Community / Research Note
Polymarket Market Analysis: Strait of Hormuz traffic returns to normal by end of April?
This post uses the sample export for this market to build a simple regime model on top of the trade history. It shows how raw trades can be turned into a structured market read, and how regime labels help organize the rest of the analysis.
Window
2026-04-02 22:59:09 → 2026-05-02 22:43:48 (UTC)
Trades
184,380
Notional
28.1M
Summary
- The market spends long stretches in calm drift, then snaps into short high-volatility windows that explain most of the move.
- The largest single-step repricing is a +0.26 jump. These moments are exactly what a regime model should surface quickly.
- Simple flow diagnostics suggest an event-driven market. Short-horizon returns are not explained by steady buy/sell imbalance.
Regimes (market moods)
Regimes are a practical way to answer a basic question: what kind of market is this right now? Calm drift and slow accumulation feel very different from a news-driven repricing. A regime model is a way to make that intuition explicit.
How the regime model works
A regime is a recurring market environment. Think "quiet drift", "choppy high volume", or "fast repricing". The point is not to predict the future. The point is to summarize the trade history into a few consistent states so it is easier to reason about execution conditions and repricing risk.
The model used here is an HMM-style regime detector. Trades are bucketed into fixed time windows, then features are computed for each bucket:
- Return: how much price changed from the previous bucket
- Volatility: absolute return (how violent the move is, ignoring direction)
- Activity: log notional volume and trade count
The HMM assigns each bucket to one of a small number of hidden states. After fitting, each state can be interpreted by its typical behavior. One state might have low volatility and low activity. Another might have high volatility and high activity, which often corresponds to a structural repricing.
Why this is useful: regimes give you a clean way to segment analysis, compare execution conditions, and avoid mixing normal periods with event periods when you summarize a market.
Price and volume after labeling regimes
Once regimes are labeled, the rest of the market read becomes easier to structure. The next step is the price path, then volume, then the timestamps that are most worth investigating.
Price path
The market trends toward near-certainty by the end of the sample. The useful information is in the path: when repricings happened, how quickly belief moved, and whether those moves came with heavy volume.
Volume clusters
Volume clusters. The market shows distinct surges where attention and information arrive. These windows matter most for execution and for explaining why the market moved.
The biggest repricing prints
This view surfaces the moments worth investigating first. If there is only time to dig into a few timestamps, start here.
Whales and concentration
This is not a one-wallet market, but the top wallets are large enough to matter. The largest prints are useful for checking whether big moves line up with size.