Hello! The game uses the standard kernel load routines, which aren't known for their speed. I am surely not a master C64 programmer, and in fact the game has been written in XC-basic (https://xc-basic.net/), not in assembly. Surely someone expert in assembly would have been able to write a fastload routine, and would have been able to write code in a more efficient way. But alas, that someone is not I.
A side effect of using compiled code, is that some revisions just do not work when compressed. I haven't a clue of what triggers this problem, and it has been a mad rush to get the game out before Xmas (although it was a self-imposed limitation, we would have never released it if we didn't set a date for ourselves). That being said, I'll make sure that the post Xmas updates will have compressed executables. Those quicken the initial load time (aka, the black screen one) by around 60%. I strongly suggest using a kung-fu flash cartridge on original hardware if that load speed won't be enough. On VICE, CTRL+W activates warp mode that will considerably speed load times up.
About the other suggestions, I will try to implement them in the next update as well. Really, the only limit is the meager amount of free memory, but I'll try :)
Thanks for the feedback!