Placing gear in soft rock and other unhelpful geology

Every piece of climbing protection carries a strength rating. That number describes the metal. It says nothing whatsoever about the rock, and the rock is what actually fails in most gear-related accidents.

Learning to read rock is slower than learning to place cams, and it matters more.

Soft sandstone

Sandstone varies enormously — some is as hard as granite, some is barely consolidated sand. The soft end is a genuine problem for camming devices, because a cam works by pushing outward, and outward force on soft rock can simply crush the crack walls until the cam slides.

In softer rock, prefer passive protection with large contact area. Nuts and hexes distribute load over a wider surface and are less prone to blowing out a crack wall. Where cams are the only option, use larger sizes than you would elsewhere — a bigger cam spreads its load — and set them in the most solid, parallel-sided section you can find.

The other rule about sandstone: do not climb it wet. Water dramatically reduces its strength, and holds and placements that would be fine when dry fail when saturated. Many sandstone areas have a local convention of waiting a day or more after rain, and the convention exists because people have broken holds off and decked.

Limestone

Limestone is generally strong but structurally unpredictable. It forms pockets, undercuts and hollow flakes, and the surface is often the weakest part of it — a thin crust over softer material beneath.

Tap it. A hollow, drummy sound means a detached flake, and a detached flake with a cam behind it is a lever waiting to pull the whole thing off. Look for the crack behind the feature: if you can see daylight, or if the flake moves at all, treat it as decoration.

Threads through natural holes are often the best protection available in limestone, and they are usually stronger than anything you can place.

Granite

Generally excellent, which produces its own hazard: complacency. Granite cracks are frequently flared, and a flared crack is where a cam can walk into a wider section and become useless. Extend placements with a sling so rope movement does not walk the piece.

Granite also produces exfoliating flakes and blocks that have separated from the wall but are held by friction. Big detached blocks look monolithic and are not. Never place gear behind a block you would not be happy to have fall on your belayer.

Choss and volcanic rock

Some rock simply cannot be protected. Loose volcanic rock, heavily weathered schist, anything that comes away in your hands — the honest response is often to not climb it, or to accept that the gear is psychological.

If you are on it, spread the load. Multiple mediocre pieces, equalised where you can, close together. Move gently. Do not fall.

General principles that hold everywhere

Look at the crack, not the piece. A perfect cam in a bad crack is bad protection. Parallel walls are what you want. Flaring means the piece can be pushed out; converging downward means it may pull through.

Consider the direction of load. Gear is placed for the pull it will get. A nut set for downward pull can lift out when the rope runs upward. If a piece could be lifted out by rope movement, oppose it with a second piece or extend it so the rope does not tug it.

Extend generously. Rope drag is not just tiring; it actively pulls pieces out of position. A quickdraw or a sling on a piece that sits off the line of the rope is not weight wasted.

Test what you place. A firm downward tug seats a nut and reveals a cam that is going to shift. Do it before you climb above it, not after.

Check the rock above your protection too. A perfect placement under a loose block means the rope will saw across that block during a fall.

The uncomfortable conclusion

Gear placement in variable rock is a judgement built from thousands of placements, most of them made on the ground or on top rope where being wrong costs nothing. That is the way to build it: place gear constantly, in every crack you pass, and have someone more experienced tell you which ones are bad.

The alternative is building the judgement on lead, one fall at a time, and the sample size required is larger than most people survive.