The CAT tests (see overview)
come in about a zillion levels, and the level that will have been taken by a ten year old (C, I think) is designed to provide meaningful and informative results over most of the ability range. An 11+ test doesn't need to provide useful information substantially below its intended range: if it yields zero for children at (to use a sloppy metric) 2c and 3a, so what? It's only really interested in providing information about children who will mostly be getting fives. The CAT test has to resolve over all the levels. So a CAT test will contain some easy questions (which in fact aren't all equivalently easy) so that there is something to distinguish children at the left end of the distribution from each other; the 11+ doesn't need to, because it wants to use as much of its range as possible to distinguish the children around the admission threshold.