here's a fun one. this is mighty bomb jack.
a lot of this game's code exhibits a strange quirk; it still works as you'd expect, but something about it is rather unusual ... can you spot it?
transcript
ROM:8420 GameState_5: ROM:8420 20 89 87 JSR sub_8789 ROM:8423 AD 40 00 LDA byte_40 ROM:8426 29 20 AND #$20 ROM:8428 D0 1F BNE locret_8449 ROM:842A 20 45 88 JSR sub_8845 ROM:842D AD 17 03 LDA byte_317 ROM:8430 F0 0B BEQ loc_843D ROM:8432 20 74 84 JSR GameState_7 ROM:8435 AD A8 00 LDA byte_A8 ROM:8438 09 40 ORA #$40 ROM:843A 8D A8 00 STA byte_A8 ROM:843D ROM:843D loc_843D: ROM:843D 20 A9 87 JSR sub_87A9 ROM:8440 20 0E A4 JSR sub_A40E ROM:8443 20 B8 97 JSR sub_97B8 ROM:8446 20 6A 8B JSR sub_8B6A ROM:8449 ROM:8449 locret_8449: ROM:8449 60 RTS
hint
the opcode bytes are important here. if you only use the disassembly, you won't see it.answer
the code does not use the more efficient opcodes for interacting with the zero page.for example, "AD 40 00" should be "A5 40".
it saves a byte and a cpu cycle each time. it does add up!
bonus fun facts once you know the answer
atlantis no nazo is another game that doesn't use zero-page opcodes (for the most part). super mario bros 2 uses zero-page opcodes most of the time, but absolute ones show up in a few spots.the assembler i use doesn't (didn't?) support forcing direct addresses for zero-page, so the disassembly has compatibility macros to fix it.