I think the original question is how can ENOB be higher than resolution without averaging, oversampling or other resolution improvement methods?
The answer is it can't be I think, and maybe you simulated the ADC without thermal, flicker, other uncorrelated noise. In simulation without uncorrelated noise the ENOB can be higher than resolution, but I wouldn't believe it.
It is not clear how did you run FFT, in simulation or on an oscilloscope under measurement.