What are P50 and P90 solar yield estimates?

P50 and P90 describe the probability of exceeding a modelled solar energy output over a specified period. P90 has a 90% probability of exceedance under the assessment's assumptions [S1]. It is usually lower than P50, but is not a fixed percentage reduction. Ask for the time horizon and uncertainty calculation alongside both values. Checked October 2026.

Last checked 18 October 2026

What do P50 and P90 mean on your solar proposal?

The letter describes an exceedance probability: the chance that output reaches or exceeds a stated energy value. P50 is the median of the modelled distribution. P90 is a lower threshold with more modelled outcomes above it [S1].

Read both alongside the same design and period. Your proposal should identify the roof layout, energy measurement boundary and losses included. Use the business solar hub to place the forecast in context, then regional solar yield for the resource background.

Keep power and energy distinct: PVGIS takes array power in kilowatt-peak (kWp) and models energy output with system losses included [S4]. Request the units beside every forecast.

How can you calculate P90 from P50?

You need an uncertainty assessment, not just the central estimate. An illustration, not a quote: assume normally distributed annual generation, a P50 of 100,000 kilowatt-hours (kWh), and a standard deviation of 5,000 kWh. The normal-distribution method places P90 approximately 1.28 standard deviations below the mean [S3].

These assumed inputs illustrate the arithmetic, not a UK yield benchmark. P90 here is not obtained by multiplying P50 by 90%. The distribution and uncertainty must fit your project.

Which time horizon should your yield report use?

Ask whether the figure describes a single operating year, a multiyear annual average or cumulative output. Solargis distinguishes the period used for variability calculations [S2]. A long-term average does not promise the same output in each individual year.

Request the assessment period beside every probability value. Keep degradation assumptions explicit when comparing different operating years. Do not accept a single unexplained percentage as a substitute for the calculation.

How accurate is a solar production estimate?

Weather variation and incomplete knowledge of the model both matter [S1]. Ask for an uncertainty breakdown covering resource data and simulation assumptions [S2]. Ask what evidence supports shading, downtime and loss inputs. A lower downside estimate can reflect wider uncertainty rather than a different roof design.

For your budget, request generation scenarios and a separate explanation of how output matches site consumption. Read solar self-consumption and solar payback. Energy probabilities alone do not establish a monetary return.

What should you do next?

Use the cut energy costs hub and site check to prepare your requirements, then get matched with an installer that covers your area.

Piggy Pineapple Ltd operates this referral service. We refer enquiries to installer partners and may be paid a fee. Our owner holds shares in at least one partner. Before sharing your details, we name the partner and disclose any shareholding our owner holds in it.