The Spark Spread & Power Plant Economics

A gas power station is a machine that buys gas and sells electricity. The spark spread is its profit margin. It is the most watched number in UK power trading.

This page explains the spark spread from scratch. You will see the formula, how to convert UK prices, a worked example and a short Python calculator. No trading background is needed.

What is a spark spread?

In trading, a spread is the difference between two prices. The spark spread is the difference between the price of electricity and the cost of the gas needed to make it.

Both prices are in £ per megawatt-hour (£/MWh). The rule is simple:

• Positive spread: the power is worth more than the gas. The plant runs.
• Negative spread: the plant would lose money. It stays off.

Why does this one number matter so much? In Britain, gas plants usually set the price of electricity. The grid switches plants on from cheapest to most expensive. The last plant needed to meet demand sets the price for everyone. Most hours, that last plant burns gas. This ranking is called the merit order, and Physical Foundations explains it in detail.

Efficiency and the heat rate

A power station cannot turn all the energy in gas into electricity. Some is lost as heat. Efficiency measures how much comes out as power.

UK traders use a standard efficiency of 49.13%. Exchanges and price reporters (ICE, ICIS) all use it. It is a market convention, not a law of physics. It means everyone trades the same contract without arguing about plant details.

At 49.13%, you need about 2.035 MWh of gas to make 1 MWh of electricity:

That 2.035 is called the heat rate. Real plants differ. A modern combined cycle gas turbine (CCGT) reaches 55-60%. Older units sit around 45-48%. Owners adjust their trading to their own plant's numbers.

The clean spark spread formula

Burning gas releases carbon dioxide. In the UK, a power station must buy a carbon allowance (a UKA) for every tonne it emits. The clean spark spread subtracts that cost too. "Clean" means the carbon is paid for.

Where does 0.375 come from? Two steps:

1. Burning 1 MWh of natural gas releases about 0.184 tonnes of CO₂. This is the figure ICIS uses for its benchmark.
2. The plant burns 2.035 MWh of gas for each MWh of power. So 0.184 x 2.035 = about 0.375 tonnes of CO₂ per MWh of electricity.

In the UK, the clean spark spread is the version that matters. Every gas plant pays for carbon. Carbon is not a small item. At £40-50 per tonne, it adds £15-19 to the cost of each MWh.

Converting UK gas prices to £/MWh

There is one catch. UK gas trades in pence per therm, not £/MWh. You must convert it before using the formula.

So gas at 70 p/therm costs 70 x 0.3412 = £23.88 per MWh of gas.

Worked example: a UK winter day

Here is the full calculation with round, illustrative UK prices:

The plant makes £12.66 on every MWh. That is positive, so it runs. But the cushion is thin. If gas rose to about 88 p/therm, the margin would be gone.

Now try a windy afternoon. Wind floods the grid and power drops to £30/MWh. Same gas, same carbon: £30 - £48.61 - £18.73 = -£37.34/MWh. The plant should stay off.

Spark spread calculator in Python

The same steps as a function. Change the inputs to run your own "what if" cases.

THERMS_PER_MWH = 34.12
CO2_PER_MWH_GAS = 0.18404  # tonnes of CO2 per MWh of gas burned


def clean_spark_spread(
    power_price: float,
    gas_price_p_therm: float,
    carbon_price: float,
    efficiency: float = 0.4913,
) -> float:
    """Clean spark spread for a UK gas plant, in GBP/MWh of power.

