Skip to content

Where the theory applies
Hitch — object-dependent

Anchor Bend

Friction on hardware · Physical hitch theory

Does not pick a knot

This page does not pick a knot for you. DNA, molecules, quantum braids and vortices stay off Decide.

Finished structure

No drawing of this exact knot yet. The written steps are the reference.

Failure modes

When this family fails, it fails as these. The recovered end still outranks the name.

Jam-to-hardware family

A hitch that cinches to the ring or shackle and is not meant to break by hand.

  • !Second turn tied outside — two half hitches, not a jam
  • !Expected to untie after a snatch
  • !HMPE without a splice — the hitch strips the cover
Fails when
  • !The end is tucked under only one turn, and the knot eases open under a pulsing load
  • !Only one pass went through the ring, so the knot is plain half hitches with no turns clamping the end
  • !The half hitch is missing or loose, and the tucked end works back out of the turns as the rode swings
  • !The tail was left unseized on a rode that stays down, and over weeks of soaking and swinging the hitch loosens unseen
Geometry
  • Single pass, no round turn

    Hitch slips on the ring in a swell

    The round turn is the hitch · step 01

  • Working end missed a turn

    Tuck does not lock and the rode walks

    The working end must pass through both turns · step 02

  • Unseized standing job

    Soak and swell work the tail loose

    A rode that lives underwater gets seized or taped · step 04

If you saw this
  • Hitch works loose after a soak

    Check

    • ?Working end through both turns
    • ?Half hitch dressed

    Then

    • Seize the tail
    • Retie wet
  • Single pass, no round turn

    Check

    • ?Hitch slips on the ring in a swell
    • ?Inspect step 01

    Then

    • Retie. Do not dress a defective structure and call it seated.
  • Working end missed a turn

    Check

    • ?Tuck does not lock and the rode walks
    • ?Inspect step 02

    Then

    • Retie. Do not dress a defective structure and call it seated.
  • Unseized standing job

    Check

    • ?Soak and swell work the tail loose
    • ?Inspect step 04

    Then

    • Retie. Do not dress a defective structure and call it seated.
Same family

Same failure class, different geometry. Open the plates — Diagnose still starts from the symptom.

This failed — start Diagnose

What holds it

  • ·A hitch is object-dependent. Remove the spar, cleat, ring or pile and the structure is usually the unknot.
  • ·Bayman / Maddocks–Keller describe a no-slip regime from wrap count and friction. The prediction is approximate, and only for hitches.

What topology does not predict

  • ·Closed-loop knot invariants (Jones, Alexander) do not say whether a clove hitch will walk on a smooth spar.
  • ·No manufacturer working load is implied by wrap count.
  • ·Wet, iced, or HMPE covers change friction — topology is silent.

Not for

  • ·Joining two free rope ends — that is a bend
  • ·A standing loop that must survive after the object is gone
  • ·A job you must undo every five minutes
  • ·Slick HMPE without a seizing

Same pattern, different job

No sourced twin among these knots. Related knots stay on the Library card.

Sources