I was reading Greg Maxwell's paper on Confidential Transactions and I think utilizing the additive homomorphic properties of elliptic curves for verifying transaction amounts is fascinating. One question I had was would it be possible to utilize scalars seeded with CS-PRNG output to "blind" values in a secure manner? I understand the necessity of signing the pedersen commitments within confidential transactions. I'm mainly referring to just securely encrypting and decrypting the data. This probably sounds confusing so let me try to give out an example. For instance if one generated the pedersen commitment C = X A Where X is a 256 bit randomly generated scalar and A is a normal integer value one wishes to blind. This can be trivially decrypted by preforming: A = C - X. If one were to preform cryptanalysis on the commitment C would it look like the random value X was mutated at all or would it still look like random data? Because I believe that even if the attacker knew that C had been mutated in some way, they would have no idea what A could be since X was a sufficiently random source. I'm not trying to implement my own encryption algorithm or anything, I'm just curious on how secure this method would be if one were to analyze the ciphertext/pedersen commitment. Thanks for all the help!