KUBO is a short 2018 NES homebrew where a cowboy turtle is on a quest to find the mysterious crystal. To do this, Kubo must retrieve a key and fight many monsters using his laser gun in order to reach his goal. This TAS speeds through his quest in just 40 seconds by manipulating RNG and taking the tightest possible lines.

The very first task is getting a key. Fortnuately, the world is extremely small and it takes barely any time at all to reach the key room but it is guarded by a monster. It's possible to open the chest without defeating it but the key itself is unobtainable. It seems like the boss' first few moves are determined by the Y value upon entry of the screen which is why Kubo moves down in the previous room. This boss takes 3 hits to kill and will be knocked back in a seemingly random direction but I was able to get it in the doorway so each hit would send it into the wall and it couldn't escape.

With this key, the final boss fight is unlocked. It consists of fighting 3 monsters that take 2 hits each. I had quite a bit of trouble getting optimal RNG for this fight; the top two monsters often walked away from the centre or they all did their own thing. Eventually, I was somehow able to get the lefmost monster to damage boost towards Kubo which would then, in turn, send Kubo flying towards the other monster at high speeds.

KUBO only allows the player to travel in one of the four directions rather than 8. However, by holding L for two or more frames, R for one then either up or down on the next, Kubo can walk diagonally to the left. It doesn't seem possible to do this to the right which is a shame given 90% of the time is spent travelling to the right. This also allows for him to moonwalk though it isn't faster than regular walking and is completely useless but a nice side effect nonetheless.

feos: Claiming for judging.

feos: For such a short and simple game, I don't believe there's a good reason to be losing 12 frames on pure movement, and 13 more frames on a missing damage boost (which you have one more hit point on). If those cost 1-2 frames each, I'd just post a suggested improvement. But for 40 seconds of basic movement around the corners, one should really look into techniques that make this game ever so slightly more interesting to TAS. This is a decent start, but I still have to reject it for sub-optimality.