I saw some charts on The Block showing that in recent months, especially the past two months, Kalshi's trading volume seems to have surpassed Polymarket. However, I'm not quite sure how these data are calculated, as there are some differences in calculation methods. I won't go into details. But I want to ask about the differences in market structure across platforms. You have also written about this in your blog. Some use traditional limit order books, while others have different structures. Can you talk about these differences and how they might affect users' trading behaviors?

Markus Thielen:

From the end-user perspective, I don't think there is much difference. They are basically using limit order book structures, fundamentally still patient market makers vs. impatient takers. Market makers will post a large number of orders in the market, waiting for retail to hit their buy or sell prices. When market liquidity is insufficient, the spreads will widen. This is very similar to crypto exchanges.

So from the user perspective, as long as there is enough trading volume, the structure itself is not important. The key is: without trading volume, the spreads will be too wide, and trading costs will be very high. Every time you cross the bid-ask spread, you are essentially losing money. This is why prediction market platforms must place a strong emphasis on building liquidity.

We have already seen that some platforms have started to establish partnerships with professional market makers. I also believe Kalshi has its own internal trading team acting as market makers in the background. Polymarket is also trying to establish a similar system. Because, again, trading volume is key. From the user perspective, whether it is a limit order book is not important as long as the liquidity is sufficient. Of course, this also brings to mind Robinhood's revenue model through payment for order flow in the stock market, where there are some similarities.

Currently, prediction markets have about $1 billion in weekly trading volume, which is still not large, but it can grow. We can also use data tools like Token Terminal to compare the changes in trading volume between Kalshi and Polymarket. It is possible that Kalshi's recent increase in trading volume is related to the start of the NFL season, as they are stronger in sports betting, while Polymarket's users are more global and may not be as interested in American sports.

But ultimately, both have raised $3 billion, which is a huge amount of capital, and they are both preparing for expansion into the U.S. market. I believe many things will happen in the coming year.

Laura Shin:

Okay, let's shift our perspective slightly to next year. We have set the stage, with these two main competitors and many new entrants. One important event we haven't discussed yet, but I think is very significant, is that Polymarket will issue its own token, POLY. Interestingly, they have chosen to do this through an airdrop, which is not the most popular method at the moment. I would love to hear your thoughts: how do you think Polymarket should make this airdrop successful and use it to solidify its leading position?

Markus Thielen:

Yes, I think from past experience, airdrops have indeed become very common. Of course, some airdrops are very successful, while others are not so successful. But here, a very interesting point is that Polymarket itself is a crypto-native platform, and everything is on-chain. Therefore, this data can be well analyzed. For example, betting behaviors can be analyzed, and we can see how funds are flowing.

This is also one of the important reasons we are writing related research reports—an airdrop is coming. This information has been confirmed by the CEO and the head of growth. An airdrop will definitely happen. And because anyone can easily open an account, you will have a wallet address, so the airdrop can be smoothly distributed directly to your account. I do believe this airdrop could be very successful. Because people are looking for airdrops, that is a fact.

This is also why, for example, investing in the BNB token itself is a strategy, because if you sell on the first day of the airdrop, you could earn about 10% more this year. So I think people are actively looking for these "free yield" opportunities.

From a crypto perspective, this is also why people prefer the Polymarket platform. Because if you trade there, you automatically qualify for the airdrop. Moreover, trading volume is actually concentrated among a few large players. We have studied this and seen some other research: if your cumulative trading volume on Polymarket reaches $50,000, you are already in the top 1% of users.

So from this perspective, at least at this stage, becoming a "big player" is not that difficult, making it easy to qualify for the airdrop. I believe the potential value of this airdrop could be quite substantial. Because competition is clearly heating up, and you want to reward your core users. Historically, rewarding users has always been a very effective strategy.

We have seen similar situations on Hyperliquid. About a year ago, it conducted a token issuance, and since then, trading volume has started to rise in tandem. A similar situation could happen here. Because anyone holding the token will, to some extent, become a "marketer" for this protocol.

