Since the matter came up last night with some spirited discussion, thought I'd throw it out here. Marsh and Ryan were both sparring about the relevance of measuring the number of hands played in cash games versus tourneys. I hold that it is best to record time spent at either a cash game or tourney and record number of hands for each as well. Then after logging into a spreadsheet you can cut the data any way you want. I am even having second thoughts about recording sessions at multiple tables as one line item since it slightly muddies the waters of hourly table yield. I do record how many tables I was playing at and can get an average take and that is close enough for government work.
So how best to determine where you should be playing? I have to think it is wherever you have the best real world hourly rate regardless of number of hands played or tables you are sitting at. I still want to record number of hands just so I can do calculations later but my punch in and punch out times are going to be the measurement of earning potential/hour. I think the $/hand measurement alone is irrelevant for deciding between ring games and touraments since it does not take into account multitabling. $/hand IS an important metric for tuning your play though so there is no way I'm going to leave that stat out of the log sheet.
Anyway, that's my take on the matter.