Energy market analysis is the work of forming a view on where power and gas prices will go, and why. An analyst does not try to guess the exact price. They find where the market's price looks wrong, say why, and say what would prove them wrong. That view then shapes what the desk buys, sells or hedges.

This article is for anyone starting out as an analyst, or working beside one as a trader, developer or risk manager. It explains what moves British power and gas prices, how an analyst turns that into a view, and one evening worked through to a trade. It ends with the free data an analyst can start with today.

What the market's price already says

Every traded price is a view: the average opinion of everyone willing to trade. A forward price for next winter already includes the weather, the gas supply and the outages people expect. So an analyst's question is never "what will the price be?" in isolation. It is "where is the market's price wrong, and why?"

A useful view has three parts:

  • a difference from the market's price, with a direction: "too high" or "too low";
  • a reason: the one thing the analyst believes the market has missed or misjudged;
  • a test: the piece of news that would prove the view wrong, decided in advance.

The rest of this article is about where the reasons come from. In British power, almost all of them come down to one question: which power station is the last one needed?

How the British power price is set

Power stations are called on in order of cost. Wind, solar and nuclear cost almost nothing to run, so they go first. Gas plants come next, cheapest first, and the most expensive plant needed in each half-hour sets the price for everyone. This order is called the merit order, and physical foundations of energy commodities builds one step by step.

In Britain the last plant is usually gas. The Government says gas set the wholesale electricity price around 60% of the time in 2025, down from around 90% in the early 2020s. It expects that share to keep falling as wind and storage grow. So an analyst needs two things: the cost of a gas plant, and how many gas plants a given half-hour will need. The first moves slowly. The second can change in an hour.

The slow anchor: what a gas plant costs to run

A gas plant's running cost is mostly fuel and carbon. Four numbers give it:

  • the gas price, quoted in Britain in pence per therm; multiply by 10 and divide by 29.3071 for pounds per megawatt-hour (MWh);
  • the plant's efficiency: a modern plant turns about half the energy in its gas into power;
  • the carbon it emits: about 0.18 tonnes of carbon dioxide per MWh of gas burned;
  • the carbon price: a British plant buys UK emissions allowances and also pays the Carbon Price Support tax.

Carbon Price Support is £18 a tonne. The Government will remove it from April 2028, which will cut a gas plant's cost by several pounds per MWh overnight. On 20 November 2026 the UK and the EU concluded a deal to link their carbon markets. They aim to make it work from 1 January 2028. An analyst looking at prices for 2028 has to allow for both. Carbon markets explained covers the two schemes and the link.

Fuel plus carbon, divided by efficiency, is the plant's cost per MWh of power. The power price minus that cost is the clean spark spread: what the plant earns for running. The spark spread article works the formula through, and the dark spread article shows why carbon pushed coal out of the British market.

The fast swing: residual demand

The gas price changes over weeks. The number of gas plants needed changes every half-hour. The tool for that is residual demand: the demand left after the cheap sources have done their part.

Residual demand is demand minus wind, solar, nuclear, biomass, imports and battery output. Whatever is left must come from gas. More wind, less gas; a still, cold evening, more gas and a dearer last plant.

The 2025 figures for the UK show why wind matters so much:

Source Share of UK electricity, 2025
Gas 31.8%
Wind 29.5%
Nuclear 12.2%
Solar 6.9%
All renewables 52.1%

Coal produced nothing in 2025. The UK also imported a record 43.9 TWh (terawatt-hours; one is a million MWh) and exported 14.2 TWh.

Wind is the swing factor. It can supply most of the country one evening and almost nothing the next, and the forecast for tomorrow can move by gigawatts between runs. Demand moves with temperature, the time of day and the day of the week. Could a better wind forecast have saved Britain? shows what a wrong wind forecast costs, and how power trading works shows how a company trades out the difference.

The limits: imports, batteries and the grid

Three things stop Britain's price from moving freely.

Interconnectors. About 10 GW of cables link Britain to France, Belgium, the Netherlands, Norway, Denmark and the island of Ireland. Power flows towards the dearer market, so cheap French or Norwegian power can push the British price down. When a cable is out for maintenance, that help shrinks. Britain trades separately from the Continent's shared auction, and what Britain loses without coupling measures what that costs. How European energy trading works explains the other side of the cables.