    Args:
        power_price: Electricity price in GBP/MWh
        gas_price_p_therm: Gas price in pence/therm (NBP)
        carbon_price: Carbon price in GBP/tonne CO2 (UKA)
        efficiency: Plant efficiency (default 0.4913, the market benchmark)
    """
    gas_price_mwh = gas_price_p_therm * THERMS_PER_MWH / 100
    fuel_cost = gas_price_mwh / efficiency
    # Emissions factor per MWh of power = CO2_PER_MWH_GAS / efficiency (0.3746)
    carbon_cost = carbon_price * CO2_PER_MWH_GAS / efficiency
    return round(power_price - fuel_cost - carbon_cost, 2)


print(clean_spark_spread(80, 70, 50))  # 12.66
print(clean_spark_spread(30, 70, 50))  # -37.34

Try an efficiency of 0.45 for an older plant. The margin halves, to about £6.50/MWh. Why so much? Both the gas cost and the carbon cost are divided by efficiency, so both rise. Small efficiency gaps add up to large sums over a season.

Why the spread changes through the day

UK power is traded in half-hour periods, 48 a day. Gas usually trades as a whole day. So gas cost stays flat while the power price moves. The spread follows the power price.

• Overnight: low demand. The spread is low or negative.
• Morning peak (about 7-9am): people wake up. The spread jumps.
• Afternoon: steady business demand. A middling spread.
• Evening peak (about 4-7pm): cooking, heating and lights. Usually the highest spread.

This creates shape risk. Say the plant only makes money for 6 hours a day. If you sold power for all 24 hours, you must deliver for 18 hours at a loss.

Wind, solar and negative spreads

Wind and solar have no fuel cost. When they produce a lot, they push gas plants out of the merit order. Power prices can fall below the cost of gas. The spark spread turns deeply negative.

Gas plants that must keep running lose money. Batteries gain: they charge cheaply and sell back at the evening peak.

Why plants sometimes run at a loss

A negative spread does not always mean the plant switches off. Real plants have limits a formula ignores:

• Start-up costs. A cold start can cost £20,000-£100,000 in fuel and wear. Running through two bad hours can be cheaper than stopping.
• Grid services. National Grid pays some plants to stay on and keep the system stable. That income can outweigh a small loss on power.
• Take-or-pay gas contracts. If the gas is paid for anyway, burning it costs nothing extra.
• Heat customers. Some plants also heat nearby buildings. They run when it is cold, whatever the power price.

How traders use the spark spread

Running the plant. Each day, generators sell tomorrow's power in the day-ahead auction. A positive spread means they offer the plant's output. A negative one means they hold back.

Locking in the margin (hedging). A trader can fix the spread months ahead. They sell power, buy gas and buy carbon for the same period. That is three trades, in a fixed ratio. For every 1 MWh of power sold, they buy 2.035 MWh of gas and 0.375 tonnes of carbon allowances.

Reading the market. Gas trades more often than power, so its price moves first. If gas rises by £5/MWh, power should rise by about £10/MWh (5 x 2.035) to keep the spread. If power has not moved, traders ask why.

Spark spread vs dark spread and other spreads

Every type of power station has its own spread. They share one idea: power price minus the cost of fuel.

A note on the US. American traders often quote an "implied heat rate" instead: the power price divided by the gas price. It is the same idea expressed as a ratio. In Britain and Europe, the spread is quoted in £ or € per MWh.

How This Fits the Curriculum

If you're new to energy markets, start with the Energy Trading 101 overview. Physical Foundations taught you merit order dispatch and why gas plants are marginal price-setters. Ships and Storage showed you where gas prices come from - the global infrastructure connecting Henry Hub to NBP. This guide combines those inputs: you now understand both sides of the spark spread formula (power prices from merit order, gas prices from global markets) and can calculate the margin that determines which plants run.

Prerequisites

Complete Physical Foundations and The Global Gas Market guides first. You need to understand merit order dispatch and global gas pricing dynamics.

Mastery Tip

Build the Python calculator above, then manually track power and gas prices for one week using free sources (Elexon BMRS for day-ahead power, National Grid for gas). Calculate the spread for morning peak (7-9am) and evening peak (4-7pm) each day. Track which hours show positive spreads versus which hours CCGTs actually dispatched (visible on BMRS). The gap between theory (positive spread = should run) and reality (plants running at losses, or idle during positive spreads) will reveal the physical constraints and commercial strategies that pure formulas miss.