Curiosities / Episode 01

The Half-Time Power Surge

How can millions of people watching football suddenly change an entire country’s electricity demand?

The whistle blows. People leave their sofas. In the control room, a very different match begins.

Follow the surge ↓
One whistle. Thousands of little switches.An original illustration connects a football pitch to household appliances and a rising demand line. It is a schematic, not recorded demand data.45:00HALF-TIMESYNCHRONISED DEMAND
One shared pause can synchronise millions of small decisions. Original schematic; no historical trace is shown.
01 / The country gets up

A football break.
A gigawatt response.

On 29 June 2021, England played Germany at Wembley in the delayed Euro 2020 tournament. At half-time, Britain’s electricity demand rose by approximately 1 gigawatt. After full-time, the pick-up was around 1.6 gigawatts.

The crowd in the stadium was only part of the story. People watching at home had been doing the same thing at the same time. A scheduled break gave them a shared moment to do something else: make tea, open the fridge, get on with the evening.

Electricity operators call this a TV pick-up. It is the synchronisation that matters. Half a million appliances scattered through an evening are a different problem from half a million switched on together.

Recorded increases reported by the system operator, now NESO. These are changes in total demand, not a count of kettles.

02 / Try the mechanism

Same cups of tea. Different surge.

Start with 500,000 people and a 2,000-watt kettle each. Then spread their start times over ten minutes. What happens to the peak? What happens to the total energy?

Build an illustrative pick-up

Model / Not historical data

Your appliance assumptions
500,000
2,000 W
3 min
All at once
Uniformly distributed across this window. Each kettle runs once.
Peak extra power
1,000 MW
Total extra energy
50.0 MWh
Illustrative extra demand over time500,000 2,000-watt kettles start together and run for three minutes, adding 1,000 MW for that period. The model uses 50 MWh.EXTRA POWER (MW)1,000500003 minTIME SINCE FIRST SWITCH-ON
Area under the line is energy. The line’s height is power. The axes rescale when your assumptions change.

All start together: a 1,000 MW peak, using 50.0 MWh in total.

Compare magnitudes, not causes

The operator’s recorded pick-ups include the net effect of many activities. Matching one with kettle arithmetic does not show which appliances actually caused it.

Every bar below shares a fixed 0 to 6,000 MW scale. Historical bars show reported increases, not full demand curves.

Forecast, not a measurement: ahead of the England–Italy final in July 2021, the operator anticipated up to 2 GW. That pre-match forecast is not presented here as an observed result. Read the forecast.
Your illustrative peak1,000 MW
Calculated from your assumptions
2021 / England–Germany, half-time≈1 GW
Reported by NESO · 29 June
2021 / England–Germany, full-time≈1.6 GW
Reported by NESO · 29 June
1990 / England–Germany, penalties end≈2.8 GW
Historic pick-up reported by NESO

Power is the pace. Energy is the amount. MW and GW tell you how fast electricity is being used at a moment. MWh tells you how much is used over time. One GW is 1,000 MW. One GW sustained for 3 minutes uses 50 MWh.

Spreading the same kettle runs out lowers the peak when fewer overlap. It does not reduce their total energy in this model. Reducing the time each runs does reduce energy.

What the kettle model assumes

Every person uses one identical appliance at constant power for the chosen duration. Starts are either simultaneous or uniformly distributed through the selected window. This is a continuous approximation, so it can represent fractional groups of users. There is no background demand, thermostat cycling, diversity of appliances or transmission loss.

All-on power (MW) = people × watts ÷ 1,000,000
Energy (MWh) = all-on power × run seconds ÷ 3,600

At each moment, extra power depends on how many starts occurred during the preceding run duration. A start window longer than one kettle run reduces the maximum overlap. The smooth ramps represent that assumed distribution, not a measured historic curve.

03 / The other side of the switch

The grid cannot wait for the tea to finish.

Electricity supply and demand must remain closely balanced. The system has some physical inertia, but it is not a giant store that can casually supply an unplanned gigawatt for minutes. In Britain, grid frequency stays near 50 hertz. If demand outruns supply, frequency tends to fall; too much supply tends to raise it.

A football match is unusual because the exact second of the break is uncertain, yet the likelihood of a shared break is predictable. Operators can prepare for the event before the whistle.

  1. Before: anticipate the audience

    Combine the demand forecast with the broadcast schedule, expected viewing and previous events. Be ready for half-time, full-time and the possibility of extra time or penalties.

  2. At the break: bring flexibility

    Keep reserve available. Fast frequency-response services react to imbalance; operators also ask generators and flexible demand to change output or consumption through balancing arrangements. Pumped-storage hydro can release stored water to produce electricity quickly.

  3. After: follow the way back down

    When appliances finish, that extra demand disappears. Supply has to follow the decline too. Preparing means managing both sides of the surge, rather than simply keeping every available generator running.

Mechanisms: NESO’s Balancing Mechanism and frequency-response services.

04 / A human-shaped power system

A small decision becomes a system event.

A kettle is ordinary. Hundreds of thousands sharing a start time are infrastructure. The same distinction matters when many electric vehicles charge at once or when heating switches on across a cold city.

The useful question is not only “How much electricity?” It is also “How much, at the same time?” Move the model’s start window once more. The energy can stay the same while the balancing task changes.

Follow the evidence

All illustrations are original SVGs. This page makes no claim to reconstruct an event’s second-by-second demand.