How to earn risk-free income on Limitless: a single strategy has generated $28.99M

Written by: @ccjing_eth

Prediction markets: the ignored arbitrage gold mine

Most people treat prediction markets like a casino—betting on elections and guessing crypto prices. But if you change your perspective and see them as markets with pricing efficiency far lower than traditional finance, you’ll find a completely different opportunity.

The core logic is simple: YES shares represent “the event happens,” and NO shares represent “the event does not happen.” In a perfect market, YES + NO = $1.00. But in reality, prices often deviate from that equilibrium point.

That’s where the arbitrage space comes from.

Limitless is one of the platforms currently best suited to execute these strategies. A decentralized prediction market built on the Base chain, it uses a CLOB (central limit order book) instead of an AMM, supports limit orders and market orders, settles instantly on-chain, and has no KYC. Cumulative trading volume has already surpassed 1.3B, raised $13M+, and investors include Coinbase Ventures, DCG, F-Prime, Arrington Capital, and Arthur Hayes’s Maelstrom.

It’s not a DEX, not an exchange, not a wallet—it’s a price discovery engine, and the pricing efficiency of this engine is precisely where arbitrageurs’ profits come from.

Strategy 1: NegRisk built-in arbitrage

This is a structural arbitrage opportunity unique to prediction markets—something that doesn’t exist in traditional finance.

NegRisk (Negative Risk) is a pricing framework for multi-outcome markets. In a three-person market for “who wins an election,” the sum of the prices of the three YES shares should, in theory, equal $1.00. But in actual trading, because each outcome’s order book operates independently in terms of supply and demand, the sum of prices often deviates from that equilibrium.

Principle

When Σ(YES price) < $1.00:

Buy the YES shares for all outcomes

At settlement, at least one YES will be assigned to $1.00

Profit = $1.00 - Σ(YES cost)

When Σ(YES price) > $1.00:

Use NegRisk’s Convert operation to convert the undervalued NO shares into YES shares for other outcomes

Sell the overvalued outcome and lock in the spread

Concrete example walkthrough

Assume a BTC price prediction market has three outcomes:

BTC > 70K: YES 0.45 BTC > 70K: is 0.45

BTC 65K-70K: YES $0.30 BTC 65K-70K: is, 0.30 USD

BTC < 65K: YES 0.20 BTC < 65K: is 0.20

Σ(YES) = 0.95, deviating from 1.00 by $0.05. Σ（YES） = 0.95, deviating from 1.00 by $0.05.

Execution path:

Buy the YES shares for all three outcomes (total cost = 0.45 + 0.30 + 0.20 = 0.95)

At settlement, one YES must be assigned to $1.00

Profit = 1.00 - 0.95 = $0.05 (5.26% risk-free return)

NegRisk’s three main operations

Split: Split USDC into a set of YES + NO shares ($1 USDC → 1 YES + 1 NO)

Merge: Merge a set of YES + NO shares back into USDC (1 YES + 1 NO → $1 USDC) Merge: Merge a set of YES + NO shares back into USDC (1 YES + 1 NO → $1 USDC)

Convert: Convert the NO shares of one outcome into the YES shares of other outcomes (NegRisk-specific)

Convert is the key weapon for NegRisk arbitrage. When you find that the NO shares of a certain outcome are undervalued, you can buy NO → Convert it into YES shares for other outcomes → sell the overvalued YES. Everything is executed on-chain end-to-end, with extremely high Gas efficiency.

“Other” trap “Other” trap

In NegRisk multi-outcome markets, there is typically an implied “Other” outcome. This means that even if you buy the YES shares for all visible outcomes, if the actual result lands in “Other,” all your YES shares will be zeroed out. Always check whether the market includes an implied outcome and incorporate it into your pricing calculations.

Strategy 2: Cross-platform arbitrage

The same event, different platforms, different prices.

This is the most intuitive form of arbitrage. Polymarket, Limitless, Kalshi, Opinion, and Myriad often quote different prices for the same event.

Why does the price spread exist?

