That is only true if:
a) You've got a simple test that -always- reproduces the bug on one kernel.
b) You're aware of atleast one kernel where the bug does NOT happen.
Most of the (suspected!) kernel-bugs I've run into in my years of running Linux (since 1.2.13)
has fulfilled neither of these 2.