Ledger Pull Test vs Visual vs Replace Decision Tree

Why this matters

Ledger failure is the single largest cause of catastrophic deck collapse in the United States, and the CPSC has tied dozens of deaths to nailed ledgers or rotted rim attachments. A field call on an older deck rarely arrives labeled "the ledger is failing" - it arrives as bounce, gap at the siding, or a homeowner who saw water behind the band board. The decision the senior carpenter has to make in the first ten minutes is whether the existing ledger is safe to leave alone, safe to retrofit with bolts and flashing, or unsalvageable. Get that wrong and the next family barbecue ends in a 911 call. This tree walks the field-probability order from "what you can see in 30 seconds" to "what the IRC R507 lateral-load math actually requires."

The decision flow at a glance:

  Ledger - pull test, visual, replace?
  |
  +-- 1. Nailed + gap / no flashing? ---> REPLACE, NO
  |                                       TEST
  |
  +-- 2. Lags, <12 yr, no movement? ----> VISUAL
  |                                       FASTENER AUDIT
  |
  +-- 3. Spacing past table row? -------> RETROFIT
  |                                       THROUGH-BOLTS
  |
  +-- 4. Movement reported? ------------> LATERAL PUSH
  |                                       TEST
  |
  +-- 5. Wrong row read (long joist)? --> TIGHTER
  |                                       SPACING NEEDED

Symptom presentation at arrival

Customers describe one of four things. First, "the deck moves when we walk on it" - usually felt at the house side, not the rim, and worst near the door threshold. Second, "there is a gap between the deck and the siding" measured anywhere from 1/8 inch to a visible daylight crack. Third, "water is getting behind the siding" with stains on the basement ceiling or interior drywall above the ledger. Fourth, the call is a pre-purchase inspection follow-up where another inspector flagged the attachment. The first two are mechanical-attachment symptoms, the third is flashing-failure, and the fourth is a documentation exercise. Treat them differently from the first question.

Quick checks before any tool comes out

Walk the perimeter and answer five questions on your tablet before you touch anything.

• Is the ledger attached to wood-frame house, brick veneer, stucco-over-frame, or solid masonry? Brick veneer attachment is non-code without an engineer's letter (IRC R507.9 does not permit a deck ledger to be attached to or supported by masonry or stone veneer; veneer is nonstructural cladding and carries no deck load), and that single answer can short-circuit the entire visit toward "replace with freestanding posts or pull engineer."
• What fastener heads are visible from below at the ledger-to-rim joint? Nail heads (round, no driver pattern) - high-probability replacement. Lag screws (hex head, no washer) - retrofit candidate. Through-bolts with washers and nuts on the interior - likely passes inspection.
• Is there visible flashing - metal Z-flashing, peel-and-stick membrane, or nothing? "Nothing" plus stucco above is a near-guaranteed water-intrusion path.
• How old is the deck? Pre-2009 decks predate the IRC's prescriptive ledger-fastener table entirely; it entered with the 2009 edition and now lives at R507.9. Pre-2000 decks predate widespread code awareness entirely.
• Are joist hangers nailed with roofing nails, common nails, or hanger nails (1.5 inch 10d or 16d Simpson SD)? Roofing nails through Simpson hangers is an automatic remediation regardless of ledger condition.

Isolation tree - which test to run

Branch 1: visible nailed ledger with daylight gap or no flashing. Skip the pull test. The ledger is non-conforming on its face. Recommend replacement with through-bolts at the IRC R507.9 table spacing for the actual joist span, or an engineered hold-down system. Document with photos.

Branch 2: lag-screwed ledger, no visible rot, deck under 12 years old, no movement reported. Run a visual fastener audit - count lags, measure spacing, check for washers under heads.

Read the spacing off the actual table row for the actual joist span, and read it in the right direction: spacing gets TIGHTER as joist span gets LONGER, because each lag is carrying tributary load from a longer joist. For 1/2 inch lag screws into a 2 inch nominal band joist, the IRC table runs roughly 30 inches o.c. at a 6 foot joist span, 23 inches at 8 feet, 18 inches at 10 feet, and 15 inches at 12 feet, continuing tighter from there. Grabbing the 8 foot row for a 12 foot deck leaves you about a third short on fasteners, and that specific one-row slip is a documented way ledgers come off houses. Confirm the row against the table in the adopted IRC before you sign anything.

