Weather Derivatives Trading Strategy Analysis
Analyze weather derivatives with Sourcetable AI. Calculate heating degree days, cooling degree days, and temperature-based contracts automatically. No complex formulas needed.
Andrew Grosser
February 16, 2026 • 14 min read
January 2024, Chicago: Utility sells HDD put at 1,200 strike, $10k per degree, collecting $350k premium. Arctic vortex hits—January totals 1,387 HDD, 207 above average. Contract expires worthless. Utility keeps full premium. This is the $50 billion weather derivatives market: trade temperature risk using objective meteorological data—HDD, CDD, rainfall, snowfall—with automatic payouts, no basis risk.
Excel makes this hell: import NOAA data, calculate daily averages, compute MAX(0, 65-temp) for each day, sum across months, analyze volatility with STDEV, model distributions with NORM.DIST—100+ formulas per contract. Sourcetable eliminates this. Upload temperature data, ask "Calculate January HDD vs 20-year average," get instant degree day totals and variance. Start analyzing weather risk for free at sign up free.
The Degree Day Accumulation Problem That Makes Excel Weather Analysis Impossible
Why does calculating HDD require daily temperature processing instead of using monthly averages?
Because degree days measure cumulative temperature deviation, and averaging temperatures before calculating degree days produces systematically wrong results. The formula HDD = MAX(0, 65°F - Daily Avg Temp) means each day contributes zero when above 65°F. If January has 15 days at 50°F (contributing 15 × 15 = 225 HDD) and 16 days at 70°F (contributing 0 HDD), the total is 225 HDD. But if you calculate monthly average first (61.45°F), then apply the formula, you'd get MAX(0, 65 - 61.45) × 31 days = 110 HDD—51% underestimate. You must process every single day individually, then sum the results.
In Excel, this means 31 rows for January (one per day), each with formulas: Column A = date, Column B = high temp, Column C = low temp, Column D = =(B2+C2)/2 for average, Column E = =MAX(0, 65-D2) for HDD contribution, and finally =SUM(E2:E32) to get monthly total. Multiply this by 12 months, 10 years of historical data, and you're managing 120 monthly calculations × 31 days = 3,720 individual degree day calculations. Add multiple cities (Chicago, Boston, New York for geographic diversification), and you're at 11,160 formula cells before even starting contract valuation.
What happens with missing temperature data or unusual readings?
Weather station data contains gaps (equipment failures, maintenance) and outliers (sensor errors reading 150°F or -80°F) that corrupt calculations if not handled. A single erroneous reading of -50°F on January 15 would contribute 115 HDD for that day (MAX(0, 65-(-50))), artificially inflating monthly totals and distorting contract valuations by thousands of dollars. Standard practice: identify outliers beyond 3 standard deviations, interpolate missing days from surrounding readings, and flag data quality issues. In Excel, this requires nested IF statements checking for outliers, AVERAGE functions calculating interpolations, and manual review of flagged data.
Sourcetable handles data quality automatically. Upload raw temperature data with gaps or errors, and ask: "Calculate HDD with data cleaning." The AI identifies statistically impossible readings (temperatures beyond historical ranges for the location and date), interpolates missing values using surrounding days and historical patterns for that date, flags data quality for review, and calculates degree days from the cleaned dataset. "Show me data quality issues" displays every interpolation and outlier adjustment with justification. What takes hours of manual Excel review and VLOOKUP formulas happens instantly with full transparency.
Real-World Example: Natural Gas Utility Hedging the 2023-2024 Midwest Winter
Let's walk through a complete weather derivative hedge for Midwest Energy Company (fictional), a natural gas utility serving 2 million customers across Chicago, Milwaukee, and Indianapolis. This example demonstrates exposure calculation, contract structuring, execution, monitoring, and settlement.
Step 1: Quantify temperature-revenue relationship (September 2023)
Upload 20 years of historical data into Sourcetable: winter HDD by city (November-March), monthly natural gas delivery volumes (million cubic feet), and monthly revenue. Ask: "Calculate correlation between HDD and revenue for each city."
