5 Signals That a Prediction Market Is Mispriced (And How to Act on Them)

You're staring at a Kalshi contract priced at 64% for a March Fed rate cut. Polymarket has the same outcome at 71%. They can't both be right.

One of them is mispriced. The question is which one, by how much, and how fast it corrects.

Most traders see this kind of gap and guess. Experienced traders recognize it as one of five identifiable signals that a market is trading away from fair value — signals that show up repeatedly, have been documented in academic research, and can be acted on systematically. Here's what to look for.

Signal 1: Cross-Platform Price Divergence

When Kalshi and Polymarket price the same outcome differently, one of three things is true: institutional flow is asymmetric, fee structures are distorting the comparison, or one platform's traders simply haven't caught up to new information yet.

That last case is the most actionable.

In early 2026, the March Fed rate cut contract was priced at 64% on Kalshi and 71% on Polymarket simultaneously — a 7-point spread. That's not random noise. Polymarket skews toward macro-sophisticated traders who react faster to Fed commentary; Kalshi's user base is broader and slower to reprice on central bank signals. The divergence told you something: the macro-aware money had moved, and the retail side hadn't followed.

Signal 2: The Favorite-Longshot Bias

This one has decades of data behind it. A 2025 academic paper analyzing over 300,000 Kalshi contracts (Bürgi, Deng, and Whelan) found a clear pattern: contracts priced below 15 cents win far less often than their price implies. Contracts priced above 80 cents win more often than their price implies.

Translation: cheap "Yes" contracts are consistently overpriced. Expensive "Yes" contracts are consistently underpriced.

Why does this persist? Retail traders love lottery-ticket bets. The psychological appeal of turning 8 cents into a dollar is powerful, and it pushes low-probability contracts above their fair value. Meanwhile, high-probability outcomes feel boring and get less attention — which means they're often slightly underpriced relative to their actual win rates.

This isn't a guaranteed edge — it's a base rate adjustment that shifts expected value in your favor over many trades.

Signal 3: Liquidity Desert

A market with $12,000 in open interest is not a reliable price signal. A market with $2.4 million is.

Prediction markets work because informed traders move prices toward fair value. That mechanism requires enough participants, enough capital at stake, and enough back-and-forth between buyers and sellers. When liquidity is thin, a single trader posting an uninformed offer can set the price — and that price tells you nothing.

Thin markets are particularly common in two places: newly created contracts that haven't attracted attention yet, and niche markets that are technically interesting but have a small trader base.

A 6-cent spread on a market trading at 50 cents implies the market maker thinks the true probability could be anywhere between 47% and 53%, which is nearly useless for positioning.

Signal 4: News Velocity With No Price Movement

Information is supposed to flow into prediction markets in real-time. When it doesn't, there's an edge.

The pattern: a news event with clear probabilistic implications (a regulatory announcement, economic data release, geopolitical development) hits the wires, and the relevant contract doesn't move — or moves by only 1–2 points when the information logically implies a 5–10 point shift.

This is most common on breaking news that's covered by one outlet before syndication, on data that requires interpretation before its implications are clear, and on markets with fewer active traders watching that category.

In practice: if you're monitoring a contract on cryptocurrency regulatory action and the EU parliament votes on a relevant measure, the market should reprice within minutes. If it doesn't, the informed traders haven't arrived yet. The window is short — often 10–20 minutes before arbitrageurs close it — but the edge is real.

Signal 5: Maker-Taker Flow Imbalance

This one requires more data but is among the most reliable signals when you have it.

Prediction markets have two types of participants: makers who post offers and set the price, and takers who accept those offers. Research shows that makers — who tend to be better-informed — earn systematically better returns than takers. When makers suddenly flood one side of a market (posting aggressive "Yes" or "No" offers), they're signaling that they believe the current price is wrong.

When a whale consistently posts maker offers at prices 3–5 cents below the current market, they're either providing liquidity cheaply or they know something. Either way, the price is about to move. The question is whether you're positioned before or after.

Why These Five Signals Matter Together

Each of these signals in isolation is useful. Together, they're a scoring framework.

A market with cross-platform divergence and thin liquidity is probably noise — the gap exists because neither side has enough traders to arbitrage it. A market with news velocity lag and maker flow imbalance is a much stronger signal — informed money is already moving, retail hasn't followed.

The traders extracting consistent positive expected value from prediction markets aren't doing it by watching one signal. They're running a composite score across signals, sizing positions to their confidence, and trading only when multiple indicators point the same direction.

That's the methodology. The problem is doing it manually across 80+ markets simultaneously is impossible. By the time you've manually checked cross-platform prices, volume, bid-ask spreads, and news velocity for twenty markets, the edge has closed on the first three.

How Presage Automates This

We built Presage to run this calculation in real-time across Kalshi and Coinbase prediction markets.

Every market gets a live edge score based on cross-venue pricing, news velocity, volume patterns, and sentiment trajectory. When multiple signals fire simultaneously, the score spikes. You see it immediately — no manual scanning, no spreadsheet, no watching 15 tabs at once.

If you're trading prediction markets with information a handful of other traders have, you're competing on luck. If you're trading with a composite signal score showing you exactly where the market price diverges from the data — you're competing on edge.

That's the difference.