A sobering reflection after a valuation of tens of billions: Why is it so difficult for the market to generate high leverage?

Original title: Everyone's Promising 20x Leverage on Prediction Markets. Here's Why It's Hard

Original author: @hyperreal_nick, crypto KOL

Original translation by: Azuma, Odaily Planet Daily

Editor's Note: While compiling a list of new projects emerging during the Solana Breakthrough cycle this week, I noticed the emergence of some prediction markets that emphasize leverage. However, upon closer inspection, the current situation is as follows: leading platforms generally shy away from leverage features, while new platforms that claim to support the feature typically suffer from issues such as low leverage ratios and small pools.

Compared to the booming Perp DEX, the leverage potential of the prediction market seems to remain largely untapped, a stark contrast to the highly risk-averse nature of the cryptocurrency market. To address this, I began researching and came across two excellent analytical articles. One was Kaleb Rasmussen's research report on this issue, "Enabling Leverage on Prediction Markets," which provided a very thorough argument, but its length and extensive mathematical calculations made it difficult to translate. The other was Nick-RZA's "Everyone's Promising 20x Leverage on Prediction Markets. Here's Why It's Hard," which was more concise and accessible, yet effectively addressed the leverage challenge of prediction markets.

The following is the original text by Nick-RZA, translated by Odaily Planet Daily.

Currently, almost everyone wants to add leverage to the prediction market.

Earlier, I wrote an article titled "The Expression Problem"—which concluded that forecasting markets limit the strength of beliefs that capital can express. It turns out that many teams are already trying to address this issue.

Polymarket's valuation has reached $9 billion after a $2 billion investment from the parent company of the New York Stock Exchange platform, and its founder, Shayne Coplan, was featured in 60 Minutes. Kalshi first raised $300 million at a $5 billion valuation, and then completed a new round of financing of $1 billion at a $11 billion valuation.

The race is heating up, with participants vying for the next level of demand—leverage . Currently, at least a dozen projects are attempting to build "leveraged prediction markets," some claiming to achieve 10x, 20x, or even higher leverage. However, when you actually examine the analyses provided by teams seriously addressing this problem (such as HIP-4, Drift's BET, and Kalshi's framework), you'll find their conclusions converging towards a very conservative figure: between 1x and 1.5x.

This is a huge gap, but where exactly did the problem lie?

Prediction Markets vs. Spot/Contract Trading

Let's start with the basics. Prediction markets allow you to bet on whether an event will happen: Will Bitcoin reach $150,000 by the end of the year? Will the 49ers win the Super Bowl? Will it rain in Tokyo tomorrow?

You are buying a "share". If you predict correctly, you will get $1; if you are wrong, you will get nothing. It's that simple.

If you believe BTC will rise to $150,000, and the price of a "YES Share" is $0.40, you can buy 100 shares for $40. If you're right, you'll get your $100 back, making a net profit of $60; if you're wrong, you'll lose $40.

This mechanism brings three completely different characteristics to prediction markets compared to spot trading or perpetual contracts:

First, there is a clear upper limit . The maximum value of a "YES share" (and vice versa for a "NO share") is always $1. If you buy in at $0.90, the maximum upside potential is only 11%. This is unlike buying a single meme coin early on.

Second, the lower limit is true zero . Not a near-zero drop, but zero in the literal sense. Your position won't slowly disappear over time—either your prediction is correct, or it goes to zero.

Third, the outcome is binary , and confirmation of the outcome is usually instantaneous . There is no gradual price discovery process; an election may be undecided one moment and the result announced the next. Correspondingly, the price does not gradually rise from $0.80 to $1, but rather "jumps" directly to that point.

The essence of leverage

The essence of leverage is borrowing money to amplify your bets.

If you have $100 and use 10x leverage, you are effectively controlling a $1000 position—if the price rises by 10%, you don't earn $10, but $100; conversely, if the price falls by 10%, you don't lose $10, but your entire principal. This is what liquidation means—the trading platform will forcibly close your position before your losses exceed your principal to prevent the lender (trading platform or liquidity pool) from bearing the losses.

The reason why leverage works on conventional assets is based on a key premise: the price changes of the asset are continuous.

If you go long BTC at $100,000 with 10x leverage, you'll likely face liquidation around $91,000–$92,000. However, BTC won't instantly jump from $100,000 to $80,000. It will only decline gradually, albeit rapidly, in a linear fashion—99,500 → 99,000 → 98,400… During this process, the liquidation engine will intervene and close your position as needed. You might lose money, but the system is safe.

Predicting the market, however, transcends this premise.

Core issue: Price jump

In the derivatives world, this is known as "jump risk" or "gap risk," while the cryptocurrency community might call it "scam wicks."

Let's use BTC as an example again. Suppose the price doesn't gradually decrease, but jumps directly—100,000 one second, 80,000 the next, with no transaction price in between, no 99,000, no 95,000, and certainly no 91,000 that you could liquidate.

In this situation, the liquidation engine still attempts to close the position at $91,000, but that price simply doesn't exist in the market, and the next available price jumps to $80,000. At this point, your position is no longer just liquidated, but deeply insolvent , and this loss must be borne by someone.

This is precisely the situation that the forecasting market faces.

When election results are announced, game outcomes are settled, or major news breaks, prices don't move slowly and linearly; instead, they jump up and down dramatically. Furthermore, leveraged positions within the system cannot be effectively broken down because there is simply no liquidity involved.

Kaleb Rasmussen of Messari wrote a detailed analysis of this issue (https://messari.io/report/enabling-leverage-on-prediction-markets). His conclusion is that if lenders can correctly price jump risk, their fees (similar to funding fees) should absorb all upside potential gains from leveraged positions. This means that for traders, opening leveraged positions at fair rates offers no profit advantage compared to opening positions without leverage, and also exposes them to greater downside risk.