Results:
- Chicago: r = 0.87 (87% correlation), each 100 HDD below average reduces revenue by $4.2M
- Milwaukee: r = 0.82, $2.8M per 100 HDD
- Indianapolis: r = 0.79, $3.1M per 100 HDD
Historical winter averages (Nov-Mar): Chicago 4,180 HDD, Milwaukee 4,380 HDD, Indianapolis 3,520 HDD. Standard deviations: Chicago ±320 HDD, Milwaukee ±280 HDD, Indianapolis ±260 HDD. One standard deviation warm winter (-320 HDD for Chicago) would reduce revenue by approximately $13.4M (320 HDD × $0.042M per HDD)—unacceptable to shareholders expecting consistent earnings.
Step 2: Structure the hedge (October 2023)
How do you determine the optimal strike price and tick value?
Balance premium cost against protection level—lower strikes provide better protection but cost more. Market analysis shows November 2023 HDD put options available for Chicago winter:
- Strike 4,000 HDD (43% probability): $850k premium, protects 180 HDD shortfall
- Strike 3,900 HDD (32% probability): $520k premium, protects 280 HDD shortfall
- Strike 3,800 HDD (22% probability): $280k premium, protects 380 HDD shortfall
Midwest Energy selects 3,900 HDD strike with $10,000 tick value. If winter HDD falls below 3,900, the contract pays $10,000 per HDD below strike. Maximum payout: (3,900 - worst case 3,200) × $10,000 = $7M. Premium cost: $520k. The hedge covers 37% of revenue exposure from moderately warm winters (1 std dev scenario: -320 HDD would trigger $2.1M payout vs $13.4M revenue loss—16% protection). For extreme warm winters (2 std dev: -640 HDD), the hedge provides $5.1M payout vs $26.8M revenue loss—19% protection.
Ask Sourcetable: "Compare strike prices from 3,700 to 4,100 HDD, show premium cost, payout probability, and expected value for each." The AI generates comparison table showing 3,900 strike offers best balance of protection (32% trigger probability) and cost ($520k vs $850k for 4,000 strike).
Step 3: Execute trades and monitor accumulation (November 2023 - March 2024)
Midwest Energy buys the HDD put on October 15, 2023. Throughout winter, Sourcetable tracks daily HDD accumulation and contract value in real-time. Upload daily temperature data and ask: "What's my current HDD accumulation and contract value?"
Daily tracking (cumulative HDD through each date, contract value based on remaining days and current pace):
- November 30: 680 HDD (vs 700 avg), pace suggests 3,980 total (above 3,900 strike), contract value $180k
- December 31: 1,520 HDD (vs 1,480 avg), pace suggests 4,060 total, contract value $80k (declining as warm winter looks unlikely)
- January 31: 2,680 HDD (vs 2,620 avg), pace suggests 4,180 total, contract value $30k
- February 29: 3,840 HDD (vs 3,780 avg), pace suggests 4,280 total, contract value $5k (nearly worthless)
- March 31 (final): 4,287 HDD, contract expires worthless (actual exceeded 3,900 strike by 387 HDD)
The 2023-2024 winter was 2.6% colder than average (4,287 vs 4,180 HDD average). Midwest Energy paid $520k premium but received zero payout—the cost of insurance that wasn't needed. Revenue came in at $487M (vs $480M budget based on 4,180 HDD), beating expectations by $7M. The hedge premium reduced net revenue to $486.5M, but provided certainty and protection against downside scenarios that would have significantly damaged earnings.
Step 4: Post-winter analysis and next year's planning (April 2024)
Ask Sourcetable: "Analyze the hedge performance. Was the 3,900 strike optimal, or should we adjust for 2024-2025?" The AI runs backtesting showing:
- Past 20 winters: 3,900 strike would have triggered payouts in 7 years (35% of the time)
- Average payout (when triggered): $1.8M
- Total premiums paid: 20 years × $520k = $10.4M
- Total payouts received: 7 years × $1.8M avg = $12.6M
- Net hedge value: +$2.2M over 20 years (+21% return on premium paid)
The backtest validates the hedge strategy—over two decades, payouts exceeded premiums by $2.2M while providing annual earnings protection. However, ask follow-up: "How would a 3,800 HDD strike have performed?" Sourcetable shows 3,800 strike triggered only 5 times (25% probability) with average payout $2.6M, total payouts $13M vs total premium $5.6M (20 years × $280k), net value +$7.4M (+132% return). The lower strike (less frequent protection) actually produced better economics due to much lower premiums, though it leaves more moderate warm winters (3,800-3,900 HDD) unhedged.
