The B.E.S.T. Challenge

I hereby challenge you to crack this code. The challenge will proceed with
three phases.

PHASE I
-------
I will provide four text files and their encrypted equivalents. A fifth file
will be encrypted, but no translation provided. The challange is to gather
information from the first four files to crack the fifth. All files
will be encrypted with the exact same key. This test will assure me that no
evident patterns exist that will assist anyone in cracking the code. By
knowing that all four data sets have undergone a transformation based on the
same key, certain statistical routines and pattern recognition software would
normally be able to provide some important information about the method and
possibly the key itself. I am hoping that the nature of this algorithm will
prevent that information from being obtained.

PHASE II
--------
If Phase I generates enough interest and a significant period of time passes
without any successful attempts at decyphering the fifth file, then I will
provide the key. Now you will have a data set, the key used to transform
that data set, and the resulting, encrypted data set. There will be four
solid examples of this method being applied to different data with the
exact same key. It is my opinion that even the most rigorous attacks using
the key as a basis will still withstand the test. This phase will assure me
that no statistical patterns appear within the encrypted stream that can
be traced back to the key, because if these transformations are evident then
it is but a small step to replicate the transformations and apply them
against the fifth and ultimate file.

PHASE III
---------
If I feel confident that the routine is impenetrable from the first two
phases, then I will take the next step. Once an encryption algorithm becomes
widely used, the method becomes well known as well. While it may be
impossible to decipher certain texts without knowing the algorithm, the true
test comes when the exact routine is known.
This phase may take awhile to begin, because I need to protect my algorithm
somehow before I publish it. I'm not sure at this point of time if it will
warrant a patent or not. My ultimate goal is to negotiate with a company
interested in this type of technology, because I don't have the time nor
the energy to attempt to market this on my own.
Phase III will provide the exact algorithm used. I will provide an encrypted
file. Using the four files from the first two phases, their keys, and the
output, and now knowing the routine, the challenge is to figure out the
key and decrypt the last file. I will not consider a brute force attack an
acceptable compromise of the effectiveness of this algorithm unless there
are an insignificant number of iterations that transpired prior to succesful
decryption. If some type of stastical attack, pattern-based attack,
transformation or other systematic method is successful, then I will know
that this encryption method does not hold par and can focus my time
elsewhere.

Thank you and good luck!