General research advice: I would also go in with the mindset "I am trying to disprove this approach as quickly as possible". This allows you to iterate quickly

General research advice: I would also go in with the mindset "I am trying to disprove this approach as quickly as possible". This allows you to iterate quickly across hypotheses and not get dogmatically fixated on one approach. Cast the net wide in the initial search, including both normal and crazy solutions, then measure what approach is working, and then narrow the search based on that, learning lessons along the way. Don't start with a narrow search. Also, be sure that the different approaches you're trying are uncorrelated and ideally thematically/qualitatively different. Many researchers get tripped up trying "different" things that are really just superficial iterations of the same core concept.