Deck Railing Spacing Calculator

A spacing tells you how far apart to put two balusters. It does not tell you where to put the first one, and that is the whole problem. Start at one post, step off 5 ⅜" at a time, and you arrive at the far post with whatever is left — a gap that matches none of the others and is the first thing anyone looks at.

This calculator runs the arithmetic backwards. It counts the openings rather than the balusters: 12 balusters in a run make 13 openings, every one the same width, and that width is the run less the balusters, divided between them. Raise the count by one and every gap narrows. The first count whose gaps come in under the sphere is the answer, and it is the smallest one that works — at 68", 11 balusters would leave 4 5/16", which the sphere passes.

Then it gives the marks. Every centre is measured from the same post face — 4 ⅝" for the first, 9 15/16" for the second, on to 63 ⅜" for the last. Measuring each baluster from the one before it is how a run picks up a sixteenth of an inch at every step and finishes visibly wrong at the far end. One tape, one datum, 12 marks.

Two numbers describe the same railing and they are not interchangeable. On centre is measured baluster centre to baluster centre. Clear is the air between them — the gap a 4" sphere has to fail to pass.

At the default run those are 5 ⅜" and 3 ⅞". Both correct, 1 ½" apart, and only one of them is the one the code measures.

The expensive version of the mistake is reading the on-centre figure as a gap. Space 1 ½" balusters at 4" on centre and the air between them is 2 ½" — legal, and 42% more balusters than the run needs. 17 where 12 pass, and every one of them bought, cut, drilled and finished.

Ask which number you are holding before you cut anything. If it came off a spacing, it is on centre. If it came off the code, it is clear.

The space between a post and the baluster beside it is an opening in the guard, and the 4" sphere (R312.1.3) applies to it exactly as it applies to the middle of the run. A 68" run with 12 balusters has 13 openings, not 11.

Leaving the two end gaps out of the arithmetic is the most common way a layout that measured right on paper fails at the post. It also fails where it is easiest to see: a wide gap in the middle of a run reads as spacing, and a wide gap against a post reads as a mistake.

Switch a section to the open side of the stairs and three separate provisions apply. The infill openings are measured against a 4 ⅜" sphere rather than 4" (R312.1.3 exception). The triangle at the foot of each baluster gets 6" of its own (R312.1.3 exception). And the guard height is measured from a line through the tread nosings rather than from a walking surface (R312.1.2). Three constants, because three provisions.

On this flight the same 68" run takes 11 balusters raked where it takes 12 level, at 4 5/16" clear against 3 ⅞".

The opening on a rake is measured horizontally, and the common claim that it is measured perpendicular to the rail is wrong for the railing most people build. Balusters go in plumb. Two plumb balusters present parallel vertical faces, and the shortest distance between two parallel vertical planes is horizontal whatever the rail above them is doing — a sphere simply slides along the rake until it finds the narrowest point, which is the horizontal one. The claim is true for balusters set perpendicular to the rail, and that is a different railing.

What the rake does change is where the marks go. Mark on the rail, not on the deck. The rail is longer than the ground beneath it, so the same balusters that sit 5 13/16" apart horizontally land 6 ¾" apart measured along the rail. Step off the horizontal figure with a tape held against the raked rail and every baluster lands short, the error accumulating down the flight.

The riser and tread depth the rake is worked from are the two numbers the deck stairs calculator returns. Take them off that page rather than measuring an angle with a protractor — this flight rakes at 30.7°, and both pages arrive at that figure from the same two dimensions.

Where the bottom rail crosses a step, the tread, the riser above it and the underside of the rail enclose a triangle. It is an opening, it has its own sphere — 6", R312.1.3 exception — and almost nothing checks it.

A sphere passes a triangle when its diameter fits the largest circle you can draw inside it, which for a right triangle is the two legs less the hypotenuse. The triangle here is the step triangle scaled by how far the bottom rail sits above the line through the nosings, so every inch the rail rises scales the opening with it.

With the bottom rail on the nosing line this flight passes 4 ⅜" — comfortably inside. Raise the rail 3" and the opening reaches 6 9/16", which fails. The calculator returns the point it turns: 2 ¼" above the nosings on this flight, and it recomputes when the riser or tread changes.

Running the bottom rail down on the nosing line is common precisely because it keeps this triangle at its smallest. If the rail has to sit higher, the opening is what closes it — an extra baluster does nothing, because the triangle is bounded by the rail and the step rather than by the infill.

A deck has several sections between several posts and they are rarely the same length. Lay each one out on its own and the gaps come out different. A 68" run gives 3 ⅞" clear; a 92" run gives 3 11/16". Both pass. They differ by ⅛", and the place you will notice is the corner where the two meet.

Lay out each section independently anyway. Forcing a matched gap across sections of different lengths means adding balusters to one of them for appearance, which costs material and buys nothing the code asks for.

If the difference is visually obvious, the fix is post positions, not baluster spacing. Equal openings between posts give equal gaps everywhere. Post positions are decided when the deck frame goes in, though, not when the railing does — so on an existing deck this is worth knowing and not worth acting on.

The alternative is to accept it, which most decks do. An eighth of an inch across a corner is not something anyone sees without two tapes and a reason to look.

A guard on the deck is at least 36" measured from the walking surface (R312.1.2). A guard on the open side of the stairs is at least 34" measured from a line through the tread nosings (R312.1.2). Same section of the code, two different datums.

The datum is what goes wrong. Measure a stair guard from a tread surface instead of from the nosing line and it reads tall by most of a riser, which is enough to build a guard that measures fine on your tape and short on the inspector's. The calculator asks which section you are laying out and checks against the matching datum rather than against a single number.

Guard height and handrail height are separate requirements that happen to share a number on the stairs. The deck stairs calculator holds the handrail — when it is required, how high, and what graspable means.

It lays out infill. Three things decide whether that guard stays where you put it, and it answers none of them:

• How far apart the posts may sit. A guard has to take a load, that load goes into the posts, and how far apart they can be depends on the post and its connection. That is a strength question and this site does not answer strength questions.
• How the guard is fixed to them. The connection is where guards fail, and it is hardware and fastening rather than geometry.
• Whether the infill itself is strong enough. This is the one most easily missed, because spacing looks like the whole requirement and it is not: the code asks the balusters and their connections to resist a load pushed horizontally against them, quite separately from the load on the top rail. A run can be spaced perfectly and still be fixed too lightly. We lay out the gaps; we do not check what holds them there, and no calculator on this site does.
• Cable and glass infill. Both have their own rules and their own hardware, and neither is laid out like a baluster run.

The baluster width the calculator starts with is the dressed width of a 2x2 — 1 ½", not two inches. Type your own if you are using something else; the field takes any width rather than offering a list of products whose dimensions this site cannot cite. The 5" figures and stock sizes on the stairs page are the same kind of thing, and are labelled the same way.