For 2024-2025 planning, Midwest Energy considers climate trends. Ask Sourcetable: "Show HDD trend over the past 20 years—is Chicago getting warmer?" The AI plots historical HDD with trend line showing -8 HDD per year average decline (winters are getting 1.9% warmer per decade). This suggests reducing strike to 3,850 HDD (splitting difference between 3,900 and 3,800) to maintain similar protection probability as climate warms.
Tick Value Selection: Why $10,000/HDD vs $20,000/HDD Doubles Your Risk
How do you determine the right tick value (payout per degree day)?
Match tick value to your revenue exposure per HDD—if you lose $5,000 revenue per HDD shortfall, hedge with $5,000 tick for 1:1 protection. The Midwest Energy example used $10,000 tick covering $42,000 per 100 HDD revenue exposure—only 24% hedge ratio. This is intentional: hedging 100% would be prohibitively expensive (premium = expected payout, meaning no risk reduction benefit after cost). Most utilities hedge 20-40% of exposure, accepting moderate revenue volatility while protecting against extreme scenarios.
Compare three tick values for the same 3,900 HDD strike:
- $5,000 tick: $260k premium, max payout $3.5M (700 HDD × $5k), 12% hedge ratio
- $10,000 tick: $520k premium, max payout $7M, 24% hedge ratio
- $20,000 tick: $1,040k premium, max payout $14M, 48% hedge ratio
Doubling the tick value doubles both premium cost and protection. The $20,000 tick provides 48% hedge coverage but costs $1.04M annually—a significant earnings drag even in years without payouts. The $10,000 tick offers reasonable protection ($7M max payout covers half the worst-case revenue loss) at manageable cost ($520k = 0.11% of total revenue).
Sourcetable makes tick value optimization easy. Upload your revenue-HDD relationship data and contract quotes, then ask: "What tick value minimizes earnings volatility at lowest cost?" The AI runs simulations across 20 years of historical data, testing tick values from $5k to $25k in $2,500 increments, calculating standard deviation of hedged earnings for each tick value, and identifying the optimal point where incremental volatility reduction costs more than the benefit. For Midwest Energy, the analysis shows $10k tick reduces earnings std dev by 18% at 0.11% revenue cost—optimal risk-adjusted trade off. Moving to $15k tick reduces std dev by only 23% (incremental 5%) at 0.16% revenue cost (incremental 0.05%)—diminishing returns.
Geographic Diversification: Why Multi-City Hedges Reduce Costs by 30%
Temperature across different cities isn't perfectly correlated—when Chicago has a mild winter, Minneapolis might still be cold. This allows portfolio hedging where you net out offsetting positions, reducing total hedge cost while maintaining protection. Midwest Energy serves three cities: Chicago (50% of customers), Milwaukee (30%), and Indianapolis (20%).
Calculate correlations by uploading historical HDD for all three cities and asking Sourcetable: "Show temperature correlation between these cities."
- Chicago-Milwaukee: r = 0.91 (highly correlated—both Great Lakes cities)
- Chicago-Indianapolis: r = 0.78 (moderately correlated—different climate zones)
- Milwaukee-Indianapolis: r = 0.74 (moderately correlated)
High correlation between Chicago and Milwaukee means warm winters usually affect both cities simultaneously—limited diversification benefit. The lower correlation with Indianapolis (0.78, 0.74) suggests 20-25% of winter temperature variation is independent. This matters for hedging: instead of buying three separate HDD puts (one per city), you can buy a weighted portfolio contract that pays based on aggregate HDD shortfall across all locations.
Compare two hedging approaches:
Approach 1: Individual city hedges (no diversification benefit)
- Chicago put: 3,900 strike, $10k tick, $520k premium
- Milwaukee put: 4,100 strike, $7k tick, $380k premium
- Indianapolis put: 3,300 strike, $6k tick, $240k premium
- Total premium: $1,140k
Approach 2: Portfolio hedge (diversification benefit)
Create weighted HDD index: 50% Chicago + 30% Milwaukee + 20% Indianapolis. Historical average: 0.5(4,180) + 0.3(4,380) + 0.2(3,520) = 3,508 weighted HDD. Standard deviation of weighted index: 260 HDD (vs 296 simple average of individual std devs—12% volatility reduction from diversification).