Different user bases: Polymarket leans toward U.S. politics, Limitless leans toward crypto prices, Kalshi leans toward traditional events

Liquidity differences: Large orders have different slippage on different platforms

Information propagation delay: News events react at different speeds across platforms

Settlement mechanism differences: Time differences between on-chain vs off-chain settlement

Tool ecosystem

Polyar.io scans price spreads across the 5 major platforms in real time

Premarket aggregates the best order routing pm.wiki

Signal Desk aggregates arbitrage signals futarchists.xyz Signal Desk aggregates arbitrage signals futarchists.xyz

okaybet.app helps identify cross-platform price spreads okaybet.app

okaybet.app helps identify cross-platform price spreads okaybet.app

Workflow

Find markets with a price spread > 2% on Polyar.io or Premarket

Buy YES on the lower-priced platform (or sell NO)

Sell YES on the higher-priced platform (or buy NO)

After both legs are filled, lock in the profit

Key risk: The biggest enemy of cross-platform arbitrage is execution delay. After you get filled on one platform, the price on the other platform may have already moved. Solutions:

Prioritize limit orders instead of market orders

Choose markets with sufficient liquidity (bid-ask spread < 1%)

Consider using Premarket’s aggregated routing feature

Strategy 3: LP spread capture

Not a market maker, but you earn the market maker’s money.

Limitless’s LP Rewards program is a unique design: you don’t need to provide liquidity pools (like Uniswap). Instead, you earn rewards by placing limit orders.

Mechanism

Place limit orders between the bid and ask prices

When your order gets filled, you simultaneously receive: trading profit + LP rewards

Rewards are distributed daily in USDC

Design inspiration comes from dYdX’s maker-taker model and Polymarket’s LP incentives

Why is this “arbitrage”?

Strictly speaking, it’s spread capture rather than risk-free arbitrage, but it provides a steady stream of returns:

Long spread: place buy orders below the Bid and sell orders above the Ask

When the price fluctuates, your orders get filled, and you earn the bid-ask spread

Add the LP rewards, and your overall return rate can be significantly higher than just spread income

Advantages

No capital lockup risk: limit orders can be canceled anytime

No impermanent loss: not an AMM pool, so no IL

Predictable returns: based on trading volume and order quality

USDC settlement: rewards are paid out directly in a stablecoin

Strategy 4: Share conversion arbitrage

Deterministic profit when YES + NO ≠ $1.00.

In a single binary market, the sum of the YES and NO prices should always equal $1.00. But because the order books operate independently, temporary deviations can occur.

Execution

When YES + NO < $1.00: When yes + no < $1.00:

Buy both YES and NO at the same time (total cost < $1.00)

Execute the Merge operation: YES + NO → $1.00 USDC

Profit = $1.00 - total cost

When YES + NO > $1.00: When yes + no > $1.00:

Execute the Split operation: $1.00 USDC → 1 YES + 1 NO

Simultaneously sell YES and NO (total revenue > $1.00)

Profit = total revenue - $1.00

Key advantages

Single-market execution: no need to cross platforms, eliminating execution risk

Instant settlement: Split/Merge on-chain atomic operations

Extremely low Gas costs: Base chain on-chain execution cost < $0.01

Automatable: suitable for bots to monitor and execute 24/7

Field data: how much money can arbitrage make?

The research team at IMDEA Networks analyzed 86 million trades on Polymarket and recorded complete data on systematic arbitrage extraction:

Total extracted amount: $29M+ (only NegRisk rebalancing arbitrage)

Markets covered: 17,218 market conditions

Capital efficiency: NegRisk arbitrage’s capital efficiency is 29 times that of binary arbitrage

Opportunity frequency: in multi-outcome markets, arbitrage opportunities exist in 8.6% of the time windows, but contribute 73% of total profit

There’s already an open-source NegRisk rebalancing arbitrage bot on GitHub. Historical records show that across 662 markets it extracted $28.99M in total.

These data show one thing: arbitrage isn’t theoretical—it’s deterministic profit happening every day. The only question is who can discover it first and execute it fastest.

Risk warning

Execution risk: Cross-platform arbitrage has a leg-risk issue—after one leg is filled, the other leg may experience slippage

Gas costs: While Base chain Gas is very low, high-frequency operations still need to account for the cost-benefit ratio

Smart contract risk: Limitless has undergone multiple rounds of audits, but on-chain interactions always carry contract risk

Liquidity risk: In niche markets, the order book may be thin, and slippage for large orders can be significant

“Other” result risk: NegRisk markets may include an implied outcome—make sure it’s included in your calculations

Summary

The essence of prediction market arbitrage is turning market inefficiency into deterministic returns. The four strategies each apply to different scenarios:

The combination of Limitless’s CLOB order book + NegRisk framework + Base chain low Gas + no KYC makes it one of the best platforms to execute these strategies today.

Market inefficiency won’t exist forever. As more participants enter, the arbitrage space will gradually shrink. But before that happens, people who understand these strategies are quietly earning the risk-free returns the market offers.

Not gambling. It’s calculating.