You are simulating pattern matching. Read the #Performance note.

Every operation can be a right fold ( one fold per input list, interwoven if more than one ).

Don't look at the type error - it's not helping in this case. You are misapplying types somewhere. Have a good look at what is a list and what is an element. ( It helps me to name lists xs instead of l. )

(cons (head x) ... x is not a list

Random tests are incorrect in Haskell, there are already issues raised about that. I assume after seeing your profile that this is the language you are using. Please say it next time.

ive looked at the Python and JS tests and they never generate duplicate elements. it would require a redefinition of the problem anyway, since [4, 5, 5] can be seen as both A or RD. i've added this to the description

[7, 7, 1, 2, 3, 4, 5] is not RA becase 7 < 7 is false; and there's always an answer by definition.