In Theorem 1.1 of his 1975 paper [4], Terras states that the -th iterate of , as presented in our introduction,
can be expressed as a "remainder representation" of the form
where represents , applied times, and . He then goes to define
We shall attempt here to rewrite this "remainder representation" as the same statement as equation in our introduction, appearing in the 1978 paper [2] by Böhm and Sontacchi (as Proposition 7),
for any number that does arrive at by the iteration of the Collatz function.
Recall that we defined an integer and constants , with and , that represent the steps taken by iterations of the original Collatz function as it goes from to :
The constant represents the number of times the rule was applied during the entire trajectory of from to .
We also defined constants as partial sums of the (, , , etc.).
Back to Terras, he uses a "shortcut" and applies the mapping to odd numbers. Therefore, a trajectory of from to is shorter (by steps) than a similar trajectory of the original function.
(If the total steps were for the function , there are only steps when using the function .)
So we have representing the total number of applications of the function , starting from and until reaching .
Notice that are precisely the positions in the vector where it contains a value of (the times the rule was used, plus the very last element corresponding to our arrival at ).
Armed with all these definitions, we have
and finally, since
we end up having
which is precisely equation above.