Playing the "Occam's Razor" card - and getting it wrong

Occam's Razor is a principle for selecting between alternative hypotheses. The statement is:

If alternative hypotheses equally explain the facts then the simpler hypothesis is preferred.

This principle is used to claim that a simple explanation is correct when, in fact the explanation is not simple but is simplistic - naive and not actually covering the facts. Such uses of Occam's Razor pick a hypothesis and then cherry pick which facts to observe and which to dismiss. Famous examples, challenged by "fringe" archaeology, are common:

The "Clovis First" hypothesis: all humans arrived in America in a single "wave" of migration, after enough of the glacier had receded to make that possible. 

  - simple [CHECK]

 - all the facts [not unless you ignore thousands of examples of pre glacial sites, ignore the absence of Clovis material in Siberia, ignore the obvious diversity of arrowhead design styles] [NO CHECK]

The "No Europeans before Columbus" hypothesis.

 - simple [CHECK]

 - all the facts [not unless you ignore hundreds of European inscriptions, the tuberculosis resilience of New England's coastal tribes, as well as the general human tendency to go long distances in boats.] [NO CHECK]

The "Indians did not build in stone" hypothesis:

 - simple [CHECK]

 - all the facts [nope, there is not a single fact supporting this statement. Actually, the statement is not a testable hypothesis, and does not rise to the level of scientific statement] [NO CHECK]

So in practice Occam's Razor is used, not to choose between alternative hypotheses, but to exclude facts from consideration. If you hear Occam's Razor used in a conventional archeological discussion you can be almost certain it is being used to gloss over key details that do not fit the conventional thinking.