Siobhan has sent us her work on this problem. Well done, Siobhan!

The problem says that t8 2 has two cycles, whereas t8 3 has one cycle and t8 4 has four cycles. I noticed that in each of these examples the number of cycles in t8 k is the highest common factor of 8 and k. I decided to test this on shuffles of lengths 9 and 10.
Powers of the cycle (1 2 3 4 5 6 7 8 9)
Powers of the cycle (1 2 3 4 5 6 7 8 9 10)
Here are the tables I got:
t9 0 = (1) t9 1 = (123456789) t9 2 = (135792468) t9 3 = (147)(258)(369) t9 4 = (159483726) t9 5 = (162738495) t9 6 = (174)(285)(396) t9 7 = (186429753) t9 8 = (198765432)


t10 0 = (1) t10 1 = (12345678910) t10 2 = (13579)(246810) t10 3 = (14710369258) t10 4 = (15937)(261048) t10 5 = (16)(27)(38)(49)(510) t10 6 = (17395)(284106) t10 7 = (18529631074) t10 8 = (19753)(210864) t10 9 = (11098765432)

This agrees with my prediction: the number of cycles in tn k is the highest common factor of k and n. (Of course, (1) really means (1)(2)(3)(n) so has n cycles.)

In tn 1 , consecutive numbers are 1 apart ( 1, 2, 3, ..., n). When we do this twice, they are two apart, because, for example, 1 goes to 2 which goes to 3 so 1 goes to 3. When we work out tn 3 , the numbers that were 2 apart become 3 apart, because, for example, 1 goes to 3 which goes to 4 so 1 goes to 4. It's quite easy to see that this will keep happening, because each time we apply tn we move the numbers 1 further apart. So in tn k the numbers are k apart. But now we can see that the number of cycles we get will be the highest common factor of k and n; this is like the Stars problem mentioned in this question.