[TAS] Decimation

Here it is. 3 months of work, from august to october. 2300 replays with an estimated 3000 bruteforcing sessions.

I think the hard limit (without uber nor noseboost) is somewhere between 30 to 35 minutes.

The "New cuts: X" displays the amount of new cuts found until the checkpoint where it's at, included. i.e if the TAS takes a checkpoint where it'll do a cut that wasn't done on the previous TAS nor is known by the community (to my limited knowledge of their decimation cut knowledge at least), the counter is increased by 1. I tried my best to categorize what is a cut and what is not, but there are a lot of situations where it is quite challenging to tell whether it's just a trick or a real shortcut, or even if skipping some intermediate blocks because of extra speed should be considered a cut. I think counting 45-55 new shortcuts is a good approximation.

For the few times my cp sector was slower than the previous TAS, it is not due to a lack of optimization on my behalf, but rather to a cut or trick used on the previous checkpoint to gain a lot of time but at the cost of some speed when passing the checkpoint. I made sure, though, that the time gain always outperformed the time lost afterwards because of lower speed.

And if you're wondering, I've smoothened the input display. It doesn't entirely reduce the input flickering due to bruteforce, but it achieves its goal of being more understandable. Note that this is only the displayed inputs, and playing those inputs wouldn't match the run. I can provide the replay or inputs upon request.

I would like to add that this run would probably have never been so fast without Shweetz and Kimura's TAS being first made, as it really helped me push the optimization up a notch to keep up and even gain time against it.

A huge thanks to BigBang1112 for adding a feature that I requested, whereby you can see time gained for each checkpoint individually (sector time gain).

Clip Checkpoint:
https://github.com/bigbang1112-cz/cli...