(The following sounds too confusing, but making it clearer doesn't seem straightforward. Posting just in case communication succeeds on some level.)

Payoffs are part of preference, preference is (aspect of) world program, and world program determines the outcome. By changing the payoffs, you change the world program, and where decision-making is based on logical inferences about it, you can thus fundamentally change the shape of possible inferences. "Nice" solutions mean parametrization of the solution by elements of the problem statement, but here parametrization is bound to talk about what can be logically inferred, so expect no nice solutions.