So I think this airdrop is very interesting, and its value logic is sound. I am confident it will happen, and it will happen sooner than many people think. If I had to guess, I would say it might happen in the first quarter of next year (Q1). Because I think Polymarket wants to gain a first-mover advantage in the competition. And being crypto-native, it doesn't need to spend much time preparing. So we believe now is the best time to participate and observe these opportunities.

Laura Shin:

Do you think we should expect other platforms to take similar actions and issue their own tokens?

Markus Thielen:

Personally, I don't think so. Because to truly become a crypto-native platform is actually very difficult. And that is Polymarket's huge advantage. If Kalshi suddenly issues a token, but it is not a crypto-native platform, it would be very hard to define what that token is actually useful for. For example: Will there be buybacks? Will there be trading fee discounts? Will there be other incentive mechanisms? In contrast, on a crypto-native platform, these things are much easier.

We have seen many such cases in the past: crypto exchanges issuing tokens to provide fee rebates, special tier benefits, etc. Of course, if you exclude the failed case of FTT, overall, the performance of crypto exchange tokens has been quite good. That is also why I am willing to draw comparisons. If you look at history, you will find that many exchange tokens have performed quite well. So if you can obtain these tokens through an airdrop, it is not a bad thing in itself.

Laura Shin:

I can't think of many right now; the first one that comes to mind is BNB. What are some other successful exchange tokens?

Markus Thielen:

You can look at Bitget; it has its own token and has performed quite well. OKX also has a token. In fact, you can list many crypto exchange tokens. Their performance this year has been surprisingly good.

For example, there is an exchange in Europe called WhiteBIT. I hadn't really heard of it until May of this year, but its token has risen about 100% since May, while the overall crypto market has not performed well during the same period. Bitget's token also performed quite well at the beginning of this year. So interestingly, even in a year when altcoins performed poorly overall, some exchange tokens have outperformed the market. Of course, Hyperliquid has also performed well since it issued its token a year ago, although it has recently pulled back a bit. But overall, holding these tokens is, to some extent, an exposure to the entire ecosystem. That is why I think it is meaningful.

As you mentioned, BNB's performance has certainly been outstanding. Moreover, users can still earn returns through some ongoing airdrop mechanisms. In the future, we may see some "permanent airdrop" structures that continuously create value for these token holders. If the token economics are designed well, they can indeed help these platforms establish a foothold in the crypto space. Of course, it should be emphasized that most trading in prediction markets is still concentrated in sports betting, which is a different market. But the crypto niche within prediction markets is still a market that can be continuously built.

Laura Shin:

Before we dive into how you specifically trade in prediction markets, I want to first have you explain the risks of trading in prediction markets compared to regular crypto trading.

Markus Thielen:

Of course. Prediction market trading involves event contracts, and these contracts have a clear end time. This is similar to options and somewhat like early crypto futures—back then, there were expiration dates, unlike the perpetual contracts that everyone trades now. Options also have expiration dates; whereas crypto assets, like altcoins, can be seen as a kind of open-ended option. As long as you continue to hold, you always have a chance to make money, even to achieve excess returns.

But in prediction markets, it is not like that. The contracts in prediction markets have a clear end time. Some contracts have long durations, while others are short. Their core operates based on probability, essentially a form of barrier option. You need to truly understand what probabilities you are implying and at what price level you are buying.

Generally speaking, these contracts are priced between 1 cent and 100 cents, that is, $0.01 to $1. If you buy a contract at 60 cents, it means you are betting that the probability of this event occurring is 60%. Moreover, most contracts are structured as binary events of "yes/no." Therefore, timing is crucial.

For example, if you bet that Bitcoin will reach $100,000 by the end of the year, that contract ends on December 31. If you just bought Bitcoin itself, even if it reaches $100,000 on January 1, you can still make money. But in the prediction market, you lose because the contract has expired. So these contracts are essentially probability-based option contracts, not the kind that can be held long-term while waiting for the narrative to play out.

