Most ice climbers know that a screw angled DOWN in good quality ice is stronger than one angled UP. How about some real data, please?

Chris Harmston and the gear testing experts at Black Diamond have some answers.

The difference in strength is dramatic - about 9(ish) kN for the screws tilted up, vs. about 22)ish) kN for the screws angled down. Most ice climbers know that a screw pointed down in good quality ice is stronger, but here are some real data to back that up.

Why is this? An Instagram comment from @willmurphy6612, explains:

”It has to do with how the load is displaced in the ice. When the threads are perpendicular or pointing UP, the load is transferred axially along the length of screw. This takes advantage of the compressive strength of ice.

When the threads are pointing DOWN, part of the screw is loaded radially which exploits the shear strength of ice, which is not very good. Combine that with the fact that the cracks initiated by the screw are propagating to the surface when angled DOWN the load strength of the system is severely compromised.”

Thank you Will, nicely said!

Note that there are many more variables and play here: the length of the screw, the type/brand of screw, is there a chance the screw might melt out, and above all the quality of the ice. Check out this link to learn more about these factors, and how the test was conducted.

Notice the strength at the purple oval / zero degrees - all above 13 kN, which is more than you would ever encounter in any climbing fall. Super good enough!

In their excellent book “The Mountain Guide Manual”, authors Mark Chavin and Rob Coppolillo write that they feel it's best to place screws at zero degrees / perpendicular.

Their reasoning: While in perfect ice tilting the screw downward can result in higher strength, in less than perfect ice, screws at 0° hold the best. Because it's often hard to judge the quality of ice for the full depth of the screw, they feel that defaulting to perpendicular is the best approach.

Along with these test data that show 0 degrees is plenty strong enough, that seems like a good choice.