I have failed to cleanse myself from the cardinal sin of a sum solve: The holdover piston. Other than that, this solve feels pretty good. Built around arms 5,7,8,9.

Anime conditionals are real! My first solve used a stick instead of a flare for metal differentiation and clocked in at 1300 sum, which I then optimised down to 1200. Then I realised flares were just better, which immediately cut it down to 1050 where I stand.

This method theoretically goes down to P15, limited by quicksilver. However, trying to get a P18 working was going to cost more than would break even. Thus I stopped at P24, limited by arm 9. Somewhere in between is probably optimal.

I predict that this solve will place 3rd and that the winner will be around 1000 sum flat.