Seven years and two publications later...
On those days when I can't sleep, I get funny ideas, like the thought that there might actually be a real [submission 3210]ten frame improvement[/submission] to the [movie 529]current five-year-old run[/movie].
Several intensive hours later, I give you
a three frame improvement to Stage 1.
This really should be a seven frame improvement - I have a five frame lead at the time I exit the second room, and pick up another two by the time of the third hit on the boss, but the boss reappears two frames later than one would expect from the timing of the third hit, and the end-of-level potion appears another two frames late (relative to the last hit). I haven't figured out why this is the case as yet (frame rules, possibly?).
Some RAM address for the curious, or for those interested in contributing (this was absurd enough to try to improve that I may not be keen on continuing):
- $0340 - X, high pixel byte.
- $0330 - X, low pixel byte.
- $0320 - X, subpixel byte.
- $0370 - Y, high pixel byte. The high byte seems to cycle in the wrong direction - it increases when the low byte underflows, and vice versa.
- $0360 - Y, low pixel byte.
- $0350 - Y, subpixel byte.
- $0390 - X speed, pixel byte. This doesn't seem to be signed - it really is a speed as opposed to a velocity (i.e. the direction isn't indicated).
- $0380 - X speed, subpixel byte.
- $03B0 - Y speed, pixel byte.
- $03A0 - Y speed, subpixel byte.
So... who's up for more work on this?