Buy single portfolio put: Strike 3,300 weighted HDD, $18k tick, premium $800k. This provides equivalent protection to individual hedges but saves $340k (30% cost reduction) because the market values the diversification benefit—the probability that all three cities simultaneously experience warm winters is lower than the probability of any single city being warm.
Sourcetable calculates optimal portfolio weights automatically. Ask: "What's the cheapest hedge structure covering 25% of my total revenue risk across all three cities?" The AI tests different weighting schemes (equal-weighted, revenue-weighted, volatility-weighted, correlation-optimized), simulates 20 years of historical outcomes, calculates premium costs for each approach from market quotes, and recommends the structure delivering target protection at minimum cost. The analysis might show that overweighting Indianapolis (despite only 20% of customers) provides better diversification because its temperature is less correlated with the larger Chicago/Milwaukee exposure.
Climate Trend Adjustment: Why Using 30-Year Averages Overprices Warming Protection
Should you use 10-year, 20-year, or 30-year historical data for strike selection?
Use recent data that reflects current climate conditions—30-year averages include decades from cooler climates and overestimate protection needs. Chicago winter HDD by decade:
- 1990-1999: Average 4,340 HDD
- 2000-2009: Average 4,250 HDD (-2.1%)
- 2010-2019: Average 4,180 HDD (-1.6%)
- 2020-2024: Average 4,095 HDD (-2.0%)
Chicago winters are 5.6% warmer than 1990s levels—245 fewer HDD annually. If you price HDD puts using 30-year average (4,268 HDD), you're setting strikes based on climate that no longer exists. A 3,900 strike represented 91.4% of 1990s average (moderately warm winter), but represents 95.2% of 2020s average (very warm winter). You'll pay higher premiums for protection against temperature levels that rarely occur in current climate.
Compare hedge costs using different historical periods for strike selection:
- 30-year data (1994-2024): Strike 3,900 (91.4% of avg), $520k premium, 32% trigger prob
- 15-year data (2009-2024): Strike 3,750 (89.9% of avg), $420k premium, 31% trigger prob
- 10-year data (2014-2024): Strike 3,700 (89.4% of avg), $380k premium, 30% trigger prob
Using 10-year climate data saves $140k premium (27% cost reduction) while maintaining similar protection probability (30% vs 32%). The savings come from aligning strike with current climate reality rather than paying for protection against 1990s temperature levels that no longer occur.
But won't climate volatility increase, making some winters still very cold despite warming trends?
Yes—this is why you need volatility analysis in addition to trend analysis. While average HDD declined 5.6%, standard deviation increased from 285 HDD (1990s) to 320 HDD (2020s)—12% more volatility. This means warm winters are more common, but occasional extreme cold winters are colder relative to the new average. The 2023-2024 winter (4,287 HDD) was 2.5 std dev above 2020s average—equivalent to the 1998 winter being 2.3 std dev above 1990s average. Extreme cold didn't disappear; it's just less frequent.
Sourcetable performs rolling volatility analysis automatically. Ask: "Show me how HDD volatility has changed over time, and adjust strike recommendations accordingly." The AI calculates 10-year rolling std deviation across your entire dataset, identifies if volatility is trending up or down, and recommends strikes that account for both mean shift (warming) and variance shift (increased volatility). For Chicago, this means lowering strike by 180 HDD (warming trend) but increasing tick value by 10% (volatility trend) to maintain similar tail risk protection despite changed climate.
Frequently Asked Questions
If your question is not covered here, you can contact our team.
Contact UsWhat's the difference between HDD puts vs calls, and which should I buy?
How are weather derivatives different from traditional insurance?
What weather stations are used for settlement, and can I choose?
How liquid is the weather derivatives market?
Can I trade weather derivatives for cities outside the U.S.?
How far in advance should I buy weather hedges?
What's a typical premium cost as percentage of notional value?