Additionally, probabilities can change. We have seen this happen; for example, when Trump suddenly started interviewing new candidates for the Federal Reserve chair, the market's probabilities for certain candidates changed. I remember the Financial Times reported that Trump was interviewing more candidates that week, and as a result, the probability of Kevin Hassett becoming the Fed chair dropped from 80% to 70% in just a few minutes.

Professional traders will see news headlines first and quickly adjust their positions, reprice, and profit from it. If you are just a non-professional trader, you might notice a day or two later, or even wonder why your P&L has changed. This is the advantage of professional traders: they are at the center of the news flow, trading volume flow, and liquidity flow, which creates a huge difference.

Laura Shin:

Okay, let's talk about a specific article you wrote titled "An Almost Certain Bitcoin Trade on Polymarket with an Annualized Return of 63%." Can you explain what this trade specifically is? How were those numbers calculated? Why can it achieve a 63% annualized return under nearly certain conditions?

Markus Thielen:

Yes. The 60%+ annualized return we are talking about actually corresponds to about 4% absolute return from now until the end of the year. For many crypto traders, 4% might not seem high; it is a very small return. But if you look at it over a period of just a few weeks, the annualized return is actually quite substantial. Especially when your capital can continuously roll between different trades, this return will be magnified. I think the key is to find these high-certainty trades.

In our report, we listed 10 different strategies for people to reference and use. The specific trade you mentioned revolves around the Bitcoin ETF. The question is: will the inflows into Bitcoin ETFs in 2025 exceed those in 2024? If you aggregate the data: last year, Bitcoin ETFs attracted $33.6 billion in inflows; this year, so far, it is about $22 billion. There is a $11 billion gap in between. So the question arises: is it possible for Bitcoin ETFs to attract $11 billion more from now until the end of the year?

We can analyze this problem mathematically. We can calculate probabilities and run Monte Carlo simulations, just like we do in option pricing. We ran 200,000 simulated paths, and the results showed that the probability of this happening is almost zero, specifically, it is in the seventh decimal place. But even so, you can still earn about 4% from this trade.

So if you believe that the inflows into Bitcoin ETFs in 2025 cannot exceed those in 2024, because that would mean needing to add another $11 billion in inflows in a very short time, then you can make this trade.

There are only 14 to 15 trading days left, and there is also the Christmas holiday in between. We know that since October, the inflows into ETFs have noticeably slowed down. This year, on average, daily inflows are less than $100 million. But to achieve this goal, we would need $700 million in inflows daily from now on, which seems extremely unlikely. Especially if at the Fed meeting, Powell appears more hawkish, which is what the market generally expects, it would further suppress institutional investors' participation. So, long story short, from a mathematical perspective, this outcome is almost impossible.

But the market still leaves you a space to earn 4% return, and you just need to stand on the opposite side of this trade. We have observed many similar trades—they are mathematically almost impossible, but the market still provides a pricing premium. This is how we view prediction markets.

Laura Shin:

Next, in another blog post you wrote, you listed many different strategies explaining how to make money in prediction markets, and even without needing to hold any opinions or judgments about the events themselves. I find this very interesting. So let's go through them one by one. The first strategy you refer to as cross-market arbitrage. Can you explain what this means and how you use this strategy?

Markus Thielen:

Of course. If Kalshi and Polymarket give different probabilities for the same event, that in itself constitutes an arbitrage opportunity.

For example, contracts related to the U.S. elections, or future midterm elections, or "who will be the next Fed chair." The probabilities on one platform may adjust faster than on another, depending on which professional market makers and participants are on the platform. These trades tend to be quite systematic and require very fast execution capabilities and sufficient capital.

Because unlike in crypto exchanges, in prediction markets, you cannot quickly move stablecoins back and forth between two platforms at any time. But it is indeed a strategy.

