What cut and fill is, and why the answer is never symmetrical
Grading a site or building a road means lowering the ground in some places and raising it in others. The volume you dig is the cut; the volume you build up is the fill. Balancing them on site is the single largest cost lever in earthwork, because material you neither import nor export is nearly free compared with material that rides a truck.
The trap is assuming that a cubic yard of cut builds a cubic yard of fill. It does not, and the reason is that the three volumes an estimator deals with measure the same soil in three different states. Bank measure is soil undisturbed in the ground. Loose measure is that soil after excavation, when fragmentation has added void space — typically 20–30% more volume for common earth. Compacted measure is the same soil rolled into an embankment, where the compactive effort squeezes it tighter than it was in the ground, so it occupies less volume than its bank state.
The two adjustments run in opposite directions and answer different questions. Swell tells you how many trucks you need, because trucks carry loose material. Shrinkage tells you how much bank volume you must dig to build a given embankment, because the fill quantity on a plan is a neat-line compacted figure. Confusing them is the classic earthwork error, and it is expensive in both directions: over-order borrow and you pay for material you cannot place; under-order and the job stops.
The average end area method, and where it goes wrong
Cross-sections are cut through the design and the existing ground at regular stations along the alignment. At each station a planimeter or the design software reports two areas: how much material lies above the finished grade (cut) and how much lies below it (fill). Those are the end areas.
Between two adjacent stations the solid is a prismatoid. The average end area method approximates its volume as the mean of the two end areas times the distance between them, and divides by 27 to reach cubic yards. Sum every interval and you have the total. With a constant station interval that sum collapses to the trapezoidal rule, where the first and last stations count half and every interior station counts once.
The approximation is exact only when the area varies linearly along the alignment — a wedge or a prism. Wherever the section is genuinely a prismatoid whose area varies quadratically, average end area overstates the volume, and the error grows with the difference between the two end areas. The correction is the prismoidal formula, which weights a mid-section area four times: V = L(A₁ + 4Am + A₂)/6. It needs a section measured at the midpoint, which is why average end area survives — it needs only what you already have, and on ground that is surveyed at close intervals the difference is usually within the accuracy of the survey itself. Where the sections are far apart and the areas differ greatly, take the extra section instead of applying a correction factor.
Two rules protect the accuracy. First, put a station at every grade break — every place where the ground or the design changes slope — because the method assumes linear variation between stations and a break violates that assumption exactly where it matters. Second, never average a cut area with a fill area. A section that is part cut and part fill contributes to both totals, and the transition point between an all-cut section and an all-fill section needs its own zero-area station or the volumes on both sides are wrong.
Worked example: 300 ft of road at 50 ft stations
Seven stations at 50 ft spacing. Cut end areas, in square feet: 0, 45, 120, 180, 95, 20, 0. Fill end areas: 60, 25, 0, 0, 15, 70, 110. Shrinkage 15%, swell 25%, 12 yd³ trucks.
- Average the cut areas pairwise. (0+45)/2 = 22.5; (45+120)/2 = 82.5; (120+180)/2 = 150; (180+95)/2 = 137.5; (95+20)/2 = 57.5; (20+0)/2 = 10. Sum = 460 ft².
- Cut volume. 460 × 50 = 23,000 ft³ ÷ 27 = 851.9 yd³ bank.
- Average the fill areas pairwise. 42.5; 12.5; 0; 7.5; 42.5; 90. Sum = 195 ft².
- Fill volume. 195 × 50 = 9,750 ft³ ÷ 27 = 361.1 yd³ compacted, at neat line.
- Bank volume needed for that fill. 361.1 ÷ (1 − 0.15) = 361.1 ÷ 0.85 = 424.8 yd³.
- Net balance. 851.9 − 424.8 = +427.1 yd³ bank — a surplus.
- Loose volume of that surplus. 427.1 × 1.25 = 533.8 yd³.
- Truckloads out. 533.8 ÷ 12 = 44.5 → 45 loads.
Check the shrinkage step in isolation, because it is the one people get backwards. You need 361.1 yd³ of finished embankment. Every bank cubic yard you place shrinks to 0.85 yd³ once compacted. So the bank volume required is the fill divided by 0.85, not multiplied — 424.8, which is more than the fill, not less. Dividing when you should multiply would have given 307 yd³ and understated the borrow by nearly 40%.
Reading the balance and the mass diagram
The net balance is the number that decides how the job is priced. A positive figure means surplus material: after the embankment is built to grade, that much bank volume is left over and has to be exported, wasted in a designated area on site, or absorbed by raising the finished grade. A negative figure means the cut cannot supply the fill and the difference has to come from a borrow pit. A designer aiming for a balanced site is trying to drive that number toward zero by adjusting the vertical alignment, and moving the profile up or down a foot over a long alignment shifts thousands of cubic yards.