Therefore, when you see a platform claiming to offer 10x or 20x leverage in the prediction market, there are only two possibilities:

• Either their fees do not accurately reflect the risk (meaning someone is bearing uncompensated risk);

• Either the platform uses some undisclosed mechanism.

Real-world case study: Lessons learned from dYdX

This is not just theoretical; we already have real-world examples.

In October 2024, dYdX launched TRUMPWIN—a leveraged perpetual market that predicts whether Trump will win the election, with up to 20x leverage, and the price oracle comes from Polymarket.

They were not unaware of the risks; in fact, they designed multiple protection mechanisms for the system:

• Market makers can hedge their exposure to dYdX in the Polymarket spot market;

• An insurance fund is established to cover losses in case of unsuccessful liquidation;

If the insurance fund runs out, the losses will be distributed among all profitable traders (although nobody likes it, it's better than the system going bankrupt; an even more brutal version is ADL, which directly liquidates the winning positions).

• The dynamic margin mechanism automatically reduces available leverage as open contracts increase.

This was already quite mature by perpetual contract standards. dYdX even publicly warned of the risks of deleveraging. Then, election night came.

As the results became clearer and Trump's victory became almost certain, the price of "YES Share" on Polymarket jumped from about $0.60 to $1—not gradually, but dramatically, a jump that broke through the system.

The system attempted to liquidate underwater positions, but there was simply not enough liquidity, and the order book was very thin; market makers who should have been hedging on Polymarket were also unable to adjust their positions in time; insurance funds were also wiped out... When positions could not be liquidated smoothly, random deleveraging was initiated—the system forcibly closed some positions, regardless of whether the other party had sufficient collateral.

According to analysis by John Wang, head of crypto at Kalshi: "Hedge delays, extreme slippage, and liquidity evaporation caused losses for traders who should have been able to execute trades ." Some traders who should have been safe—with proper positions and sufficient collateral—still suffered losses.

This is not a worthless DEX without risk control, but one of the world's largest decentralized derivatives trading platforms, with multi-layered protection mechanisms and clear warnings issued in advance.

Even so, the system still experienced some failures in real market conditions.

Industry solutions

Regarding the leverage issue in prediction markets, the industry has split into three camps, and this split itself reveals each team's attitude toward risk.

Camp 1: Limit Leverage

Some teams, after seeing the mathematical reality clearly, chose the most honest answer—providing almost no leverage.

HyperliquidX's HIP-4 proposal sets the leverage limit at 1x—not because it's technically impossible, but because it's considered the only safe level for binary outcomes.

• DriftProtocol's BET products require 100% margin, meaning full collateral and no lending.

John Wang, head of crypto at Kalshi, also suggests in his framework that the security leverage is approximately 1–1.5 times without additional protection mechanisms.

Camp Two: Utilizing Engineering to Counter Risk

Another group of teams attempted to build systems that were complex enough to manage risk.

• D8X dynamically adjusts leverage, fees, and slippage based on market conditions—the closer to settlement or in extreme probabilities, the stricter the restrictions.

dYdX built the protection mechanism we just saw fail on election night and is still iterating on it;

PredictEX's solution is to increase fees and reduce maximum leverage when the risk of price jumps rises, and then loosen it again when the market stabilizes. As its founder Ben put it bluntly: " If we directly apply the perpetual contract model, market makers will be completely wiped out in the one second when the probability jumps from 10% to 99%."

These engineering teams did not claim to have solved the problem; they were simply trying to manage risk in real time.

Camp Three: Go in first, fill in later.

Some teams choose to launch quickly, directly claiming leverage of 10x, 20x, or even higher, without publicly disclosing how they handle jump risks. Perhaps they have an elegant, yet-to-be-revealed solution, or perhaps they want to learn in a production environment .

The crypto industry has a long tradition of "getting it running first and then hardening it," and the market will ultimately test which approach will stand the test.

What will happen in the future?

We are facing a problem of an extremely open design space , which is precisely what makes it so interesting.

Kaleb Rasmussen's Messari report not only diagnosed the problem but also suggested some possible directions:

• Do not price the risk of the entire position all at once; instead, charge rolling fees based on changes in conditions.

• Design an auction mechanism to return value to liquidity providers when prices jump;

• Build a system that enables market makers to continue to profit without being overwhelmed by information advantages.

However, these solutions are essentially improvements on the existing architecture.

Deepanshu of EthosX offers a more fundamental reflection. He previously researched and built clearing infrastructures such as LCH, CME, and Eurex in JPMorgan Chase's global clearing business. In his view, attempting to leverage prediction markets using perpetual contract models is fundamentally addressing the wrong problem.

Prediction markets are not perpetual contracts, but rather extreme exotic options —more complex than products typically handled in traditional finance. Exotic options are not traded on perpetual trading platforms; they are generally settled through clearing infrastructure specifically designed for their risk. Such infrastructure should be able to:

• Provide traders with a time window to respond to margin calls;

• A mechanism to allow other traders to take over positions before they get out of control;

• Multi-tiered insurance funds allow participants to clearly accept the socialization of tail risk.

These are not new—clearinghouses have been managing jump risks for decades. The real challenge is—how to do all this on-chain, transparently, and at the speed required to predict the market.

Dynamic costs and leverage decay are just the beginning. Ultimately, the teams that truly solve the problem will likely not only create a better sustainable engine, but also build a "clearinghouse-level" system. The infrastructure layer remains unresolved, while market demand is already very clear.