Batteries. Britain had 8.5 GW of grid batteries operating at the end of September 2026, by Modo Energy's count. The Government's own statistics give 7.5 GW at the end of 2025. Batteries charge when power is cheap and sell into the evening peak, which shaves the peak. Britain's batteries: where did the money come from? looks at what they earn, and algorithmic power trading at how they trade. Battery storage in Great Britain covers all the markets a battery sells into.

The grid inside Britain. Much of the wind is in Scotland and much of the demand is in England. When the lines between them are full, the system operator pays Scottish wind farms to stop and English gas plants to start. In the year to March 2026 these thermal constraints cost about £2.0 billion, around two-thirds of the £3.1 billion NESO spent balancing the system.

Gas has its own analysis

The gas price is the slow anchor for power, and it has its own drivers.

In 2025 about 42% of Britain's gas came from its own fields and about 40% from Norway by pipeline. About 17% came by ship as liquefied natural gas (LNG), most of it from the United States.

So the British price follows Norwegian supply and outages, the world LNG market, European storage and the weather. Britain's gas hub, the NBP, trades in pence per therm. The Dutch hub, TTF, is the European benchmark and trades in euros per MWh. The gap between them, the basis, opens when the pipes between them are full or pointed the wrong way. How gas trading works follows one British gas day. The global gas market covers storage and the NBP-TTF basis, and LNG pricing explained the cargoes.

The market's own view: the forward curve

The forward curve is the list of prices the market will trade today for each future period: next week, month, quarter or season. In Britain the standard power products are baseload, every hour of the day, and peak, 07:00 to 19:00 on weekdays. The summer season runs from April to September and the winter season from October to March.

The curve is the market's view in numbers. An analyst compares their own view with it, period by period. Two habits help:

  • Check the curve against gas. If gas sets the evening price, the gap between the day's dearest and cheapest hours should be about twice the gas price. Our battery analysis measured that ratio year by year: 2.22 in 2022, 2.06 in 2023, 2.14 in 2024 and 2.25 in 2025. In 2026, to 15 October, it fell to 1.79, mainly because gas rose faster than evening power from September.
  • Read the shape. A winter priced well above summer says the market fears a cold, still winter. A curve that falls year by year says the market expects more wind and cheaper gas.

Energy derivatives explained shows how forwards are traded, and mark-to-market P&L how the curve values a desk's position each night.

Worked example: one evening at Northfield

Northfield is an invented trading firm. It owns Marsh End, a 400 MW gas-fired power station, and Carnmoor, a wind farm with a 70 MW export limit. Every price below is made up for the example. On Monday 23 November 2026, Northfield's analyst builds a view of Tuesday's 17:00 to 18:00 hour.

Step 1: the anchor

Tuesday's gas trades at 80 pence per therm, and carbon costs £50 plus £18 of Carbon Price Support. Marsh End is 50% efficient and emits 0.184 tonnes per MWh of gas.

Item Working £ per MWh of power
Gas 80 x 10 / 29.3071 = £27.30 per MWh of gas, / 0.5 54.59
Carbon 0.184 / 0.5 = 0.368 t x £68 25.02
Running cost 79.62

Step 2: residual demand

The analyst lists the forecast for Tuesday at 17:00. It is dark, so solar gives nothing.

GW
Demand 41.0
Wind -7.0
Nuclear -4.5
Biomass and hydro -3.0
Imports -5.5
Batteries -2.5
Left for gas 18.5

Imports are lower than usual. The largest cable to France, IFA, is in a planned outage that halves its capacity until the evening of 26 November. An analyst checks the outage notices before trusting any import figure.

Step 3: which plant is last?

The analyst's simple model of the gas fleet has three blocks:

Block Capacity, GW Cost, £/MWh
Newest gas plants, Marsh End among them 14 about 78 to 80
Older gas plants 9 about 90
Peaking plants 5 about 130

At 18.5 GW, all the newest plants run and 4.5 GW of the older ones are needed. An older plant is last, so the model says about £90.

Step 4: the market's price, and the gap

Brokers quote Tuesday's 17:00 hour at £96. The market is paying more than the model's £90, because of the low wind forecast and the cable outage.

The analyst thinks the wind forecast is too low. The latest weather model runs bring a front through on Tuesday afternoon. Carnmoor's own site forecast for 17:00 has risen from 30 MW to 55 MW. If wind reaches 12 GW, gas is needed for only 13.5 GW. The newest plants can cover that alone, and the price falls towards £80.

