A Forecast That Remembers the Cycle express

Original: "Symmetry-Constrained Forecasting of Periodically Correlated Energy Processes"

· Cyril Voyant, Candice Banes, Luis Garcia-Gutierrez, Gilles Notton, Milan Despotovic, Zaher Mundher Yaseen

arXiv:2602.18949 · 2026-02-21 · CC BY 4.0 · 1 min

A new way to predict sunlight and wind is more accurate than older methods.

A solar panel on a roof produces little current in the morning, a lot at noon, and stays silent at night. Wind doesn't blow randomly either: it's often stronger in the morning and evening than during the day. Energy providers need to know in advance how much energy will be available in an hour or three, so they can switch on backup plants or store electricity in batteries in time.

The usual forecasting method is 'as it is now, so it will be': if the sun is shining brightly, they assume it will be just as bright in an hour. But this fails when day turns to evening or when the wind dies down. The new formula adapts to the daily rhythm: it compares not just the current value with the past, but the current value with what it was at the same time of day yesterday or a few days ago. It's like an experienced gardener who knows that cucumbers should be watered not 'when dry' but every evening after sunset.

The formula doesn't require training on large datasets. It works immediately, without powerful computers, and gives a result that's easy to verify by hand.

Tests on real data from solar and wind plants showed that the new method is more accurate than the old one, especially on horizons of several hours. This means we can keep fewer gas turbines in reserve and trust the renewable energy forecast more.

Whole graph →

Solar energy reaching Earth in one hour exceeds the entire world's annual energy consumption — the problem is just knowing exactly when and how much will arrive.

\tilde{\lambda}(t,\tau)=\tfrac{1}{2}\bigl(1+\rho(t,\tau)\bigr)

The forecast coefficient equals half the sum of one and the local correlation between the current value and the value shifted by time τ, accounting for the phase of the daily cycle. If the correlation is maximal (1), the coefficient is 1—the forecast repeats the past; if there is no correlation (0), the coefficient is 0.5—the forecast averages.

Scientists

Enrico Fermi Max Planck Ludwig Boltzmann Claude Shannon John von Neumann Josiah Willard Gibbs

Concepts

Analogue forecast probabilistic forecast real time stochastic processes stochastic time reversal symmetry climate energy conservation thermodynamic stability automated analysis seasonal forcing Hurst exponent symmetry automation prediction accuracy forecasting free energy product operator predictability drought

I am the author

A Forecast That Remembers the Cycle express