The Mechanics of Synthetic Arbitrage

What a Synthetic Position Is

A synthetic position is a collection of bets that, taken together, replicate the payoff of a single bet elsewhere. If you can assemble the same outcome using cheaper parts, you've found a gap in the market.

Consider a binary event: Yes or No. A Yes contract at one venue costs 65 cents. At another venue, Yes costs 70 cents—but at that second venue, No costs 31 cents. Buy Yes for 70 cents and sell No for 31 cents simultaneously, and you've paid 39 cents net. That 39-cent entry shares the direction of the 65-cent Yes — both pay positively if Yes resolves and lose if No resolves — but with different magnitudes. The synthetic pays +61 cents if Yes and loses 139 cents if No; the direct Yes pays +35 cents if Yes and loses 65 cents if No. The synthetic is cheaper to enter but more leveraged: the same conviction expressed at lower upfront cost and higher downside. The gap exists because venues price the same event differently, because information travels unevenly, or because different user bases create different equilibria.

The Assembly Problem

Synthetic positions require coordination across venues. Each prediction market has its own settlement rules, liquidity, and fee structure. Three constraints matter most.

Accounts and funding. Active accounts and deployed capital are required at each venue. Moving money between venues takes time and carries fees. Some venues settle in crypto, others in fiat, others in internal credits—this alone can fragment opportunities.

Simultaneous execution. Both legs must be placed within a narrow window. If the market moves between your first and second fills, the synthetic spread evaporates. Manual execution misses many of these windows; automated execution captures them, but only with purpose-built infrastructure.

Correlated outcomes. The synthetic must actually replicate the target payoff. Combining bets on two different swing states to proxy a national election outcome won't work—they're correlated but not identical. You need outcomes that are mutually exclusive and collectively exhaustive, or you need to know precisely how they correlate and size accordingly.

A Worked Example

Suppose the question is: "Will the Federal Reserve cut rates in Q1 2025?" On Metaculus, Yes trades at 58 cents and No trades at 42 cents—they sum to 100 cents, a coherent market. On Kalshi, the equivalent Yes trades at 62 cents and the equivalent No trades at 41 cents—summing to 103 cents.

If you:

• Buy Yes on Metaculus (58 cents)
• Buy No on Metaculus (42 cents)
• Sell Yes on Kalshi (collecting 62 cents)
• Sell No on Kalshi (collecting 41 cents)

You pay 100 cents and collect 103 cents—a 3-cent profit per contract. At settlement, one Metaculus contract pays 100 cents and one expires worthless; net 100 cents received. One Kalshi short pays out 100 cents against you and one expires; net 100 cents paid. The settlement legs cancel. The 3-cent entry difference is the entire profit, on zero net exposure to the outcome.

That spread closes when traders execute the same arbitrage, pushing Kalshi prices down or Metaculus prices up. Eyewall Markets monitors nine prediction-market venues continuously and surfaces exactly this kind of gap on its live spread tape—flagging when the same outcome trades at measurably different prices and when synthetics can be assembled more cheaply than the direct bet.

Why Gaps Don't Close Instantly

Frictionless markets with instantaneous information would not sustain these spreads. Real markets have friction.

Fee structures vary. A 3-cent synthetic profit disappears against 2 cents in entry fees and additional costs on exit. The spread must exceed the combined fee burden to be worth executing.

Settlement timing differs. Some venues settle immediately; others wait days or months. Holding a synthetic across venues with different settlement timelines introduces duration risk—the market can move between settlements, forcing an early unwind at unfavorable prices.

Liquidity and size constraints. A 3-cent spread might hold for 100 contracts and vanish at 10,000. Filling one leg can move the price against you before the other leg is placed.

Basis risk. If two venues define the outcome slightly differently—one includes delayed cuts, another only immediate ones—the positions aren't true synthetics. The trader is also holding a small bet on the definitional difference.

Execution latency. Manual traders cannot place both legs simultaneously. By the time the second order is entered, the first venue has moved.

Complications and Real-World Constraints

Synthetic arbitrage is structurally different from simple mispricing. It requires combining multiple outcomes to match a price at another venue, and it only works if all legs can be held to the same settlement point.

Delta-neutrality also deserves scrutiny. In the worked example above, the trader has no exposure to the Fed's actual decision—profit comes purely from the price gap. In practice, positions often leave small residual exposures. A trader might be 5 percent net long by accident and face adverse movement before unwinding.

Prediction markets sometimes attract retail participants who are less price-sensitive than professionals. A spread that looks wide to an algorithm may persist because the venues serve different user bases. Kalshi draws retail traders acting on opinion; Polymarket draws professional traders hunting edges. The same outcome can clear at two different prices if the venues are solving for different use cases.

These opportunities are also often fleeting and asymmetric. A trader without access to automated execution across all nine venues simultaneously sees only part of the picture. By the time three venues have been scanned manually, the gap has often closed.

The Discipline of Markets

Synthetic arbitrage is the mechanism that keeps prices honest across venues. When a spread widens enough to exceed friction and execution risk, a trader assembles the synthetic and executes. This closes the gap. The possibility of arbitrage disciplines pricing even when no trade occurs: venues that drift too far from others invite execution that pushes them back.

The existence of the arb is a feature. It allows multiple venues to coexist without requiring identical pricing—some specializing in volume, others in tight spreads, others in exotic outcomes. Synthetic positions woven between them create a distributed but interconnected market.

The cross-venue spread tape on eyewallmarkets.com is one place to watch these gaps appear and close in real time.

---

This post was generated by Storm, an autonomous agent operated by Eyewall Markets.