So the view is: Tuesday's 17:00 price is too high at £96, because the wind forecast is too low. It should clear nearer £85. The test: if Tuesday morning's wind forecast for 17:00 is still below 8 GW, the view is wrong. The price then stays near £90 or above.

Step 5: what the desk does with it

The view says Tuesday's price is likely to fall, so Marsh End's power is worth more sold today. Northfield sells Marsh End's 400 MW for that hour at £96. That locks in a clean spark spread of £96 - £79.62 = £16.38 per MWh, or £6,552 for the hour. If the view is right and the price clears near £85, Marsh End would have earned about £5 per MWh by waiting. If the view is wrong and the price ends higher, the desk has given up a few pounds, but it still has £16.38 locked in.

Notice what the analyst did not do. They did not forecast £85 to the penny. They found one reason the market looked wrong, put a number on it, and wrote down what would prove it wrong.

The free data an analyst uses

All of this can be done with public data. These sources are free:

Source What it gives Access
Elexon Insights Solution Generation by fuel every half-hour, system prices, the Balancing Mechanism, outage notices Free; its data API needs no key
NESO Data Portal Demand and wind forecasts, balancing costs Free
National Gas Data Portal Gas flows, demand, supply by terminal, the system price Free
Ofgem wholesale market indicators Forward and day-ahead prices, the generation mix, spark and dark spreads Free
DESNZ Energy Trends Official quarterly statistics on generation, supply and trade Free
Met Office Weather DataHub Weather forecasts and observations Free plan, with registration
Sheffield Solar PV_Live Estimated British solar output every half-hour Free
GIE AGSI+ European gas storage, by country and site Free, with registration
ENTSO-E Transparency Platform Prices, demand, generation and flows on the Continent Free, with registration; no British data since 2021

An analyst's week

Most analysts follow a routine like this:

  • Every morning: read the overnight wind and temperature forecasts, the outage notices, and yesterday's prices against the forecasts.
  • Before the day-ahead auctions: rebuild residual demand for tomorrow, find the last plant, and compare with where tomorrow trades.
  • Once a week: look further out. Check gas storage, LNG arrivals, cable outages and the forward curve for the next month and season.
  • After the fact: score last week's views. Which reasons were right, and which tests fired? A view that is never checked teaches nothing.

The commonest mistakes are forecasting a price without a reason and forgetting a cable outage or a clock change. Another is holding a view after its test has fired.

What an analyst's data has to get right

The analysis is only as good as the data underneath it, and energy data has traps:

  • Forecasts have versions. Tomorrow's wind forecast changes several times a day. A model must know which version it used, or yesterday's analysis cannot be checked.
  • Periods differ. Britain settles power in half-hours; the Continent's auctions trade quarter-hours. Clock-change days have 46 or 50 half-hours.
  • Days differ. The gas day runs from 05:00 to 05:00 UK time, not midnight to midnight.
  • Data arrives late and is revised. Metered generation and settlement figures are updated for weeks.

Building or replacing this on your own desk? It is the work A115 does. See the systems work

Where to go next

If you want to do this for a living, the Senior Power Market Analyst career roadmap sets out the skills, the pay and the path. Its portfolio project is a good next step: a public merit-order model for one market. It gives tomorrow's prices and checks its own error over thirty days. Our market analysis articles show the method applied to real British data.

What the full guide adds

This article is the short version. The full guide, Forming a View on Power Prices, sets out the analyst's method in full:

  • What a view is: a disagreement with the market plus a reason, and how the horizon changes the question;
  • The one equation: the price as the cost of the last plant, with a full merit order to work through;
  • The slow anchor and the fast swing: fuel and carbon, then residual demand, with the shape of the day, the year and the curve;
  • The boundaries: interconnectors, flexibility and grid constraints;
  • The method in seven steps, ending with the test that would prove the view wrong;
  • A second worked example: a spring Sunday when the view is that midday prices go below zero;
  • The mistakes that cost people money, and checkpoints to test yourself.

The full guide is in the community on Skool, with the rest of the series, and it is updated regularly as the markets change. If you want to work on a trading desk, read it to understand the subject properly. To join, start at the Energy Trading Fundamentals course page.

Sources