If two platforms give very different odds for the same event, you can go long on one platform and short on the other to hedge your risk. This is one of the lowest-risk strategies we listed. In our report, we rank these strategies from lowest risk to highest risk. We tend to participate in those lowest-risk strategies because historically, these strategies have the highest probability of making money. Those "moonshot" trades, which bet on extreme outcomes, often do not succeed. Smart traders and market makers usually place their funds on high-probability trades, rather than low-probability bets.

Laura Shin:

In practice, does this mean you are betting the same amount on both markets? Then regardless of which outcome occurs, you earn the difference between the winning side and the losing side?

Markus Thielen:

Yes, that's basically it. Of course, you also need to consider transaction costs. For example, Kalshi will have some fees, while Polymarket may not, or the fee structures may differ. There are indeed some small differences between the two platforms, and they haven't designed arbitrage to be that easy. It’s a bit like the early crypto market: in the initial stages, arbitraging between different exchanges was not easy. But arbitrage opportunities do exist.

Especially during last year's election, these opportunities were very obvious. Because at that time, the demographic structure of people trading on these platforms was different, and as Polymarket entered the U.S. market, these differences may gradually disappear.

Laura Shin:

I want to confirm, when you say the platforms haven't made arbitrage easy, do you mean they have deliberately designed it that way?

Markus Thielen:

No, I mean that on Polymarket, you can transfer crypto assets very quickly, and funds are almost instantly available. But on Kalshi, you need to transfer funds in advance, and the funds are "stuck" there. This is different from the arbitrage environment in modern crypto exchanges. In the crypto market, you can transfer USDC or USDT between different exchanges at any time.

Laura Shin:

Got it. The next strategy you refer to as endgame sweep, or late-stage arbitrage. How does this work?

Markus Thielen:

Yes, this strategy is somewhat similar to the time decay we just discussed. I already explained this with the Bitcoin ETF example. When we are very close to the expiration of a contract, some probabilities are mathematically almost certain, but the spreads in the market have not fully converged yet. The results are basically decided a day or even a few hours before the contract expires. For example, right now, it is almost impossible for the Bitcoin ETF to suddenly see $11 billion in inflows. Mathematically, this is not possible. So we are willing to bet at this time because we can earn 4% return. And a 4% return, when annualized, is actually quite a good return.

Laura Shin:

Yes, I felt a similar sense while reading these strategies. It’s like an intraday trader's mindset: as long as you do enough, the returns will accumulate. So the next strategy you mentioned, time decay capture, is actually in the same category as the previous one, right?

Markus Thielen:

Yes, they are indeed somewhat similar. Generally speaking, time decay capture corresponds to a situation where the market is still overpricing volatility. For example, do you think Bitcoin will outperform gold this year? So far, Bitcoin has basically been flat this year, while gold has risen about 60%. Now there are only three weeks left. If you bet that "Bitcoin will not outperform gold this year," you can still earn about 4% return. This is the time decay capture strategy. From an options pricing perspective, we find that the implied volatility of these trades is priced too high.

Laura Shin:

Next, there is a strategy you refer to as maker spread harvesting. Can you explain what this means and how you use this strategy?

Markus Thielen:

Of course. Maker spread harvesting is more of a strategy aimed at professional traders. This strategy usually presents opportunities when the market is very volatile but liquidity is relatively low. In this case, some traders will buy directly using market orders instead of limit orders. The market price will move rapidly as a result. We have seen this happen: retail traders buy directly at market prices, and at that moment, the bid-ask spreads have widened because the market is volatile. This provides arbitrage space for market makers.

However, this type of strategy is more suitable for professional traders with mature trading systems. It is similar to the strategies used by market makers in crypto exchanges and traditional financial institutions. This strategy only works effectively in high volatility, low liquidity environments, when the volume is too large relative to liquidity.

Laura Shin:

It sounds like this strategy somewhat relies on news events triggering emotional responses from market participants. As a market maker, how do you prepare for such situations in advance? Do you place limit orders ahead of time, waiting for market fluctuations to occur?

