AI Prompt Library for Natural Gas Trading

Natural gas is the most weather-sensitive commodity in the energy complex. A single polar vortex forecast revision can move Henry Hub front-month contracts 8-12% in a session — faster than most discretionary traders can build a thesis. AI changes the speed equation, but only if you’re asking the right questions.

The gap between traders who use AI generically and those who use it precisely is widening. Asking ’what drives natural gas prices?’ returns a textbook. Asking an AI to compare current EIA storage draws against the five-year seasonal average and flag deviation thresholds returns an edge. The prompt is the strategy.

This library gives natural gas traders — from Henry Hub basis traders to LNG freight desk analysts — a structured set of prompts mapped to real workflows: storage analysis, weather demand modeling, supply-side disruptions, spread trading, and risk framing. Copy, adapt, deploy.

Why Natural Gas Demands Its Own Prompt Architecture

Natural gas doesn’t trade like crude oil or gold. Its price is driven by a layered interaction of pipeline constraints, LNG export capacity utilization, power burn demand, residential heating degree days, and storage injection/withdrawal cycles — all on different time horizons. A prompt framework built for equities or macro FX will miss most of this structure.

Henry Hub is the benchmark, but basis differentials at Waha, Algonquin, and SoCal Citygate can diverge hundreds of percent from the prompt during cold snaps or pipeline outages. Effective AI prompts for natural gas need to be regionally specific, seasonally aware, and anchored in the actual data cadence — EIA Weekly Natural Gas Storage Report drops every Thursday at 10:30 AM ET, and the market moves before most analysts finish reading it.

Building prompts that mirror how the gas market actually works — not how a generalist AI assumes it works — is the difference between signal and noise. Every prompt in this library is structured around a specific data input, a defined analysis task, and an actionable output format.

• Henry Hub spot and futures: front-month, 12-month strip, winter/summer spread
• EIA Weekly Storage Report: actual vs. consensus vs. five-year average
• LNG feed gas nominations: Sabine Pass, Freeport, Corpus Christi export terminals
• Weather models: HDD/CDD forecasts, GFS vs. European model divergence
• Production data: Haynesville, Marcellus/Utica rig counts and output estimates
• Power burn demand: gas-to-coal switching thresholds by region

Storage Analysis Prompts: Reading the EIA Report Faster

The EIA storage number is the single highest-impact weekly data release for natural gas. Consensus estimates from surveys of analysts and traders set the market’s expectation. The price reaction is almost entirely a function of the surprise — how far the actual draw or injection deviates from consensus, and how that deviation shifts the trajectory toward the five-year average band.

AI can compress the post-release analysis from 20 minutes to 90 seconds. Feed it the actual number, the consensus, last year’s comparable week, and the five-year average, and a well-constructed prompt will output the implied supply/demand balance shift, the seasonal trajectory implication, and the directional bias for the front-month contract.

Use the prompt below after each Thursday release. Modify the numerical inputs each week. The structure stays constant; the insight compounds over time as you build a log of AI-generated storage summaries to backtest against price action.

Act as a natural gas market analyst. This week's EIA storage report showed a [actual draw/injection] of [X] Bcf against a consensus estimate of [Y] Bcf. The five-year average for this week is [Z] Bcf. Current total storage is [A] Bcf versus the five-year average of [B] Bcf.

Analyze: (1) the surprise magnitude and its directional implication for Henry Hub front-month, (2) whether the current storage trajectory is tightening or loosening relative to seasonal norms, (3) what storage level at the end of injection season (Nov 1) this trajectory implies, and (4) the key weather or demand variable most likely to shift this trajectory in the next 30 days.

Output as a structured brief: Surprise Assessment, Seasonal Trajectory, End-of-Season Projection, Key Risk Variable.

Weather Demand Prompts: Translating Forecasts into Price Bias

Heating degree days and cooling degree days are the translation layer between meteorology and natural gas demand. A 10-HDD week in the Northeast consumes roughly 2-3 Bcf/day more than a neutral week. When the GFS and European (ECMWF) models diverge on a 6-10 day outlook, that spread itself becomes a tradeable signal — the market prices toward the European model, which has historically shown better skill at longer ranges.

AI won’t replace a meteorologist, but it can rapidly synthesize weather data into demand implications and cross-reference them against current storage positioning. The prompt below is designed for the Monday morning setup, when weekend weather model runs often diverge sharply from Friday’s positioning.

Pair this prompt with actual HDD data from NOAA’s Climate Prediction Center. The more specific your inputs, the more precise the output. Vague weather descriptions produce vague demand estimates.

Act as a natural gas demand analyst. The current 6-10 day weather outlook shows [GFS model HDD forecast] versus [ECMWF model HDD forecast] for the Northeast and Midwest combined. The current week's actual HDD came in at [X] versus the 30-year normal of [Y].

