Submission #1401: Randil's NES Mickey Mousecapade in 07:16.20

This run is 962 frames (16.03 seconds) faster than Michael Fried's run. This is due to optimized movement, better luck manipulation, and not killing the witch boss in the first level.

Mickey Mousecapade is a platformer that looks easy, but is actually both hard to play, and to TAS. You play as Mickey, and Minnie is your companion, always following you, throughout the game. You can shoot stars with your wand too, and later on Minnie can shoot stars too. In short, this game is pretty trivial, and doesn't need any extensive description. You'll understand the game when watching the run.

Comments on movement: Optimizing movement in the stages with ladders (first level and the boat level) is a whole science in itself. The main thing to consider is that Minnie has to stay close to you all the time, since you can only proceed to the next room when Both you and Minnie reach the exit.

Getting on and off ladders as fast as possible is quite a challenge in itself too. I've tried hard to optimize these parts.

Luck manipulation: To sum up this game's randomness: This is the most pseudorandom game I've ever TASed. Really. Enemies placement differ a lot depending on when you enter a room. Other things such as how many points you have can also effect randomness. Almost every factor in this game can affect the outcome of something. Because of this high dose of randomness I sometimes need to wait for a few frames before entering a room, in order to get the enemies in that room into better positions. There are two places in this run where I have to wait about 17 frames before entering the room, due to misfortune in luck manipulation. I tried hard to avoid these waiting times, but with no luck. So these waiting times you see before I enter a new room sometimes are purely for luck manipulaton, and not sloppiness.

Lag: This game lags when there are two many moving objects on the screen. If I enter a room that lags, I manipulate luck to get the enemies in better positions so I can finish them off quickly, and get rid of the lag.

Some small comments:

As always, I did most of this run in frame advance. Memory Viewer was used at times, to check my position and other relevant things. The rerecord count is quite high as you'll see. This is both because I tried hard to optimize this run, but also due to this game's randomness and the difficulty to optimize movements.

I can't think of much else to mention in these comments now. If I think of anything else, I'll just add it in later on. Enjoy the run, and once again, a merry christmas to all of you!

PS: This game has one of the worst and most boring endings of all NES games, I think.

Bisqwit: Accepting. Congratulations, timewise this improvement is greater than I suspected there's room by visually inspecting Michael Fried's movie... It would have on my own todo list if it weren't improved for another 4 years ;)
Bisqwit: Reverting acception because of found improvements.
Randil: I'm cancelling this run because of known improvements. I will start working on a faster run.