When automatic layout is unreliable

Software can try thousands of arrangements in a second. It still can't be guaranteed to find the best one, and it only knows what you tell it.

It's a hard problem

Arranging irregular pieces on fabric with no overlaps, to minimize length, is the nesting or marker-making problem. It is NP-hard, which means no known method finds the best layout quickly for all inputs. Practical systems use heuristics and metaheuristics (Bennell & Oliveira, "The geometry of nesting problems: A tutorial", EJOR 2008; Luzzi, "Exact and Heuristic Methods for Nesting Problems", 2003). Even packing plain rectangles onto a strip is NP-hard in general.

What Bolt & Grain does, and doesn't

• It packs rectangles using MaxRects, trying 12 sort orders and placement rules and keeping the shortest result. That's a heuristic: a careful hand layout can sometimes beat it.
• Curved garment pieces are planned by their bounding rectangles. Real garment markers interlock curves (a sleeve cap beside an armhole), so this tool over-estimates fabric for garments. For garments, follow the pattern envelope's yardage chart or lay out the real pattern pieces.
• No bias placement, no fold layouts (cutting "on the fold" or two layers at once), no matching of plaids or stripes sideways across seams.
• It trusts your numbers. A wrong usable width or forgotten shrinkage makes every result wrong.

Trust your own layout when…

• the fabric has a flaw, a placement print or a border print;
• plaids, stripes or large motifs must line up across seams;
• the pattern gives a specific layout (especially "with nap" or fold layouts);
• pieces are cut on the fold or in mirrored pairs from folded fabric.

Use the planner to estimate yardage and test what-ifs, then do the final check on the real fabric.