I remember a method to make a reasonable estimate of the magnitude of a given star X by using two other stars of known magnitude as references. The method used evaluation phrases like "the star X and the reference star appear to have the same brightness, and this sensation remains even after prolonged observation" or "X and the other star appear to be the same, but after some time, X appears slightly brighter" or "X and A are completely different in brightness". Each of these phrases had a value (IIRC from 1 to 5) and then with a formula you could estimate the magnitude of X from the magnitude of the reference stars and the values obtained from the comparison phrases.
Do you know the name of the method, and the exact protocol ?