Hi Christof,
That all makes sense to me. Thanks for your research. It's on a higher level than mine.
I think it's important to point out that this logic applies to any potentially skewed fields where the values are poorly distributed. Additionally, I have a lot of SELECTs with < m.SomeDate predicates with an index on SomeDate. The < m.SomeDate in some cases does a poor job of eliminating records. I believe I would do better to use the HAVING clause in those situations.
Charlie
Charlie