How wrong can the simple average be?

On a pool with a long flat shallow section and a short drop to the deep end, noticeably. The average assumes half the pool is deeper than the midpoint depth, and on that floor shape far less than half of it is. The direction of the error is always the same: a simple average overstates the volume for every floor shape except a uniform slope.

How do I do better than the simple average?

Measure the depth at several points along the pool and average those instead, weighting them by how much of the pool each represents. Depths at 5 or 6 points along the length, averaged, is much closer than two depths for any floor that is not a plain ramp.

A deep end shaped like an inverted truncated pyramid, with sloping walls on all four sides down to a small flat bottom, rather than a floor that simply slopes down to a flat deep end. It is common on UK domestic pools and it holds substantially less than a rectangular deep end of the same maximum depth.

Does the shallow end really matter that much?

It matters more than the deep end does, because it is usually the larger area. A pool is wider than it is deep, so a difference of a few centimetres across a long shallow section moves more water than the same difference at the deep end.

What is the exact method?

Measure the water rather than the pool. Fill it through a meter, or record the volume added when filling from empty, and you have the capacity with no geometry and no assumptions at all. Every method on this page is an approximation of that one.

How to measure average depth on a real pool floor

Adding the shallow and deep depths and halving them is exact for exactly one floor shape: one that slopes uniformly from one end to the other. Most UK pools do not. A flat shallow section that then drops away, or a hopper in the deep end, both hold less water than that average implies, so the calculated volume comes out high.

Last reviewed 14 August 2026

The short answer

Average depth is the input the volume is most sensitive to, and the standard shortcut only fits one shape of floor.

Floor shape	Simple average is	Effect on the volume
A uniform slope from end to end	Exact.	None. This is the case the formula is for.
Flat shallow section, then a slope down	Too high.	The average assumes depth increases steadily along the whole pool, and for the flat part it does not.
A hopper deep end	Too high.	A hopper narrows on all four sides as it goes down, so the deepest part occupies a small fraction of the plan area.

The three common floor shapes, and what the simple average does

The error always goes the same way. Every real floor shape other than a uniform slope holds less than the simple average predicts, so the calculated volume is a ceiling rather than a best estimate.

How this works in practice

Two better methods, and the second one removes the question rather than improving the answer.

Multi-point averaging. Take depths at 5 or 6 evenly spaced points along the length and average those. Each reading then represents a similar slice of the pool, which is what makes the average meaningful, and it tracks a flat section or a hopper far better than two end readings can.
Measuring the water. Fill the pool through a meter, or note the volume added when filling from empty. That is the capacity, exactly, for any shape, with no geometry involved at all.

The multi-point method is worth the ten minutes it takes on any pool with a shaped floor, which is most of them. Take the readings with a weighted line or a pole marked in centimetres, from the water surface rather than from the coping, because the coping sits above the water and adds a constant error to every reading.

For what this site does with the figure, the surface area matters more than the depth anyway. why the bucket test needs the surface area explains why an error there carries straight into a loss and an error in depth does not.

What changes the answer

Four things, and the first two are the ones people measure wrongly rather than estimate wrongly.

Measuring from the coping instead of the water surface. The coping is above the water, so every depth comes out too large by the same amount, and the error does not cancel.
Measuring at the wall rather than at the floor. A pool with a curved transition between wall and floor is shallower at the very edge than a step in from it.
A shaped shallow end, such as steps or a beach entry, which occupies plan area at almost no depth and pulls the true average down further than a slope does.
Whether the pool is full. Depths taken at a level below the normal water line understate every reading, and the calculation assumes a full pool.

None of these is a large error on its own. They all point the same direction, which is what makes them worth listing: several small overstatements compound rather than cancelling.

Assumptions and sources

Litres in a cubic metre

Exact by definition

SI definition. The litre is defined as one cubic decimetre, so one cubic metre is exactly 1,000 litres.

Nothing here rests on a constant. That the mean of two end depths equals the mean depth only for a linear slope is arithmetic, and the rest describes what real pool floors are shaped like.

The floor shapes listed are the common ones on UK domestic pools rather than a surveyed classification. If yours is a shape not described here, the same principle applies to it: the simple average is exact only if depth increases at a constant rate along the pool.

What to do next

Two things, and the first takes ten minutes.

Take five or six depth readings along the pool rather than two, and average those. On any pool that is not a plain ramp it is a materially better figure for very little work.
If the volume genuinely matters, for dosing or for expressing a loss as a percentage, measure the water going in instead. It is the only exact method and it makes every assumption on this page irrelevant.

Then take the surface area, rather than the volume, into the leak test. work out your pool surface area in square metres is the calculation that uses it.

Common questions

How wrong can the simple average be?: On a pool with a long flat shallow section and a short drop to the deep end, noticeably. The average assumes half the pool is deeper than the midpoint depth, and on that floor shape far less than half of it is. The direction of the error is always the same: a simple average overstates the volume for every floor shape except a uniform slope.
How do I do better than the simple average?: Measure the depth at several points along the pool and average those instead, weighting them by how much of the pool each represents. Depths at 5 or 6 points along the length, averaged, is much closer than two depths for any floor that is not a plain ramp.
What is a hopper?: A deep end shaped like an inverted truncated pyramid, with sloping walls on all four sides down to a small flat bottom, rather than a floor that simply slopes down to a flat deep end. It is common on UK domestic pools and it holds substantially less than a rectangular deep end of the same maximum depth.
Does the shallow end really matter that much?: It matters more than the deep end does, because it is usually the larger area. A pool is wider than it is deep, so a difference of a few centimetres across a long shallow section moves more water than the same difference at the deep end.
What is the exact method?: Measure the water rather than the pool. Fill it through a meter, or record the volume added when filling from empty, and you have the capacity with no geometry and no assumptions at all. Every method on this page is an approximation of that one.