How Our Commodity Edge Engines Work

One framework, four commodities — silver, gold, oil, and bitcoin. Same idea each time: Kalshi sells a binary contract on a public settle price; the listed options on the same underlying are priced by a different room full of market makers; the two don't always agree. We surface the gap in real time on every active strike.

The Core Idea

Kalshi's commodity contracts ask one question per strike: will spot finish above K at settle?The YES contract pays $1 if it does. The market price of that YES contract is — by construction — the market's implied probability that the event happens.

The same probability is priced a second way by the options market. An out-of-the-money call on the underlying ETF (SLV for silver, GLD for gold, USO for oil, IBIT for bitcoin) has its own implied probability of finishing in the money, which we can back out from its price using Black-Scholes. Two markets, two pricing channels, one underlying number. The gap between them is the edge.

Round-trip trading cost on a typical Kalshi commodity contract runs about 5pp once you include the spread plus fees. So a 12pp gap leaves roughly +7pp of expected-value edge after slippage. That's what each engine surfaces, with a confidence tier on top: HIGH if the liquidity guards pass (volume, spread, distance from spot), MEDIUM/LOW if they don't.

What Each Engine Actually Does Under the Hood

1. Pulls live spot. XAG/USD and XAU/USD from Pyth for silver and gold, the published WTI tape for oil, BRTI for bitcoin. Spot updates every few seconds during market hours.
2. Pulls the listed options chain on the matched ETF (SLV / GLD / USO / IBIT) in real time. Derives a probability that spot will finish above each Kalshi strike at the event close, using the risk-neutral measure (Black-Scholes N(d2)).
3. Pulls the live Kalshi YES contract price on every active strike of the matched event.
4. Computes the signed edge in percentage points — options-implied probability minus Kalshi YES price. Positive edge = the options market thinks the event is more likely than Kalshi is pricing.
5. Applies liquidity + risk guards. Edge must exceed round-trip cost. Strike must be near spot. Kalshi volume must be real. Then we tier the result and surface it.

Why This Setup Has Edge

Kalshi's commodity flow is mostly retail traders sizing in 1–10 contract clips. They have views, but they're not running a calibrated probability book. The options market on the same underlying is run by professional market makers who are running a calibrated probability book — they have to be, or they get picked off.

When the two prices disagree by more than the round-trip trading cost, the structural answer is usually the same: the options book is closer to right, and the Kalshi quote will drift toward it as smart money flows in. We're not always right — no edge is — but the calibration table on each tool page (and the full plots linked below) shows how often the model has been right across two years of replay.

The Calibration Evidence

We replayed every engine across every settled daily snapshot from January 2024 to present — 7,685 settled signals across silver, gold, and oil — and scored model probability against realized outcome. The hit rate is monotonic across every confidence bucket on all three commodities: higher model probability, higher realized hit rate, in the right shape.

How to Trade a Signal

1. Start with HIGH-tier rows. Those passed all four liquidity guards. MEDIUM and LOW rows are still actionable but carry a specific caveat in the rationale field — read it before sizing.
2. Verify on Kalshi. Click through to the live book before entering. Snapshots are real-time but books change second to second on liquid strikes.
3. Size with Kelly, fractional. Use the Kelly calculator at 0.5× — full Kelly on commodity edges is a fast way to blow up. Cap at 2% of trading account per position, 5% across correlated strikes.
4. Optionally hedge. Each row links to the matching ETF option position on Robinhood. Entering both legs turns the trade into a pure mispricing capture instead of a directional view on the underlying.
5. Hold to settle.Don't day-trade the Kalshi leg — the edge is structural and shows up at settlement, not on minute-to-minute price action.

A Few Honest Caveats

• Tail strikes carry a risk premium. Far out-of-the-money options have a known risk-premium gap vs the real-world probability. The engine flags these as MEDIUM or LOW and the rationale field calls it out.
• Macro events change the regime.OPEC meetings on oil, FOMC days on metals, ETF flows on bitcoin. The engine applies a confidence haircut around scheduled events but it can't price unscheduled headlines.
• USO is a futures-rolling ETF. Oil Edge applies a contango multiplier before computing the strike ratio. The tool-page methodology spells out the exact adjustment.
• Calibration is not a guarantee.The 2-year backtest shows the model has been monotonic and tight on strong calls. Past calibration is the best evidence available; it isn't a promise about the next signal.

Where to Go Next