Ah. It was mostly to get the antithetical 'bit more [loss]' → 'lot more [gain]' pattern in.
But, since you ask, the Q of a tuned system (inversely proportional to its associated power loss) is the issue. The higher the Q, the more precisely tuned it is - which is not what you want if you'd like the frequency to vary rather more than a semitone. If this device were more sensitive to the differences in whatever it was measuring, you might be able to get a decent tune out of it. I'm quite sure it could be tweaked enough to get at least an octave out of it.