The fencepost, and which way bars run
Two errors account for most bad rebar takeoffs, and both are counting errors rather than arithmetic. The first is the fencepost: a 20 ft width at 18 in on center has thirteen gaps and fourteen bars, not thirteen. Divide and forget the plus-one and you are a bar short in every direction, every time.
The second is crossing the dimensions. Bars that run along the length are spacedacross the width — so the width, not the length, sets how many of them there are. It is easy to write down the right two numbers and pair them the wrong way, and because the result still looks plausible it usually survives all the way to the delivery. The working panel above spells out which dimension drove which count for exactly this reason.
Weight is the order, length is the layout
You place rebar by the foot and you buy it by the pound. ASTM A615 Table 1 fixes the nominal weights — 0.668 lb/ft for #4, 1.043 for #5, 2.670 for #8 — and those are what the invoice is built from. The consequence is that bar size dominates the delivered weight far more than the mat layout does: the same 20 × 30 slab that takes about 536 lb of #4 takes about 838 lb of #5, on an identical bar count and identical lineal footage.
A detail worth knowing about those dimensions: A615's own footnote states that the nominal dimensions of a deformed bar are equivalent to those of a plain round bar of the same weight per foot. They are equivalent-round figures back-calculated from weight, not measurements taken across the deformations. Put calipers on a #4 bar over its ribs and you will read more than 0.500 in, and nothing is wrong.
Bar numbering has a quirk too. From #3 to #8 the designation is eighths of an inch, so #4 is exactly half an inch. Above #8 the numbering switches to an area basis, which is why #9 is 1.128 in rather than the 1.125 in the eighths pattern would predict. Assuming the pattern continues is a small error on paper and a real one when checking a bar tag.
Laps are designed, not allowed for
This calculator asks for the lap length as its own input rather than folding it into the waste percentage, and that separation is deliberate. A lap splice is a designed quantity: it depends on bar size, concrete strength, whether the bar is epoxy-coated, the spacing and cover around it, and whether it sits in a top-cast position with a lot of concrete below it. The same #4 bar can call for very different laps in two different pours in the same building.
So the number in that field is a placeholder to be replaced with what your drawings say, and the overage percentage beside it covers only offcuts and damage. Burying laps inside a blanket 10% would quietly under-order on any slab long enough to need splicing, which is precisely where the error costs the most.
What this leaves out
It is a rectangular mat in one layer. It does not handle top and bottom mats as a pair, openings, thickened edges, corner bars, dowels, stirrups, ties or column cages, and it does not bend anything — bend and hook lengths add material this straight-bar model has no way to see. It also does not check that the spacing you entered is achievable around the aggregate, or that your cover meets the exposure requirement.
For the concrete going around the steel, use the concrete bag calculator for bagged work or the Sonotube calculator for round piers, and the gravel calculator for the base course underneath.