First cave completed. Stats:
Me: 17343
Randil: 17427
Sami: 17742
So, I'm beating Randil by 84 frames and Sami by 399 frames. Not bad! I might attempt to navigate to Chapter 3 before crashing.
Before I continue with that, I want to explain where this new 12 frame improvement to Randil's version came from. The most important one is in the boss fight: I jumped towards the boss before throwing an axe at it, allowing me to hit it earlier. In addition, I used the side of one of the skulls to get a subpixel boost like the one Randil uses in his jump glitch. It's a small improvement, but everything counts, right?
The above improvement takes care of 11 frames. Where did the last one come from? Well, while navigating through the maze, for some reason, Randil's version is one frame slower than mine, just like in the scene where you have to jump from platform to platform. Clearly, I didn't employ any kind of new strategy, but yet the improvement exists. It is my theory that it might be more optimal to turn during some frames than others, and that there's probably a relation between the animation and this. As it would be pretty obnoxious to have to check this every time, I'm wondering if a script could be constructed that checks whether movement is optimal or not. If anyone has any ideas, I'd love to hear them. Alright, back to work.
P.S. When I work on non-trivial public projects, I often monologue a lot. Get used to it. That said, if this triple post is out of order, then please let me know. I figure that this post is substantial enough to not be an edit.
Edit: Navigated to the next cave. I really thought this would be the most trivial part of my TASing day, but to the contrary, it raises many questions.
- Throughout this section, the difference between Randil and I's times fluctuate. Before this begins, it is 84, but it seems to go up after the points are tallied for some reason. In fact, I'd suspect that points would add time, but in this case, it seems to subtract it. If this is not the cause, it could be lag or inaccuracies in text navigation, though I figure Randil used the same script as I am for this section, so the latter wouldn't make sense. What caused this improvement, how do I maximize it, and how are points calculated anyway?
- On the other side, at some point, the difference became 54, a whopping 30 frames less than it started. After some examination, I think I can safely attribute this to the filling of the half a heart I lost performing Sonik's enemy jump glitch. Sure, the glitch saved more than 30 frames, but this is still disconcerting. Is there any way I can recover that half a heart in the first level?
- Again, somewhere before I made it to the gave, the difference became 56, improving my run by 2 additional frames. Why would this be? Unlike the other changes, this really doesn't seem to have a suspected culprit.
I'm hoping that Randil might be able to step up to the plate once more and help address these questions. Upon reviewing the remainder of his movie, I see that I am closer to completing the part he completed than I expected. With at least one use of the subpixel boost jump glitch available, I'm hoping to improve his movie by at least 2-3 seconds. Only time will tell...To see the latest version of my run, as always, please check out my Google Code Mercurial repository, linked several times above.
Furthermore, I got some kind of message when I wanted to load a state that was farther on in the game then I currently was at. That seemed odd to me. Should I assume states work very differently here? Instead of storing the actual state, does it just play back the input to that point or something? Even so, is this message really necessary?
Update: I'm also often having issues with sound and controls. The only button that seems to be working for this game is the Enter button. So, I was able to get 600 frames into my run; I successfully told the game that I didn't want to configure my joypad!