In my book, I have used the term "experimental observation" to describe any particular mathematical pattern discovered and supported by extensive computational evidence. For example, my first experimental observation concerns the possibility of transforming a composite number into a prime by changing a single decimal digit of the given number.

No, although two different experimental observations were applied to all the RSA challenge numbers we do not obtain a prime in every case. There are eight RSA challenge numbers, but only seven of them can be transformed into primes. The exception is RSA-896 where 896 stands for the number of binary bits. In other words, RSA-896 represents a 270 decimal digit number which has not being factored to date.

My third experimental observation states that if you obtain a prime by subtracting some power of ten from a given Fermat number, then the decimal size of one of its prime factors must be either equal to that power of ten or one plus that power of ten. Furthermore, please note the possibility that a certain Fermat number can be transformed into a prime but no corresponding prime factors exist for such number. Nevertheless, this observation provides (when applicable) an indirect way to uncover the decimal size of one of the prime factors of any Fermat number without actually attempting to factor the number itself.