some good suggestions here. I hadn't thought about fuses for the Flukes. I will have to shop around online. Do multimeters lose calibration often?
Flukes don't typically lose their calibrations. There are many examples of 20+ year old Fluke meters that are still holding their calibration (87-V for example). That said, there are exceptions, particularly if they've been highly abused (electrically speaking especially, as components that have been regularly pushed past their design limit would be more likely to drift).
As per other brands, it depends, and there may not be a lot of long term data to determine how well they'll hold either.
Unless you have access to calibration equipment, the most common trick to see if your meter is anywhere near what it should be, is to have multiple meters to check against. This has other advantages too, such as being able to make multiple measurements simultaneously, or having a spare meter if your primary is damaged.
There are also some reference voltage and resistance products out there to check against as well (not technically a calibration standard, but close). Not horribly expensive either (particularly compared to the cost of a traceable NIST calibration), so something you might want to consider (even if you do have multiple meters).
Couple of reference examples:
VoltageStandard.com
Gellar Labs
If you want to research this further, you might want to look at the
EEVBlog.com forum. Great videos on meter reviews as well (main page, EEVBlog.com, look at the older videos for Multimeter reviews). Worth a look IMHO.
Do you guys trust ebay "new"? Vs retail new
Generally speaking, it depends. I research it carefully, and ask questions of the seller if I'm not sure. Depending on the answer (or lack of one), I decide from there.
In the case of a Fluke, be warned that new meters sold on eBay are considered "grey market" by Fluke, so they don't have to honor the warranty. Sometimes they do, but officially, it's only valid if it's purchased through an Authorized Retailer/Distributor.