Maybe everyone is different. Here is how I get to do this for the interval of perfect 5th. Basically, I realized most of the perfect 5th is going to the same colored key, i.e. white to white, or black to black, with only two exceptions Bb->F, and B->F#. Knowing that, finding the key of perfect 5th above or below a given key I can do just by almost by instinct (well, there still is the eyeballing of distance, but that does not seem to require much thinking, either).
The goal is to be able to do the same for perfect 6th. If given time, I can figure out where they are without problem, by counting semitones, or by finding the 5th and add/subtract a whole step to it. But that's too slow!
I looked at this problem again tonight, and have found some rules that might help. Still, it's about first determining whether the 6th lands on the key of same color of different color, then finding the distance. If I look at the 12 keys in an octave as two parts, each is symmetrical with white keys on outside (C-C#-D-D#-E, F-F#-G-G#-A-A#-B), then:
- for white keys, going up a 6th
-- the first two of each group will land on white key (e.g. C-A, D-B, F-D, G-E)
-- the third and fourth of each group will land on black key (E-C#, A-F#, B-G#)
- for black keys, going up a 6th
-- the first one of each group will land on black key (C#-A#, F#-D#)
-- the other black keys in each group will land on a white key
Then, going down a 6th is a mirror image of this. The rule remains the same, but when I think "first" of each group, it's "first from the right side".
As I write this, I realized this is harder to write but actually not so complex when looking and poking at the keyboard. But if you try, you might get an idea what I trying to internalize.
And I am not sure if this is the most effective way to achieve instinctive identification of any interval's destination, hence my original question.