I still don't understand what you mean by fake. Do you mean poor performance or disguised carbon film resistors?
As a relative test I compared two 11k resistors:
1) Chinese metal film
2) Digikey
http://www.digikey.com/product-detail/en/MFR-25FBF-11K0/11.0KXBK-ND/13227Their initial values after stabilization in a 2 ohm measurement on a 6.5 digit meter were within 2 ohms of each other.
Anyway, I nulled out the meter and performed a finger pinch heating test, being careful not to touch the leads. I held them for a while (it took lets say 20+ seconds to stabilize).
The digikey one, listed at +- 100ppm, stabilized at ~ +1.6 ohms.
The ebay one, listed as 100ppm (i think?), stabilized at ~ -3 ohms.
I don't know how hot my hand makes em but they are similar size and I was holding it similarly. When I released my fingers the resistors went back to null (as good as they will get in a slightly drafty room) and the results were fairly repeatable (lets say to 0.2 ohms).
So it seems that worst case, assuming the digikey resistor is in spec, the chinese one has -200ppm temperature coefficent.
I'd imagine that the digikey resistor is slightly better then its spec, (as dave jones demonstrated that the tolerance of resistors listed as 1% was more like 0.5 percent) then I don't think the chinese ones are too bad.
Perhaps I will try pressing them against a light bulb in a styrofoam box or something. Getting a stable thermal environment seems like a complete motherfucker.