Markus Thielen:

Generally speaking, market makers will continuously place limit orders in the market. When prices start to move, the trading system can automatically withdraw these orders from the market. We often see this in the crypto market. For example, there are a lot of liquidation stop-loss orders above a certain key price level, and when the price approaches those levels, those orders suddenly disappear.

Professional traders also do similar things. So ultimately, it depends on how your trading engine is set up. The key is whether you are trading against those orders that are already posted in the market or against those market orders that "must be executed immediately." Many times, some traders will place a very low limit buy order in the market and then step away from their computer or phone. When the market suddenly fluctuates violently, those orders may get filled. These strategies are not unfamiliar to professional traders, but not everyone can execute them.

Laura Shin:

The next strategy you refer to as probability compression play. What does this specifically refer to?

Markus Thielen:

A good example is after the October FOMC meeting. At that time, the market saw a huge change in the probability of whether there would be a rate cut in December. Before that, the market almost considered a December rate cut a done deal, with probabilities around 80% to 90%. But after the October FOMC meeting, that probability suddenly dropped to 30%. Then, New York Fed President Williams came out and stated that a rate cut is still the baseline scenario, and the probability jumped back from 30% to 80%.

So you can see that probabilities can fluctuate dramatically in a very short time. This strategy is not completely risk-free because you need to make a judgment to some extent. But if the probability has been compressed to a very low level, like 30%, and you believe the Fed at least wants the market to maintain a 50/50 pricing structure before making a decision, then buying into that probability makes sense.

The same logic applies to the question of who will be the next Fed chair. Everyone knows that Trump usually leaves decisions until the last minute. If a candidate's probability is priced in at 80% or even higher, then you can choose to trade in the opposite direction, waiting for the probability to fall back. This is probability trading conducted when there is still a considerable amount of time left, but the market pricing has not yet reflected the real uncertainty.

Laura Shin:

You also mentioned earlier that you deliberately avoid so-called "long shots," which are extremely low-probability trades. Can you explain in more detail why?

Markus Thielen:

Yes. Low-probability trades typically refer to contracts priced below 10 cents, meaning the market believes the probability of occurrence is below 10%. Research shows that 60% of lost funds occur in these low-probability trades. The psychology behind these trades is very similar to buying a lottery ticket: you spend 5 cents to buy a contract, and if the event occurs, you can get $1, which seems like a very enticing return.

But in the long run, this strategy is almost destined to lose money. Just like studies on U.S. lottery purchases show: if you analyze lottery buying behavior by postal code, you will find it is highly correlated with income levels. This psychology is the same in prediction markets. Many people are attracted by the possibility of "excess returns," but the real probability of these trades is extremely low.

Professional traders do not do this. They place their funds on high-probability trades that gradually converge to certainty. That is where the real money is made.

Laura Shin:

This naturally leads us to the next category of strategies. I find this category interesting because from my perspective, it resembles "copy trading." One of them you refer to as liquidity imbalance trading, or following whale flow. Can you explain what this means and how ordinary people can utilize this strategy?

Markus Thielen:

Of course. These platforms often claim to provide what is called "the wisdom of the crowd," which aggregates information through crowdsourcing. But in my view, a more accurate statement would be: "the wisdom within the crowd." Because among these participants, there will definitely be some large players. They may have better information and can place larger positions. If they are very confident about a certain outcome, they will place a large bet. This is indeed somewhat like copy trading, but the premise is that you need to identify truly excellent traders.

We have seen similar situations in the crypto market, where some high-performing wallets are tracked and analyzed by various tools. In prediction markets, similar things can be done. For example: who has a high historical win rate? Who consistently places large bets in a specific area and performs steadily? Who seems to have some advantage? Following the flow of these individuals' funds is a viable strategy. This is not blindly copying a single trade but identifying those participants who have long-term advantages in a particular category.

Laura Shin:

Is it equally difficult to identify these accounts across different platforms? For example, is it easier on Polymarket because it is more crypto-focused?