Two base conditions ride with that table and both get dropped in the field: the values are for a 2 inch nominal band joist of the listed species, and the lag and bolt columns are different numbers. Do not read a bolt spacing and install lags at it.

If field spacing exceeds the table row for the actual span, or washers are absent, retrofit with through-bolts at code spacing. If spacing meets the table and lags have washers, leave alone but inspect flashing.

Branch 3: deck movement reported, no visible gap, lags present. Run a manual lateral push test - apply 50 to 75 lb horizontal force at the rim mid-span and measure deflection at the ledger-to-rim joint with a dial indicator or paper shim. Movement over 1/16 inch under hand-pressure indicates fastener pull-through or wood crushing behind the lag. Escalate to drill-and-probe.

Branch 4: stains on interior, no movement. The mechanical attachment may be fine; the failure is flashing. Probe the rim band joist with an awl through any access point (basement, crawl, or removed siding). Spongy band joist plus mechanical attachment intact is still a replacement call because the wood holding the lags has lost strength.

Confirming diagnosis - the drill-and-probe sequence

When branch 2 or 3 ends ambiguous, pull two lag screws on opposite ends of the ledger run. Insert a 3/8 inch drill bit into the lag hole and feel for resistance. Crisp resistance with dry shavings - sound wood, retrofit acceptable. Soft punky resistance, no shavings, water comes out - the rim joist behind the ledger is rotted. The ledger comes off, period. Pull a third lag at mid-span to confirm uniform vs localized rot. Localized rot at one end may allow a partial repair with sister blocking; uniform rot is a full ledger tear-off and rim sistering. Measure embedment - lags engaging less than 1.5 inches of solid wood in the band joist do not meet the IRC ledger provisions at R507.9 regardless of spacing.

For a fully documented call, follow with a Resistograph or moisture-meter survey. Pin moisture meters reading over 28 percent in band-joist wood indicate active wetting; under 19 percent is dry. Anything in between needs root-cause flashing fix even if you save the ledger.

Remediation by branch

Replace: full tear-off, rim sistering with new PT 2x kiln-dried-after-treatment lumber, peel-and-stick + cap flashing per AAMA 711, new ledger fastened with through-bolts and washers at the R507.9 table spacing for the joist span actually built, lateral-load device per R507.9.2 (Simpson DTT2Z or equivalent ICC-ES ESR connector at minimum two locations). Retrofit: add through-bolts adjacent to existing lags, do not remove the lags. Install Simpson DTT2Z lateral connectors per ICC-ES ESR-2611 to satisfy R507.9.2. Install or replace flashing - this is mandatory even on a "retrofit only" job. Visual sign-off: only when fasteners, spacing, and embedment all pass and band joist probes as sound. Document with three photos minimum and a written letter to the homeowner.

References

• IRC 2021 Section R507.9 - Deck Ledger Connection to Band Joist, including the fastener spacing table (spacing tightens as joist span lengthens; separate columns for lags and for bolts) and R507.9.2 for lateral load connection
• IRC 2021 Section R507.9 - no ledger attachment to or support by masonry or stone veneer
• AWC DCA 6, Prescriptive Residential Wood Deck Construction Guide - the same ledger fastener table in a working format
• ICC-ES ESR-2611 - Simpson Strong-Tie DTT2Z deck tension tie acceptance criteria
• AAMA 711-20 - Voluntary Specification for Self-Adhering Flashing Used for Installation of Exterior Wall Fenestration Products
• ASTM D5764 - Standard Test Method for Evaluating Dowel-Bearing Strength of Wood and Wood-Based Products. Note what this is and is not: dowel-bearing strength underpins LATERAL (shear) connection capacity. Lag and bolt WITHDRAWAL design values come from the NDS and are keyed to species specific gravity, not from D5764.
• US CPSC Publication 5429 - Deck Collapse Prevention guidance and field inspection checklist