Deck calculator

Every deck calculator on the internet does the same thing: area divided by board width gives linear feet, add ten per cent, done. It is a reasonable-sounding number and it is wrong in both directions at once.

Decking is not sold by the linear foot off a roll. It comes in fixed lengths — eight, ten, twelve and sixteen feet for treated pine, and only twelve, sixteen and twenty for most composite — and a row of decking is covered by one of those lengths or it is not. The offcut is the waste, and the offcut depends entirely on how your run length divides into what the yard stocks.

Read down that table and the point makes itself. A deck at fourteen feet wastes a seventh of its decking in pine and a seventh in composite; at sixteen feet both are zero. Nothing about the deck changed except two feet of a line on a drawing.

If the deck is still on paper, this is the cheapest decision available to you, and it is made before anyone has argued about the colour.

A deck is a floor built outdoors with nothing underneath it. The boards are the surface; the joists, beam, posts and footings are the structure, and on a treated deck they cost about as much as everything you can see.

Two numbers decide the frame: how far the joists have to span, which is the depth of the deck out from the house, and how far apart they sit, which is decided by the decking going on top.

Joist spacing looks like a framing decision and it is not. It is set by whatever is going on top, because the decking has to span between joists without flexing underfoot.

A 5/4 board — an inch thick when it is finished — is rated to sixteen inches and no more. A 2 × 6, half an inch thicker, will do twenty-four, which removes a third of the joists from the job. Composite is sixteen inches whatever its thickness, because it is considerably less stiff than wood and it creeps under sustained load.

And anything laid on the diagonal drops to twelve. A board crossing the joists at forty-five degrees spans about forty per cent further between supports than the same board laid square, which is a change of stiffness, not of appearance.

Almost everything that can go wrong with a deck goes wrong slowly and visibly. The ledger does not. It is the board bolted to the house carrying one whole edge of the structure, and when it lets go the deck comes away from the wall in one piece, fully loaded, with everybody standing on it.

Two things cause it. The first is nails. A nailed ledger relies on the shear strength of smooth shanks in end grain, and it is not adequate — the connection is through-bolts or structural lag screws on a specified schedule, staggered top and bottom.

The second is water. An unflashed ledger holds moisture against the sheathing behind it for its whole life. The bolts stay perfectly sound while the wall they are anchored into turns soft, and there is no way to see it happening.

The material question is almost always asked as a price per board, and answered that way. It is the wrong horizon. A deck is a thing you own for decades, and two of these materials ask for a weekend of your time every couple of years for all of them.

Treated pine at about $1.75 a foot needs cleaning and sealing every 3 years. Over 25 years that is eight cycles — eight weekends, or eight invoices. Composite at three times the board price asks for a wash.

The comparison panel runs the whole build in each material on your own dimensions, including the board waste, and then adds the refinishing twice: once at the cost of the materials, which is you on a Saturday, and once at what a contractor charges.

The result is less flattering to composite than the marketing, and more interesting. At nothing an hour for your own labour the treated deck usually stays slightly ahead for the whole twenty-five years — the refinishing materials are only a couple of hundred dollars a cycle, and they never catch up with three times the board price. Put a contractor’s rate on the same work and composite is clearly cheaper well before the end.

So the composite question is not really about money. It is about whether you intend to spend eight weekends of the next twenty-five years on your knees with a brush.

Deck stairs are the one part of the job with no tolerance in it. People climb stairs without looking, on a rhythm set by the first two steps, and a riser that differs from its neighbours is what a fall is made of. The code limit on variation across a flight is 0.375 of an inch.

Three things follow from that and are routinely missed. The run comes out further than people expect — nearly four feet for a three-foot deck — and it has to land on a footing or a pad, not on soil. The stringers go at no more than 16 inches apart for wooden treads and 12 for composite ones, because a composite tread bends underfoot where a 2 × 6 does not. And the total rise is measured to the finished surface, so if you are adding decking over an existing frame, measure after.

Above 30 inches of drop a guard is required, 36 inches high, with no gap anywhere that a 4 inch sphere can pass through. That last rule is about a child’s head, which is why it applies to the gap under the bottom rail as well as between the balusters.

With a standard 1.5 inch baluster that puts the centres at 5.5 inches at the absolute most, and it is worth working slightly tighter: the last gap before a post is where an evenly spaced run usually goes over.

The part nobody thinks about is the load. A guard has to resist a substantial horizontal force at the top rail — a person falling against it, not leaning on it — and all of that force ends up in the connection between the post and the frame. A post bolted through the rim joist with blocking behind it holds. A post screwed to the face of a rim joist is decoration.

Split a deck estimate and the lumber divides almost evenly between the decking you can see and the frame you cannot — and then labour dwarfs both. Railing and stairs add a surprising amount on top, because both are slow work on a small quantity of material.

The frame being about half the lumber is the part that changes decisions. Moving from treated pine to composite raises the total by much less than the board price suggests, because the joists, beam, posts, footings and the days of work underneath do not change at all. The showroom compares boards; the estimate compares decks.

Only labour is adjusted for where you are. Lumber costs much the same nationally, with a freight allowance for the weight, and it moves with the commodity market far more than with geography; what changes between states is the hour. The regional price index is the figure being applied.

Joist and beam spans, the overhang limit and the ledger schedule follow the wood council’s deck guide; the guard height, sphere rule, riser and tread limits follow the residential code; joist spacing for composite follows the manufacturer’s own installation instructions, which are stricter than the code and are what voids the warranty. Board lengths are what shops actually stock, which is the whole basis of the waste figure. Everything is on the sources page, dated.

Three assumptions worth stating. The deck is taken as a rectangle attached to the house along its length, with a beam and posts at the outer edge — the common case, and the one the span tables are written for. Butt joints are avoided wherever a single board covers the run, because a joint costs blocking and shows; the jointed alternative is offered separately when it saves enough to be worth the trade. And the lifetime figure prices refinishing as materials only, on the assumption you are doing it yourself.

Nothing you type is sent anywhere — the whole calculation runs in your browser.

Ledger and span are the two that decide whether the deck is safe; the rest is arithmetic and taste.