Methodology

Four levels of confidence

Not all numbers are the same kind of number. A maximum joist span is a code requirement. A waste factor is a rule of thumb. Presenting both with the same authority would be dishonest, so every value in our data layer is tagged with one of four confidence levels, and pages say which they are showing.

• Code — read directly from a published building code or a code-referenced standard. Cited on the page that shows it.
• Derived — calculated from a code value by a rule we state explicitly. Footing bearing area is tributary load divided by soil bearing capacity; both inputs are code, the division is ours.
• Typical — a defensible industry figure where no single authority exists. Stain coverage is the clearest example.
• Estimate — a planning figure, usually a price. Volatile, dated, and labelled as such wherever it appears.

The build enforces this. A constant without a source and a confidence level fails the data guard, and a component holding a number of its own fails the build guard. Neither is a matter of discipline; both are checked mechanically on every deploy.

Structural numbers, and what we have not finished checking

Span thresholds decide whether a deck holds people up, so they get treated differently from everything else here. Our data layer currently holds 120 joist and beam span thresholds, and 4 structural records are still awaiting a second reading against the published document.

No calculator that renders an unverified structural number will ship. The build fails rather than publishing one. We would rather have fewer calculators than a fast wrong one.

We store threshold values — the maximum span for one species group, one member size, one spacing — with attribution, and render them as calculator output. We do not reproduce the source tables, and you will not find a copied span table anywhere on this site. If you want the tables, get them from the source: AWC DCA 6 is free.

Design loads

Every structural calculation on this site assumes the standard residential deck case unless you tell it otherwise:

These take a deck outside the prescriptive tables entirely, and our calculators say so rather than quietly returning a number:

Soil bearing capacity

Footing size comes from load divided by what the ground under it can carry. Where a soil test is not available, these presumptive values are what the code allows you to assume.

Every row above is entered from IRC Table R401.4.1 and none has been read back off the published table by a second pair of eyes, which is what the last column records. They are here because this page exists to show you what we hold and how sure we are of it — not because a calculator uses them. None does: no page on this site currently computes a footing size.

Board dimensions and gaps

Nominal lumber sizes are names, not measurements. Every quantity calculation here works from dressed dimensions, which is why our board counts sometimes disagree with calculators that use the number on the label.

Board coverage also accounts for the gap between boards, and the gap depends on the material. Wet-treated pine is laid nearly tight because it shrinks as it dries; everything else is laid with a deliberate gap because it will not.

Species and materials

Hardness and natural decay resistance come from the USDA Forest Products Laboratory Wood Handbook, which is public domain. Service life and maintenance cycle are industry typicals for a well-maintained deck.

Fasteners and ledger fastening

Screw counts are stored as a rule, not as a per-square-foot lookup table. A table goes stale the moment someone picks a different board width or joist spacing; the rule reproduces any of those rows on demand and cannot drift out of step with the board dimensions above.

screws per sq ft = 2 × (12 ÷ board coverage width in inches) × (12 ÷ joist spacing in inches)

Face-screwed decking takes 2 fasteners at every joist crossing and 2 extra at a butt joint. Hidden clips take 1 per board per joist. Screw length is governed by penetration — the screw has to reach at least 1 ½" into the joist below the board:

Ledger fastening is the one connection on a deck that fails catastrophically rather than gradually, and it is the one thing on this page we will not give you a number for. The spacing depends on the joist span the ledger carries, and the schedule that sets it belongs to AWC DCA 6 — Prescriptive Residential Wood Deck Construction Guide (2015 edition) , which is free to download. We had that schedule on this page as a table. It came down for two reasons: reproducing a published table is exactly what this site says it never does, and not one of those values had been read back off the document by a second pair of eyes. A figure that decides whether a deck stays on a house is the last place to carry either of those.

Stain coverage

There is no standards body for how far a gallon of deck stain goes, so every figure below names the product sheet it was read off and links to it. Coverage is per gallon, first coat, on prepared wood in sound condition — the condition the sheets themselves assume.

Chemistry, not pigment level, is what moves the number. The trade sells stain on a scale from transparent to solid, and every deck stain calculator we have looked at models coverage on that scale alone. The sheets say it is the wrong scale. Behr's semi-transparent penetrating oil and Behr's semi-transparent stain and sealer sit at the same pigment level: the oil covers 350–400 sq ft per gallon in one coat, the acrylic 250–300 in two. Same pigment, 100 sq ft and a whole coat apart. So we split the list by how the finish sits on the wood, and we ask you which one you are holding.

