Flooring waste by pattern: what 2,250 simulated floors actually measured

Across 2,250 simulated floors, the patterns split into two clear groups: linear lays (straight, running bond, random stagger, strip plank) wasted 5.8% to 8.7% on average, while the block patterns wasted a near-constant 20% to 25% (chevron 24.5%, basketweave 21%, windmill 19.9%). The biggest surprise: optimised herringbone averaged just 6.1%, nowhere near its 15% to 20% folklore range. Every number below was measured by running our layout engine over the same 250 rooms in every pattern, not copied from another waste table.

How much waste does each pattern really produce?

Waste here means the percentage of purchased board material that does not end up on the floor: rip offcuts, end offcuts and notch cut-outs all count. Each figure is the optimizer's best strategy for that room, so these are floors where every cut was planned before the first board was bought.

Two things stand out. The gap within the linear group (5.8% to 8.7%) is driven by how rows start, not by the pattern name. More on that below. And the block patterns cluster tightly around their averages: chevron's middle 50% spans just 1.5 points, because its waste is built into the geometry of the pattern, not into your room.

Why does optimised herringbone beat the 15% to 20% rule?

Herringbone's folklore range assumes the angled wall cuts are lost. In a 45° herringbone, every board that meets a wall is cut at 45°, and the offcut is a mirror image of a piece needed somewhere else along that wall or the opposite one. A layout engine pairs those pieces across the whole floor; a fitter cutting board by board almost never can, because the matching position may be thirty boards away.

Three honest caveats before you budget 6%:

• Board format matters. We simulated parquet-format boards (600 × 120 mm). Herringbone laid with long laminate planks wastes substantially more; a single test with 1285 mm planks came out near 15%.
• The pairings assume planned cuts. If you're cutting by eye, keep budgeting 15% to 20%.
• Chevron gets no such rescue (24.5%): its mitred ends are cut into every board before it ever meets a wall, so there is nothing to pair. The difference is explained in chevron vs herringbone.

How much does room size matter?

More than pattern choice, for linear lays. The wall cuts are the waste; walls grow with the perimeter while the floor grows with the area, so small rooms pay proportionally more:

A room under 12 m² wastes roughly twice the percentage of a room over 25 m² in every linear pattern. If you're pricing flooring for a hallway or a small bedroom, the waste line in your budget matters more than it does for an open-plan floor, and generic "add 8%" advice under-buys exactly there.

Do L-shaped rooms and obstacles increase waste?

Far less than folklore says, if the layout is planned. Across the corpus, L-shaped rooms wasted within half a point of plain rectangles in almost every pattern (straight lay: 7.3% vs 7.1%). A planned layout treats the L as two runs that share offcuts, exactly as described in flooring an L-shaped room.

The exception: rooms with a fixed obstacle (we simulated a 1.2 m chimney breast) pushed the offcut-starter strategy from 8.5% to 10.4%, because the obstacle interrupts rows, and interrupted rows hand you offcut lengths that fight the stagger rules.

Which row-start strategy wastes least?

The same random-look floor can start its rows two ways, and they measure differently:

The offcut method is the classic thrifty advice, and when nothing checks your joints, it does save material. But under enforced craftsmanship rules (minimum 300 mm joint offset, no H-joints, no staircases), the offcut the room hands you regularly lands too close to a joint two rows back, and the fix, trimming the starter, turns the saved offcut into waste plus a trim. Randomised starters chosen with the rules in view kept both the look and the material. The trade-off is the subject of the random ashlar guide.

How we measured it

• Corpus: 250 simulated rooms (137 rectangles, 75 L-shapes, 38 rooms with a fixed obstacle), areas 6 to 35 m², aspect ratios 1.0 to 2.2, generated from a fixed random seed.
• Patterns: all nine of the designer's generators, default parameters, run on the identical room set (2,250 layouts total).
• Boards: 1285 × 192 mm planks for the linear patterns; 600 × 120 mm parquet-format boards for herringbone, chevron, basketweave and windmill. Minimum piece length 300 mm, minimum rip width 50 mm.
• Waste definition: share of purchased board material not on the floor, for the optimizer's best-ranked strategy per room, with craftsmanship rules enforced (joint offset, no slivers, no H-joints).
• Not modelled: saw kerf (≈0.2% on a full-length plank), miscuts, damaged boards, batch and grain matching. Real projects should add a buying margin on top; see how much flooring do I need.

The dataset regenerates from a seeded script against the same engine that plans real rooms, and we refresh it annually. To see where your own room lands in these distributions, draw it in the designer: the layout, the waste percentage and the full cut list for your floor is what the first free project computes.