Death to the Resistors!

Chris, K4FH, has been gracious enough to give me 5 of his 100 ohm ebay resistors to test for their actual watt rating.
I would like to solicit comments on the criteria people would consider 'good enough' during this test.
They are marketed as 1/4 watt, 5% tolerance resistors, from ebay.  I will try to identify the manufacturer and I will try to acquire a datasheet for them.
Here's the procedure I will use to test them:
1.  measure resistance. The measurement shall consist of two readings from my elcheapo centech digital VOM; 1 taken as close to the epoxy body of the resistor as possible, and the other as close to the ends of the resistor leads as possible.  Any significant difference in measurement will be confirmed using an analog meter.  "Interesting" results will be posted to QRP-L and my blog before I do any experiments on the resistor, and that resistor will be isolated.  I will take these measurements at the following times:
   a) before doing any testing
   b) immediately before testing
   c) immediately after testing
   d) next day after testing
2.  Label resistors, Resistor 1-5, based on their resistance from measurement A, 1 being lowest, 5 being highest.  In the event of a tie, the first resistor measured gets the lower number.
3.  Test the power dissipation of the resistors.  I will probably only do one test a night over the next couple of weeks.  I usually get home from work sometime between 04:30 and 05:00 UTC, and spend about 1.5 to 2 hours 'winding down'.  I try to be in bed by 06:00/07:00 UTC (2am local time whether daylight savings or not).  That gives me about 1 hour to test resistors each night.  During the day, I watch my kids, so I won't be able to test them outside of that time frame.  Here are the tests:
    a) full bench supply for 1 hour: I'll post the exact model number and voltages read before the test, but a 100 ohm resistor across the 13.8 VDC terminals will output  1.9044 watts and draw 138 mA of current, if I done did the math right. I haven't measured the terminal voltage of the supply in a while, and this will be as good an excuse as any to figure out howthe list high quality my 10 year old supply is.  Voltage will be measured every 5 minutes during the test.  After the test, I will publish my results via QRP-L, and my blog before continuing.
    b.) I will set up a series/parallel resistance circuit.  The series part will consist of a lump parallel section, and a variable series section.  The parallel part will be 5 resistors wide, the test resistor being one of them, and is switched in and out of circuit with another resistor of equal resistance and higher power rating.  The circuit will be constructed in such a way that the total power handling of the circuit is at least 4 Watts without the inclusion of our test resistor.  Each of the tested resistor will be tested by varying the voltage available to the parallel section until the resistor fails.  When the resistor fails, it will be switched out and the 4th resistor in the circuit switched in, the voltage across the fourth resistor will be measured and used to calculate the power dissipated.

Circuit theory: because resistances in parallel share voltage and resistances in series divide voltage, it's possible to use a series/parallel circuit to produce a measurable voltage at the terminals where a test resistor is located.  This measurement can be used to record the power dissipated by the resistor at the moment of failure. 

My circuit and calculations will be published on my blog and QRP-L before testing for peer review. 
    c.) After each test I will publish my results within 12 hours of testing. 


So there you have it, Death to the Resistors!
thanks everybody!