Seven product sheets from three manufacturers is not the market. Where the calculator gives one number for a whole grade, that number is a synthesis of the sheets listed above and nothing more — it is not a standard, and the tin in your hand governs. Two things in that table are worth reading twice: the solid stain's range is unusually wide because a solid on rough old boards and a solid on smooth new ones are barely the same job, and its midpoint sits above the semi-transparent acrylic. That is not an error. It is the clearest evidence we have that pigment level does not predict coverage.

Clear sealers are the widest spread of the lot and the place our single figure is least safe. We publish 300 sq ft per gallon, the midpoint of Olympic WaterGuard's 250–350. Thompson's Waterproofer Plus is rated 150–200 — half as far. If you are buying a clear sealer, read the tin before you trust our number, because the two ordinary products on the same shelf disagree by a factor of two.

Expected service life — 1–2 years for a clear sealer, up to 4–6 for a solid — is trade practice rather than a manufacturer warranty figure, and the calculator does not compute anything from it. Prices come from our own retail survey, not from the sheets.

Two adjustments then apply, and they are the reason most stain estimates come up short. The first is surface condition. No manufacturer publishes this as a number — the sheets say coverage will be lower on rough or weathered wood and stop there — so the multipliers below are ours, marked as estimates in the data layer, and printed here so you can disagree with them:

The second adjustment is the one we got wrong until we went and read the sheets, so it is worth being blunt about. We had the second coat of a film-forming stain covering 1.4 times the ground of the first. Every sheet we checked says 2.0 — Behr's semi-transparent acrylic at 250–300 then 500–600, Behr's solid at 200–400 then 400–800, Behr DECKplus the same, KILZ's semi-transparent at 250 then 500. Four products, three manufacturers, all exactly double, and Behr's sheets explain the mechanism: the second coat is meant to go on thinner than the first. At 1.4 we were putting close to half a coat of stain in your trolley that you did not need. The ratio is now stored on each product rather than as one figure for the whole site, because a penetrating oil does not have one — its sheet specifies a single coat and publishes no second-coat rate, so if you ask for extra coats on an oil we cost them at the full rate instead of inventing a discount.

What we deliberately do not add

We do not add surface area for board edges and gaps, and it is worth saying why, because it makes our figure lower than it would otherwise be and lower than some competitors'. Behr rates its semi-transparent acrylic at 250–300 sq ft per gallon on a prepared deck. That is measured on real decking, in plan, gaps and all — not on an imaginary solid sheet of wood — so the published figure has already absorbed the edges. Our surface multipliers then handle extra real surface by reducing coverage, never by adding area, so putting 12% of edge area on top would count the same thing twice through a second mechanism. The figure is still in the data layer, unapplied, with that reasoning attached: it stands only for somebody deliberately flooding the gaps with a brush or a gap tool, which is not the job most people are doing.

When to apply it

More finishes fail from being applied at the wrong moment than from being the wrong product. There is no single window — the products genuinely disagree, and by a lot — so here is what each sheet actually says rather than one set of numbers averaged into uselessness.

Across those two sheets the floor sits somewhere between 40°F and 50°F, the ceiling at 90°F, and the rain-free stretch you need runs from 4 hours to 24 hours. A six-fold spread on the rain figure is exactly why a single number would not have helped you. Read your own tin — and remember that air temperature is not surface temperature, because a deck in direct sun runs far hotter than the air around it.

Two more figures, both from the Behr penetrating oil sheet: it is dry to touch in 2–4 hours, ready for foot traffic at 24, and fully cured at 72. And after washing the deck, give it at least 24 hours to dry before you coat it — longer if you soaked it rather than damped it.

New pressure-treated lumber is its own case, and the usual advice — "wait until it is dry" — is not something you can put in a calendar. Thompson's publishes a number: 30 days before sealing new pressure-treated stock. Kiln-dried-after-treatment boards may be ready sooner and soaking-wet ones later; the practical check is to sprinkle water on the boards, and if it beads rather than soaks in, the wood is not ready for you.

Stairs and guards

Stair geometry has the tightest tolerances on the whole deck. The riser-variation rule is the one people fail inspection on, and it is three rules rather than one: the same ⅜ inch allowance applies separately to risers, to tread depths and to nosing projections, and a flight can pass on one and fail on another. It is also not a tolerance per step — it is the difference between the largest and the smallest in the entire flight.

Every value below is from the 2021 IRC and carries the section it came from. The IRC is a model code. Your jurisdiction adopts an edition — often an older one — and may amend it, so these are the values to design against and your building department has the ones you will be inspected against. Two things in particular move: section numbering (riser height was R311.7.4.1 before 2018 and is R311.7.5.1 from 2018 on) and individual provisions (Pennsylvania's Uniform Construction Code amends the nosing requirement). One of the documents cited below shows that amendment beside the model text, which is why it is cited at all.

