This is my first entry into an OM event, but I've been following along since the 2023 weeklies. This iteration of the solve improves upon the previous one, which had an extra track segment below arm 2 (the input arm). My original idea was to add as many track segments as possible to delay how often arm 2 actually grabs an input, but 5 track segments ends up being better than 6 because 6 is also the number of track loops it takes for the arm's angle to repeat, whereas 5 and 6 are coprime.