Why decks fall off houses — and when we refuse to use a ledger board.

The one board holding everything up

A typical attached deck hangs half its weight on a single piece of lumber: the ledger, bolted against the house. Every joist connects to it. Every person at your cookout loads it. If the ledger lets go, the deck doesn't sag or creak in warning — it rotates away from the house and drops, all at once, usually when the most people are standing on it.

Forensic investigations of deck collapses point overwhelmingly to the same culprit: not broken joists, not snapped posts, but the ledger connection — either pulling away from the house or rotting silently behind its own siding. Researchers who study these failures have noted for years that the majority of catastrophic collapses trace back to this one connection. That's why modern code devotes an entire section to it.

The four ways ledgers fail

1. Nails instead of bolts. For decades, decks were nailed to houses. Nails are strong in shear but nearly worthless in withdrawal — and a loaded deck constantly works its fasteners outward. Today's residential code (IRC R507) prescribes a specific schedule of ½-inch lag screws or through-bolts. A nailed ledger isn't a lesser deck; it's a countdown.

2. No flashing. The ledger sits in the worst moisture position on your house: a horizontal wood surface pressed against the wall, catching every drop of rain that runs down the siding. Without properly layered flashing, water wicks between the ledger and the rim board and rots both — from the inside, where you can't see it. A rotted rim board means the bolts are gripping mush. The deck looks perfect until it doesn't.

3. Attached to something that can't carry it. Brick veneer looks structural. It isn't — it's a decorative skin, and code prohibits hanging a deck ledger on it. The same goes for cantilevered floors, bay overhangs, and some manufactured-home rim details. A ledger is only as strong as what it's bolted to, and on many houses, what's behind the siding was never designed to carry a deck full of people.

4. No lateral connection. People on a deck don't just push down — they sway, rock, and pull the deck away from the house. Code now calls for dedicated lateral-load connections (hold-down tension devices tying deck joists to the house floor structure) precisely because downward-strength alone has proven insufficient in real failures.

When we build freestanding instead

There's a second answer the industry doesn't talk about enough: skip the ledger entirely. A freestanding (self-supporting) deck carries its house-side load on its own beam and footings, standing millimeters from the house without touching it. Code explicitly permits this, and in several situations we consider it the only defensible choice:

On brick veneer homes, where a ledger is prohibited outright. On houses with cantilevered floors or overhangs, where there's nothing sound to bolt into. On older homes where the rim board's condition can't be verified without opening the wall. And whenever a homeowner wants zero penetrations through their siding and weather barrier — because every ledger bolt is a hole in your house's rain jacket, and flashing done wrong causes the very rot it's meant to prevent.

The cost of freestanding is honest: one more beam, two or three more footings — typically a few hundred to a couple thousand dollars depending on size. What it buys is the permanent elimination of the failure mode responsible for most deck collapses, and a house envelope with no new holes in it. On the wrong wall, it's not an upgrade; it's the only correct answer.