Thanks for the answers. There's no more to see than what's in that screen, really. Anyway, I think I find what's causing the bug: lag. In order to test if the changes are supported, I replaced a bunch of lsl/lsr #6 with mulu/divu #sstsize (sstsize now being $4c). While I knew it would cause major slowdowns, I didn't know those slowdowns would also break the object. It's apparently a problem with an interruption causing a loss of data when the children addresses are being calculated. I was planning to change the children system later, but it looks like I have to do it sooner than expected.