Estimate: (1) the implied daily demand variance in Bcf/day if the GFS scenario verifies versus the ECMWF scenario, (2) the cumulative storage impact over the 6-10 day window under each scenario, (3) which scenario is currently more reflected in Henry Hub prompt month pricing based on [current price], and (4) the trade setup if model convergence occurs toward the [colder/warmer] scenario by Wednesday.

Format as: Demand Variance Estimate, Storage Impact Range, Market Pricing Assessment, Trade Setup.

LNG Export and Supply-Side Disruption Prompts

U.S. LNG export capacity now exceeds 14 Bcf/day of feed gas demand when all terminals run at full utilization. That has structurally tightened the domestic supply/demand balance and created a direct transmission mechanism between global TTF prices and Henry Hub. When European storage is low and TTF trades at a large premium to Henry Hub on an energy-equivalent basis, LNG export demand pulls hard on domestic supply.

Supply disruptions — Haynesville freeze-offs, Appalachian pipeline maintenance windows, Gulf Coast hurricane impacts — hit the market fast. AI prompts designed for rapid scenario analysis help traders frame the magnitude of a disruption before price discovery fully runs.

The most underused prompt category in natural gas is the supply-side shock scenario. Most traders focus on demand. The asymmetric moves often come from supply.

Act as a natural gas supply analyst. A reported production disruption in [region] is estimated at [X] Bcf/day for an expected duration of [Y] days. Current Henry Hub front-month is at [$Z/MMBtu]. LNG feed gas nominations are running at [A] Bcf/day against nameplate capacity of [B] Bcf/day.

Analyze: (1) the total supply shortfall implied by this disruption, (2) the required storage draw to offset it assuming demand holds flat, (3) the implied Henry Hub price sensitivity based on current storage surplus/deficit versus five-year average, and (4) whether LNG export curtailments are likely as a pressure relief mechanism and at what Henry Hub price level that historically occurs.

Output: Supply Shortfall Estimate, Storage Impact, Price Sensitivity Range, LNG Curtailment Threshold.

Spread Trading Prompts: Calendar Spreads and Basis Differentials

The November/March spread — colloquially ’the widow maker’ — is the most volatile calendar spread in commodity markets. It captures the market’s pricing of winter supply adequacy: November is pre-winter, March is post-withdrawal season. When storage enters winter below the five-year average, this spread widens aggressively. AI can help frame the historical distribution of this spread relative to current storage positioning.

Basis differentials between regional hubs and Henry Hub are equally important for physical market participants and sophisticated financial traders. Waha basis — West Texas gas at the Permian Basin hub — has gone deeply negative during pipeline constraint periods, trading at discounts exceeding $3/MMBtu to Henry Hub. Prompts that map current pipeline capacity utilization against production growth give traders early warning on basis blowouts.

Use AI not to predict the spread, but to define the scenario envelope — what range of outcomes is historically consistent with current fundamentals, and where is the current market price relative to that range.

• Nov/Mar Henry Hub spread: reflects winter supply adequacy pricing
• Summer/winter strip spread: captures seasonal storage economics
• Waha basis: Permian takeaway capacity vs. production growth
• Algonquin basis: Northeast demand premium during cold snaps
• Henry Hub vs. TTF: LNG export arbitrage window monitor
• 12-month strip vs. spot: backwardation/contango structural signal

Risk Framing Prompts: Position Sizing in a High-Volatility Commodity

Natural gas has a 30-day realized volatility that regularly exceeds 60-80% annualized — two to three times crude oil in active weather seasons. Standard position sizing models built for equity volatility regimes will systematically oversize natural gas exposures. A prompt that forces explicit volatility-adjusted risk framing before entering a trade is not a luxury — it’s a risk management requirement.

The key inputs for a natural gas risk prompt are: current implied volatility (from NYMEX options), the anticipated holding period, the specific catalyst that defines the trade thesis, and the storage/weather scenario that represents the trade being wrong. AI can rapidly construct the scenario matrix and back out the implied position size consistent with a defined risk budget.

Natural gas rewards traders who size correctly almost as much as those who are directionally right. A well-sized losing trade is survivable. An oversized correct trade that gets stopped out on intraday volatility before the thesis plays out is the most common way capital leaves the natural gas market.

Act as a natural gas risk manager. I am considering a [long/short] position in Henry Hub [month] futures at [$X/MMBtu]. My thesis is [brief thesis]. The primary risk scenario that invalidates this trade is [risk scenario]. Current 30-day implied volatility is [Y%]. My total risk budget for this trade is [$Z].

Calculate: (1) the 1-standard-deviation daily price move implied by current volatility, (2) the recommended stop-loss level based on a [1x / 1.5x / 2x] ATR placement, (3) the number of contracts consistent with my risk budget given that stop distance, and (4) the key data releases or events in my holding period that could accelerate the invalidation scenario.

Output: Daily Vol Estimate, Stop Level, Contract Sizing, Key Event Risk Calendar.