Markus Thielen:

Yes, that's absolutely correct. It can be done on Polymarket. On Kalshi, I believe it is currently not possible.

Laura Shin:

Okay, the next strategy you refer to as price sensitivity screening. Can you explain what this means and how you utilize it?

Markus Thielen:

Here we return to the concept of one-touch barrier options. Taking Bitcoin as an example. Suppose there is a contract: will Bitcoin touch $100,000 before December this year? This is essentially a one-touch barrier option. We can compare this probability with the implied volatility in traditional options markets like Deribit or IBIT. In Deribit, whether the option touches a certain price at expiration is different from "as long as it touches once, it counts as a win" in prediction markets. Therefore, the implied volatility in prediction markets is inherently higher.

But the question is: how much higher is reasonable?

We can compare the option surfaces of different markets to see if the probabilities given by prediction markets are significantly higher than those in institutional options markets. When we published this trading idea, the implied probability on Polymarket was about 60%, believing Bitcoin would touch $100,000. But in the traditional options market, the implied probability was only 10% to 15%. Even considering that the one-touch structure requires higher volatility, the gap between the two is still very large.

This indicates that the pricing in prediction markets clearly reflects retail optimism. In this case, you can sell that probability in the prediction market while hedging in the options market. This is a form of cross-market probability arbitrage. So far, this strategy seems to be running quite well.

Laura Shin:

The next few strategies seem somewhat similar. One of them you refer to as conditional hedging. Can you explain how this works?

Markus Thielen:

Yes, this is more related to macro event risks. This is actually one of the core arguments for Kalshi when it obtained CFTC approval in 2024: prediction markets can be used to hedge real-world outcomes. For example, if you want to hedge whether oil prices will fall below a certain level, it is not easy to do so in traditional futures markets because of complex settlement rules and high delivery risks. In prediction markets, such issues can be simplified into a "yes/no" event. This is very similar to the structure of certain insurance contracts.

Laura Shin:

The final strategy is event calendar positioning. How is this different from the conditional hedging strategy?

Markus Thielen:

Event calendar positioning is more centered around known time points. For example: Federal Reserve meetings, elections, major political events. Conditional hedging usually means you are exposed to some risk in the real economy, and prediction markets are a hedging tool. The event calendar strategy is more like trading around the event itself. But I don't want to complicate things too much. From the outcome perspective, the strategies that truly make big money are still concentrated in endgame sweeps and time decay captures. Not those "moonshot" extreme bets.

Laura Shin:

Thank you very much for explaining all these strategies. I think this really shows how detailed prediction markets can be analyzed. Many people's first reaction to prediction markets is, "You are guessing what will happen in the future." But in reality, you don't need to have any information about the events themselves; you just need to judge whether the probabilities of different outcomes are reasonably priced.

Laura Shin:

Before we wrap up, is there anything important about prediction markets that I haven't asked but you think the audience should know?

Markus Thielen:

I think we have covered a lot. The key is: do you view prediction markets as a form of entertainment or as a probability market? If you are just in it for entertainment, then like any entertainment expense, you are paying an entry fee and shouldn't complain about the final outcome. But if you look at it from a probability perspective, you will find that more and more professional traders are entering this market, pricing these events using probability models.

This is also the core message we want to convey to our subscribers: prediction markets are a probability market, not an opinion market. Making money does not come from 100% certain outcomes, but from the higher probability side. Ultimately, the core remains that saying: it is not the wisdom of the crowd, but the wisdom within the crowd. This is a theme that will continue to grow until 2026.

Airdrops could be a very interesting catalyst, and as trading activity increases, prediction markets will become increasingly important. That is why we see companies like Robinhood starting to enter this space—everyone wants a piece of this pie. We want to get ahead of this trend, which is why we study these trades and explain how to make money in prediction markets.

Laura Shin:

Markus, it has been a pleasure chatting with you today. Thank you very much for joining Unchained.

Markus Thielen:

Thank you for having me.