Do you suppose there might be a finite number of highest numbers?
No, but I have found after running several million tests, that there are major clusters similar to 9232 found in higher regions. One thing that I think is cool is that with the exception of numbers that are powers of 2, (powers of 2 are always the shortest route to 1 for obvious reasons) there is no easy way to guess how many steps a given number will take to get to 1 without running it.
From the list I posted previously...
25 takes 23 steps, and 88 is the highest number reached.
26 takes 10 steps, and 40 is the highest number reached.
27 takes 111 steps, and 9232 is the highest number reached.
28 takes 18 steps, and 52 is the highest number reached.
29 takes 18 steps, and 88 is the highest number reached.
The number 27 is definitely an outlier in the number of required steps.
There are other variations on the Collatz Conjecture. One multiplies by 5 and divides by 2, and has the same result. Strangely that 9232 number pops up there and in other variations as well. I did some searching on the 9232 number and couldn't find anything specific about it.
I wish I'd known about stuff like this when I was young and stupid.