Fixed.

The outputs for "expected" and "but was" are apparently swapped in the Java version, which is a bit confusing.

Very elegant use of elementary geometry!

No need to feel that way. Had to glance at the description for a sec to remind myself. You're good :)

Indeed. Thanks. Definitely one of my less bright moments...

One subarray represents the command to execute if the bit at the current position of the head is a 0 and the other subarray represents command to execute if the bit at the current position of the head is a 1.

I don't understand your list of 1-state TMs. They include TMs that have two different commands for state A and reading symbol A, such as this one:

const states = { "A": [
[0, L, "A"],
[0, R, "A"],
]};

So, are we talking about indeterministic Turing machines here? If so, does this mean that all subsets of possible pairs of movement direction and new state are admissible as the command for one state-symbol-pair?

I find it strange that "null" is to be assigned to the number of items appearing 0 times in the list. Everything that is not in the list does that, and that is a lot. ;-)

Getting to know the libraries? I am always happy to see such solutions, I learn a lot from them.

You need to take into account that the circle has 4 sectors, and then the origin of the coordinate system.

It carries out one last step of the recursion.

Learned something new from this one, (namely "modulo 1"), thanks!