If you’re blocked by a problem, you may be able to reframe the problem in a way where you can make progress, or at least establish some bounds which can help you to better understand your approach.
A deterministic problem may be possible to solve, but explodes in complexity through large numbers or exponential paths (Arimaa) and it’s not realistic to work through the Cartesian product of possibilities. Sampling predicted results could make the calculation manageable again…Sieve of Eratosthenes vs the Miller–Rabin primality test.
deterministic versus probabilistic (is this equivalent to frequentist vs bayesian?).
A change of perspective, view things from another angle, think outside the box, etc.
Three types of constraint relaxation.
working backwards (there’s a term for this)