The code itself is copyrighted, so we store threshold values with their attribution and never reproduce a table or a section's text. The values were read off building-department handouts that quote the provisions — public documents published by authorities having jurisdiction — and each record names the ones it came from: City of Boise — Treads and Risers (handout 437, March 2022) , Town of Timnath, Colorado — 2018 Residential Stair Guide , Town of Essex, Connecticut — Residential Stairways and Ramps (2015 IRC) , weknowcodes — Residential Stair Geometry (R311.7.5, with Pennsylvania UCC amendments noted) .

A stairway with 4 or more risers needs a handrail (R311.7.8), which is a shorter flight than most people expect.

Waste factors

Two different things get called waste and conflating them is why material lists come up short. Offcut waste is the ends you cut off because stock lengths do not divide evenly into your deck. Layout waste is extra material a layout demands — a 45-degree run needs longer boards for the same deck and every board is a compound cut. Our factors cover both.

These add on top:

• Picture-frame border — 3%. Mitred perimeter board. Adds material and requires extra blocking under the mitres.
• Breaker board — 2%. A board run across the field so long runs can be butted neatly rather than staggered.
• Random-length stock (typical for ipe) — 5%. When you cannot choose your board lengths, you cannot optimise the cuts.
• First deck / learning curve — 5%. If this is your first deck, buy the extra. Everybody miscuts something, and matching a board four weeks later is worse than storing a spare.

These are trade rules of thumb. There is no standards body for deck waste. We publish the exact figures so you can disagree with them, and if you have better data we would genuinely like to see it.

Regional adjustment

Where a calculator accepts a ZIP code, it is always optional and always offered after a complete result. It never gates anything.

We apply two multipliers rather than one. Lumber is close to a national commodity and barely moves between Ohio and Oregon; labor moves enormously. Blending them into a single "cost of living" number is how national-average deck costs end up wrong everywhere. Our data covers 29 regions, against a united states average baseline of ×1.00.

The same records carry typical frost depth and ground snow load, because those change what you build and not just what you pay. Both are regional typicals for planning. Frost depth in particular is set by your local jurisdiction, and the only correct source for it is your building department.

Costs, labor and prices

Prices were last surveyed in August 2026. That date is a single field in our data layer, stamped onto every priced record and rendered visibly on every calculator that outputs dollars. We re-run the survey every 6 months, and a build attempted with prices more than 9 months old fails outright rather than deploying stale figures.

Labor rates are what a competent contractor charges, not the cheapest bid on a lead-generation site. Lowballing labor makes DIY look better than it is, which is the opposite of useful when you are deciding whether to take a project on. The DIY hours column is what the same unit of work costs you in time.

Concrete, by how it is bought:

Project costs people forget until the invoice arrives. The permit is not optional, whatever the internet tells you — an unpermitted deck surfaces when you sell, and it can force a teardown:

All prices are national planning figures before the regional multipliers above. They are not quotes, they are not offers, and they do not bind any supplier or contractor.

Sources

Codes and standards

• International Residential Code (IRC), 2021 edition — Model code. Your jurisdiction may adopt a different edition or amend it.
• IRC Table R301.5 — Minimum uniformly distributed live loads
• IRC Table R401.4.1 — Presumptive load-bearing values of foundation materials
• IRC R311.7 — Stairways — The authority. ICC publishes it read-only at the URL above; the values on this site were read off the building-department handouts that quote it, which are named in each record's derivedFrom.
• IRC R312 — Guards — Same route as R311.7: the authority is the code, the documents actually read are the handouts named in derivedFrom.
• IRC R507 — Exterior decks
• AWC DCA 6 — Prescriptive Residential Wood Deck Construction Guide (2015 edition) (read 2026-09-12) — Free from the American Wood Council, and the prescriptive basis for most residential deck permits in the US. Every span, overhang, cantilever and stringer figure citing it was read back off the 2015 edition on the date above, and each record names its table and page. Pages matter here because the document numbers two different tables 3A: the beam spans on page 6 and the joist-hanger capacities on page 9. Southern Pine design values changed in June 2013, so the 2012 edition disagrees with this one for that species; municipal mirrors of the 2012 edition are common, and a value that does not match this site should be checked for edition before anything else.

Other references

Manufacturer pages change without notice and without a version number, so each one carries the date we last read it. If the sheet now says something different from what we have stored, the sheet is right and we want to hear about it. Two entries below are still aggregations rather than documents — fastener and composite guidance — and we have not yet done to them what we did to the coating figures. They are marked as aggregations for that reason rather than quietly presented as citations.