The mass diagram is how earthwork engineers read the same information spatially. It plots cumulative cut minus the bank volume consumed by fill against distance along the alignment. Rising sections are net cut; falling sections are net fill. Any horizontal line drawn across the curve cuts it at two points where the accumulated volume is equal, which means the material between those two stations balances and the haul distance between them is the distance you actually pay for. The curve's overall trend tells you at a glance whether the job exports, imports, or balances, and where the free-haul limits fall.
Treat the truckload count as a floor. Trucks fill by volume or by legal weight, whichever comes first, and saturated clay reaches its weight limit well before the body is full. Ask the hauler what payload they can legally carry, and compare it against the loose volume figure here. For the bedding and surfacing quantities that follow the grading work, the gravel tonnage calculator and the asphalt tonnage calculator pick up the imported materials.
Typical swell and shrinkage factors by material
| Material | Typical swell | 1 bank yd³ hauls as | Typical shrinkage | 1 bank yd³ builds |
|---|---|---|---|---|
| Sand and gravel | 10–15% | 1.10–1.15 yd³ | 10–12% | 0.88–0.90 yd³ |
| Common earth / loam | 20–30% | 1.20–1.30 yd³ | 10–20% | 0.80–0.90 yd³ |
| Clay, dense | 30–40% | 1.30–1.40 yd³ | 20–30% | 0.70–0.80 yd³ |
| Rock, blasted | 50–60% | 1.50–1.60 yd³ | Negative — rock fill occupies more than bank | 1.2–1.3 yd³ |
These are the ranges published in earthmoving equipment handbooks for planning purposes. Blasted rock is the exception that proves the rule: it never returns to its bank volume, so a rock cut always produces more embankment than the cut volume suggests.
Neat-line volumes are not the whole job
Cross-section end areas describe the geometric difference between existing ground and design grade, and nothing else. A real earthwork quantity also has to account for topsoil stripping (usually 4–8 in over the whole disturbed area, stockpiled and respread, and normally excluded from the fill because organic material cannot be used in structural embankment), unsuitable material removed below grade and replaced, subgrade preparation and any over-excavation, pavement and base thicknesses that occupy part of the section, and settlement of soft foundation soils beneath the embankment.
Each of these is a separate quantity added to or subtracted from the neat-line figures. On soft ground the settlement allowance alone can add several percent to the fill. Treat this calculator's output as the geometric core of the estimate, not the estimate.
Mistakes that ruin an earthwork estimate
- Multiplying by the shrinkage factor instead of dividing. To build 100 yd³ of fill at 15% shrinkage you need 100 ÷ 0.85 = 117.6 bank yd³, not 85.
- Using swell where shrinkage belongs. Swell sizes trucks and stockpiles; shrinkage sizes borrow. They are different numbers answering different questions.
- Averaging a cut area with a fill area. They are separate accumulations. A section that is part cut and part fill contributes to both totals.
- Missing the transition stations. Where a section goes from all cut to all fill, insert a station at the zero point or both adjacent volumes are wrong.
- Stations too far apart on rolling ground. Average end area assumes the area varies linearly between sections. Add sections at every grade break.
- Forgetting topsoil. Stripping is normally excluded from structural fill and respread at the end, so it is a separate double handling, not part of the balance.
- Pricing haul by volume alone. Cost depends on haul distance and cycle time far more than on the number of cubic yards, which is what the mass diagram exists to reveal.
Other methods and when they beat average end area
Average end area is the standard method for linear work — roads, ditches, levees, pipelines — because cross-sections come free with the design. For an open site being graded to a surface rather than an alignment, two other methods fit better.
The grid or borrow-pit method lays a square grid over the site, records the cut or fill depth at every grid node, and computes the volume as the grid cell area times the average of the corner depths, weighting each node by how many cells it touches. It is well suited to parking lots, building pads and pond excavations, and it is what most survey software uses for a surface-to-surface comparison.
The contour method measures the plan area enclosed by each contour on the existing and design surfaces and treats the contour interval as the station spacing, applying the same average end area arithmetic vertically instead of horizontally. It is the fastest way to volume a stockpile or a pond from a topographic map.
Modern practice computes all three from a triangulated surface model, comparing the design TIN against the existing TIN directly. That is more accurate than any hand method, but it hides the arithmetic — which is exactly why running the numbers by hand at a few stations remains the best check on a software quantity that looks wrong. For quantities that follow the grading, see the trench excavation volume calculator for utility work, the paver base calculator for hardscape, and the topsoil tonnage calculator for the stripping and respread quantity.