An honest look at Polymarket as a business — volume, risks, and why we still build here

If you're building trading bots for a platform, you should understand that platform's business trajectory. Polymarket has had a remarkable run — and a few legitimate risk factors that anyone building serious tooling on top of it should understand. This is my honest read.

The volume story

Polymarket's volume growth through 2024 was extraordinary. The 2024 US election cycle drove billions in monthly volume, briefly making Polymarket the most liquid prediction market in the world by a significant margin. Post-election, volume dropped sharply — which was entirely expected. Election-cycle spikes are structural.

The more interesting question is the floor. Non-election volume on crypto and sports markets has stabilized at a level that supports active trading infrastructure. Daily volume regularly sits in the hundreds of millions of dollars across all markets, with BTC and ETH binary markets among the most consistently liquid individual books.

For a bot builder, liquidity floor matters more than peak volume. The floor is high enough to run scalping strategies at meaningful sizes without moving the market.

Oracle disputes and settlement risk

Polymarket resolves markets against the UMA optimistic oracle. The resolution mechanism works like this: a proposer submits an outcome, a dispute window opens, and anyone can challenge with a bond. If challenged, UMA token holders vote to resolve.

In practice, most markets resolve cleanly. The edge cases are:

Ambiguous resolution criteria. Market descriptions that don't precisely define the resolving event create dispute risk. If you're trading a market and the description language is fuzzy, that's a risk factor you should price in.

Late oracle commits. For time-sensitive crypto binaries, the Chainlink oracle commit timing relative to expiry matters (covered in depth in the oracle latency post). Settlement disagreements between the oracle price and what traders "saw" on exchange are the most common source of confusion, but the oracle is always correct by design.

Liquidity pool disputes. A small number of markets have seen genuine disputes where the outcome was genuinely ambiguous — "did X happen before Y?" type questions. These are rare but real. For algorithmic traders, the mitigation is to stick to markets with precise, objectively verifiable resolution criteria: crypto price levels, weather measurements, sports scores.

Regulatory risk

The US restriction is the biggest elephant in the room. Polymarket does not allow US-based traders. In practice, enforcement is VPN-bypassable and the platform has not taken legal action against individual US users — but the regulatory exposure is real.

The core risk is not that Polymarket gets shut down outright. It's that US regulatory action forces a restructuring that affects platform economics or liquidity. If a significant portion of current volume comes from US-adjacent participants using VPNs, regulatory tightening could reduce that volume meaningfully.

The secondary risk is precedent. As Kalshi and other regulated US prediction market operators grow, they will push for clearer enforcement against unregistered competitors. This is not imminent, but it's a multi-year headwind.

Kalshi's growth

Kalshi is the most relevant competitor — a CFTC-regulated US prediction market with genuine legal standing. Its growth has been real but from a much lower base. As of mid-2025, Kalshi's BTC binary volume is a fraction of Polymarket's equivalent markets.

What Kalshi has that Polymarket doesn't: US user access, regulatory clarity, and growing institutional interest. What it lacks: depth, variety of markets, and the crypto-native user base that creates the retail-vs-sophisticated-money dynamic that scalpers rely on.

My read: Kalshi and Polymarket serve different primary audiences right now. Kalshi is better for regulated US participants who care about legal certainty. Polymarket is better for serious traders who want depth and variety. The two may converge, but the timeline is years, not months.

predtools will add Kalshi support when the market depth justifies it. The APIs are different enough to require separate bot implementations, but the strategy logic is largely transferable.

Why we still build here

Three reasons:

Depth and variety. Polymarket has the deepest books and the widest variety of markets. For strategy development, variety is essential. You can test a scalping approach across dozens of different market structures. Kalshi's current catalog is narrow by comparison.

Retail-sophisticated dynamic. The median Polymarket trader is not running an alpha model. They're placing opinion bets. This creates persistent mispricings that a systematic strategy can exploit. As the platform matures and attracts more sophisticated participants, this dynamic will compress — but it hasn't yet.

Infrastructure maturity. The CLOB API is stable, the Gamma API covers market discovery and resolution cleanly, and the Chainlink oracle is well-documented. Building production-grade bots on Polymarket is feasible today. The API surface is reasonable enough that you can go from zero to live trading in a few days.

The honest summary: Polymarket is the best prediction market venue to build on right now, with real regulatory and competitive risks on a 1–3 year horizon. Anyone building serious tooling here should have a Kalshi migration path in mind. The core trading logic — signal generation, fee management, resolution handling — transfers cleanly. The API layer doesn't.

predtools is building with this in mind. The strategy modules are platform-agnostic where possible; the exchange-specific connectors are the thin layer that will need to be replaced.