• City of Boise — Treads and Risers (handout 437, March 2022) (read 2026-08-29) — Covers R311.7.5 — risers, treads and nosings. The document does not state which IRC edition it quotes, so it corroborates the values rather than fixing the edition; the section numbers here are 2021 numbering.
• Town of Timnath, Colorado — 2018 Residential Stair Guide (read 2026-08-29) — Covers R311.7 and R312 together, which is why it backs both the stair and the guard limits. R311.7 and R312 numbering is unchanged between 2018 and 2021 for every value stored here.
• Town of Essex, Connecticut — Residential Stairways and Ramps (2015 IRC) (read 2026-08-29) — An older edition, kept deliberately. It is evidence that section numbering and some values move between editions, which is the reason this site never states a limit without naming the edition it came from.
• weknowcodes — Residential Stair Geometry (R311.7.5, with Pennsylvania UCC amendments noted) (read 2026-08-29) — Shows the Pennsylvania UCC amendment to the nosing provision alongside the model text. Kept as direct evidence for the local-amendment caveat: a reader in Pennsylvania is not governed by the model value.
• ALSC Policy for Evaluation of Recommended Spans for Span Rated Decking Products (November 5, 2004) — What DCA 6 (2015, DECKING REQUIREMENTS, p. 3) defers to for decking spans, and what a span-rated decking stamp is issued under. DCA 6's Commentary (p. C3) says the policy assigns "usually 16" or 24"" and is not meant for spans above 24". The policy itself has not been read; nothing citing it is verified.
• American Softwood Lumber Standard PS 20 — nominal vs dressed sizes — Why a 2x8 is actually 1½ by 7¼ inches.
• USDA Forest Products Laboratory, Wood Handbook (FPL-GTR-282) — Public domain. Source for hardness and natural decay resistance.
• AWPA U1 — Use Category System for treated wood — Defines which treatment retention is required for ground contact vs above ground.
• ACI 332 — Residential concrete construction
• NOAA / state frost depth records, as adopted by local jurisdictions — Frost depth is set locally. Our figures are regional typicals, not a substitute for asking your building department.
• ASCE 7 ground snow load maps
• US Bureau of Labor Statistics — regional price and wage differentials — Basis for the shape of our regional multipliers, not the exact figures.
• Behr Premium Semi-Transparent Waterproofing Stain & Sealer — product page (read 2026-08-25) — 100% acrylic, film-forming. 250–300 sq ft/gal first coat, 500–600 second, two coats required.
• Behr Premium Solid Color Waterproofing Stain & Sealer — product page (read 2026-08-25) — 100% acrylic, film-forming. 200–400 sq ft/gal first coat, 400–800 second; two thin coats recommended.
• Behr Premium Semi-Transparent Waterproofing Wood Finish, Penetrating Oil — product page (read 2026-08-25) — Penetrating oil. 350–400 sq ft/gal, one coat. Apply 40–90°F; resists rain 4 hours after application. Dry to touch 2–4 hours, foot traffic 24 hours, full cure 72 hours.
• Behr Premium Transparent Waterproofing Wood Finish, Penetrating Oil — product page (read 2026-08-25) — Penetrating oil. 350–400 sq ft/gal, one coat. Same chemistry as the semi-transparent oil at a lower pigment load, and the same coverage — which is the point.
• Olympic WaterGuard Clear Wood Sealer — product page (read 2026-08-25) — 250–350 sq ft/gal.
• Thompson's WaterSeal Clear Wood Sealer — product page (read 2026-08-25) — Surface and air above 50°F during application and for 48 hours after. Do not apply if rain is forecast within 24 hours. Allow 30 days before applying to new pressure-treated lumber.
• Thompson's WaterSeal Waterproofer Plus Clear Wood Protector — retail listing (read 2026-08-25) — 150–200 sq ft/gal per manufacturer support — materially below the other clear sealers, which is why we publish the spread rather than one number.
• Deck fastener manufacturer installation guidance (aggregated) — No single document. Still to be replaced with named products the way the coating figures were — until then, treat these as typicals rather than citations.
• Composite decking manufacturer published product dimensions (aggregated) — No single document, and composite profiles are not standardised. Always check the specific product.
• Established deck-building trade practice — No single authority. Our exact figures are published on the methodology page so you can disagree with them.
• National retail price survey (DeckCalculators.com) — Volatile. Every priced record carries the date it was last checked.

The model code is not your code

The IRC is a model. Your jurisdiction adopts an edition of it, often amends it, and sometimes replaces sections wholesale. Local frost depth, snow load, wind exposure, seismic category and soil all vary, and your inspector has the final say regardless of what any of us calculated.

Use this site to arrive at the permit office already knowing roughly what you need. Do not use it instead of going.

Tell us we are wrong

If a number here does not match the source we cited, or your inspector told you something different, we want to know. Send the page, the number and what you believe it should be to our contact page. Corrections to structural data get looked at the same week.