Most published pull figures are measured in ideal conditions that a product rarely reproduces.
Correlated Magnetics Research's FAQ names the two common test setups. Case 1 is the maximum pull between a magnet and an effectively infinitely thick piece of steel. Case 2 is a magnet sandwiched between two infinitely thick steel pieces, pulled against a third. CMR says data from these idealized setups can overstate the force achievable in product design by 2 times or more.
Real joints lose force in several ways. The steel is thin, so it saturates. There is a gap from paint, plating, plastic or dirt. The load is in shear, not straight pull. And the magnet tilts as it engages or releases. CMR notes that its own published data comes from in-house force testing with magnets held parallel and aligned, compared against conventional magnets of the same size and grade, and that unconstrained testing may give lower numbers.
Use catalog numbers only to compare options. Then build a test with your real target material and thickness, the real gap, and the real load direction, and design around the lowest result you measure, with a safety factor.
WholeMagnetics, 2026-10-04. Polymagnet facts are taken from Correlated Magnetics Research’s public pages, listed below; vendor performance figures are CMR’s own.