The Super Bowl of Prediction Markets: Kalshi vs Polymarket

Main Paper

The Super Bowl of Prediction Markets Kalshi vs Polymarket - Jay (Pantera), Rongxin (WPRC)                                        17 Feb 2026

Studying NFL game markets from the 2025 season, we observe that Kalshi, with more notional volume and open interests, exhibit faster in-game pricing reactions to shock events based on the settled trades compared to Polymarket, while the latter demonstrates higher implied liquidity depth. This reveals the structural differences between the two platforms, likely due to the “3s speed bump” design on Polymarket together with other factors.

The divergence between Kalshi and Polymarket proves that "more volume" does not automatically mean "better liquidity." While Kalshi handles more trades, and responds to news 7 seconds faster (median from the data sample), its prices are significantly more volatile and sensitive to individual shocks. For traders, this means choosing between Kalshi for execution speed during live events or Polymarket for price stability and lower slippage when moving large positions.

Market design dictates the trade-off between speed and depth. Kalshi’s instant execution model prioritizes price discovery, leading to a 7-second lead in repricing but higher volatility. Conversely, Polymarket’s 3-second "speed bump" on live sports orders protects liquidity providers from adverse selection, resulting in a market that is 3.5x deeper and more resilient to trading pressure, despite having lower overall volume.

1. Platform Design Trade-offs: The key design decisions by the platforms that impacted these results. Kalshi (immediacy-first): (i) immediate off-chain matching + settlement (ii) no explicit delay → favors faster repricing, but exposes resting quotes to fast/informed flow → LPs widen/pull → higher volatility / thinner effective depth during shocks. Polymarket: (i) off-chain order book + on-chain settlement (Polygon PoS ~2s block time) (ii) ~3s delay on marketable orders in live sports → reduces adverse selection risk → supports deeper “near-price” liquidity (lower price impact), but feels laggier for latency-sensitive takers. Takeaway: (i) small execution rules can create fast-but-thin vs slow-but-deep markets (ii) speed and depth don’t have to move together.

2. Measurements: Dataset: (i) 282 NFL games on both platforms (9/4/2025–1/18/2026) (ii) trade-level price/size/timestamp from chain/API (iii) window: ~30 min pre-kickoff → end of game (highest info intensity) (iv) Kalshi lacks trader IDs → no trader-level profiling. 2. Scale stats (NFL sample): (i) total notional: Polymarket ~$359M vs Kalshi ~$1.3B (ii) avg trades/game: 2,718 vs 19,024 (iii) avg bet size: $469.88 vs $224.83 (iv) median bet size: $18.00 vs $23.97. Pricing-speed metric: (i) align price series to timestamped on-field events (ii) compare onset of repricing after large probability moves (iii) result: Kalshi leads ~80% of large moves; median lead ~7s (robust across swing sizes). 4. Volatility check: (i) realized in-game volatility higher on Kalshi across underdogs / evenly matched / heavy favorites (ii) surprise outcome explains limited variance (R² ≈ 0.118). 5. Liquidity inference: (i) no full order book snapshots → infer depth from executed trades (ii) compute signed $ flow (buy +, sell −) (iii) measure 60s post-trade price change (iv) per game regress price change on signed flow → estimate Kyle λ (v) define LD = 1/λ and compare distributions. 6. Liquidity result + trend: (i) median LD ≈ 10^6.96 (Polymarket) vs 10^6.42 (Kalshi) → ~3–4× more flow needed to move Polymarket over 60s (ii) Polymarket shows fewer low-depth outliers (iii) depth distribution stable over season (no meaningful upward drift).

(i) Ally Zach (@0xallyzach) (Pantera Capital): Prev Enterprise team at Messari, worked as a Structural Engineer at Magnusson Klemencic Associates. (ii) Danning Sui (danning.eth) (Pantera Capital): Host of the Indexed Podcast, prev Data Scientist at Flashbots, prev at Consensys.

Pantera Research Lab, “The Super Bowl of Prediction Markets: Kalshi and Polymarket’s Battle for Price vs Liquidity,” Feb 5, 2026. https://panteraresearchlab.xyz/research/the-super-bowl-of-prediction-markets-kalshi-and-polymarkets-battle-for-price-vs-liquidity/