Roof orientation and pitch: what actually matters

Two numbers describe a roof

Pitch is the angle from horizontal: 0° is flat, 90° is a vertical wall. UK domestic roofs are mostly 30–45°.

Orientation — or Azimuth — is the compass direction the roof plane faces. UK and European modelling gives it in degrees from due south, so 0° is south, 90° is west, −90° is east and 180° is north.

The two interact. A steep pitch helps a south roof and hurts an east one, for a reason worth holding onto: tilting a panel points it at where the sun is at midday. If the roof already faces away from midday, tilting it points it harder at empty sky.

The whole picture in one table

Output as a percentage of that location’s own south-facing 35° figure, averaged across all 25 of our UK reference locations, with the spread in brackets.

Source: our own modelling from PVGIS 5.3, European Commission Joint Research Centre, at 20% system losses following the method in MCS MIS 3002. Each location is normalised against its own 35° south figure before averaging, so the table measures the cost of orientation rather than the difference between places. Dataset: data/irradiance.yaml.

Three things to take from it.

Orientation matters more than pitch. Moving from south to east at 35° costs 21%. Moving from 35° to flat while staying south costs 17%. But moving from 20° to 60° — a change that spans nearly every real British roof — costs nothing at all: both give 96%.

The ratios barely move across the country. Look at the brackets: mostly two or three percentage points wide across 25 locations spanning ten degrees of latitude. The absolute totals differ by 50% from Shetland to Sussex; the relative penalty for facing the wrong way is effectively a national constant. So this single table is usable anywhere in the UK, which is not true of any absolute figure.

Flat is orientation-blind. The 0° row is flat at 83% across every column, because a horizontal surface has no orientation to get wrong.

The result nobody mentions: flat beats pitched, off-south

Follow the columns downward rather than the rows across.

For anything from east round to north, the best available pitch is horizontal, and the advantage becomes enormous on the north side. A north-facing 35° roof gets 54% of the south-facing optimum; the same panels laid flat get 83%.

This holds at all 25 locations, so it is a property of the geometry rather than an artefact of one place. And it has a real application: on a flat roof, you choose the mounting angle. The received wisdom of “tilt everything south at 35°” is right if you can, but the framing choice on an east–west flat roof layout is between a modest tilt and near-horizontal, and near-horizontal is not the poor relation it is usually presented as.

What is worth paying to change

Almost nothing, on generation grounds alone.

The last row is the point of the table. Orientation is fixed by the building; shading frequently is not. A tree that can be pruned is generally a better investment than any amount of re-angling.

Location-by-location figures

The percentages above are national averages. If you want the absolute figure for your own roof, this is every location at 35° facing south:

Source: our own modelling from PVGIS 5.3, European Commission Joint Research Centre, at 20% system losses following the method in MCS MIS 3002. Full dataset: data/irradiance.yaml.

And the same locations for an east or west-facing roof at 35°, where the whole table shifts down by around a fifth:

Source: our own modelling from PVGIS 5.3, European Commission Joint Research Centre, at 20% system losses following the method in MCS MIS 3002. Full dataset: data/irradiance.yaml.

Every pitch and orientation combination is in the full table on our plug-in solar calculator, and both our calculators run off the same dataset.

The 35° convention, and where it comes from

35° is close to optimal for UK latitudes and it is also close to the pitch of a great many British roofs, which is why it has become the default assumption in quotes and in modelling. That is a convenience, not a physical law.

The optimum tilt for annual output rises with latitude, and shifts with what you want to optimise for: a steeper array does relatively better in winter, when the sun is low and demand is high, at the cost of some summer peak it might not have been able to use anyway. Our dataset holds pitch at five discrete steps rather than searching for a per-location optimum, because the difference between 30° and 40° is inside the noise and a table pretending otherwise would mislead.

Practical checks for your own roof

Find your orientation properly. A compass app on a phone gives magnetic north; the model wants true. In Great Britain the difference is currently small enough to ignore at this resolution, but a roof you think is “south” is often 20–30° off, and that is worth knowing before you accept a modelled figure.

Measure the pitch, do not guess it. Most phone spirit-level apps will do it against a rafter or the underside of the roof in a loft.

Count the planes. Many houses have two or three usable roof planes, and the question is rarely “what is the best one” but “which combination of them is worth using”. A south plane plus a west plane usually beats either alone, which is the subject of east–west versus south-facing.