On the first set of cogs in TTC, a question was posed if climbing these cogs, rotating hexagonal right prisms, was possible without pressing A or using a bloated bob-omb to skip the walkway above the last cog. This meant entering a misalignment, which occur on corners facing away from the center of the course, on the last cog by using a dive recover, landing 4 or less frames after the dive recover to get sufficiently strong VSC, and activating the VSC next to the walkway to grab it. Pannenkoek investigated and initially concluded that it was unlikely to be possible, and later, a strat was found that made this idea unneeded.
A few years later, this idea became useful again when bringing a bob-omb past the amp's pole could save an A press in The Pit and the Pendulums by clipping into the cage.
Pannenkoek tested this idea under ideal conditions where both cogs were stopped. However, having the cogs move gave more possibilities. Also, after investigating on STROOP's Model tab, it was found that every vertex on each cog had a different distance from the center because the center was really offset by 0.707 units. This gave more options where the vertex and misalignment can be. Although it still needed to be facing away from the center of the course, choosing the closest corner from the center of the hexagonal platform made the dive recover into the misalignment less awkward.
Simplifying the problem by making the lower cog just a circle with a radius equal to the furthest corner from the center, a simulation was made to gauge if Mario can enter the misalignment within 4 frames. Because the floor of the lower cog ends so close to the misalignment, it was in fact better to attempt to reach the last cog in 3 frames instead.
What especially makes this misalignment awkward to enter is that, in all available dive recover launch positions we have, one of the walls is obnoxiously in the way of the direct path into the misalignment. Pannenkoek's attempts had to launch towards the wall, slow down, and strain to turn at this wall if he wanted to reach the misalignment.
However, by having the cog be turning, 1 frame before the misalignment is reached, we could have the walls be at a different position that were blocking less of Mario's path to the misalignment as it would be otherwise.
| To Be Continued >>>
When the misalignment was finally reached under hacked conditions, Mario immediately had his VSC activated as if he immediately left the floor. It was quickly found that this was because of NUT spots, edges of moving platforms that swap from being floors to not. Normally useful, these spots occur because edges of platforms are not smooth, but are quantized squares, and when the platform turned, the orientation of these squares were not accounted for. Landing further inside the misalignment solves this problem, but this meant that the problem was constrained even further.
[skip skip a few paragraphs]
Executing the strat would mean setting up the ideal conditions given in the simulation. Pannenkoek used a brute forcer and
"TTC, dive recover, cog, walls extend in the x or z direction,"
"we were tasked with launching Mario into a corner of a rotating hexagonal right prism platform at an angle, which seemed impossible due to the angle constraint and the lack of ground. I found in the 3D model that all 6 vertices of the hexagon were different lengths away from the rotation axis".
"Due to wall behavior, the corner only opens up at certain points in the platform's rotation. After setting rotation to max speed, I planned for Mario to launch into the closed corner and get pushed out by a wall, which allows a better angle to strain into the now open corner. I modeled this in Desmos and we were able to successfully replicate this within the game."