Here is output:
https://pastebin.com/M17sJM8A
It works a good amount of times. Next I'll break down each of these individual steps and do an individual search just with those. The depth things took wicked long but I might run it in the background for depth=4 for more primes while I write a paper.
I broke it down into 13 moves
"EDX(e,a,d)/", "EDX(x,e,a)/", "EDX(d,e,n)/", "EDX(n,e,d)/", "EDX(x,d,n)/", "EDX(d,b,e)/", "EDX(b,d,e)/", "AB(a,n)/", "EDX(x,e,d)/", "AB(a,b)/", "EDX(d,x,e)/", "EDX(x,d,e)/", "EDX(e,d,b)/"
Then went for all the products of these primes
[3, 5, 7, 11, 13, 17, 19, 23, 29,31,37,41,43,47,53,59,61,67,71,73,79,83,89,97]
this is a total of 184 coprimes. I found 3 patterns that work 41/184 times:
41 184 C(c)/EDX(x,e,d)/AB(a,b)/EDX(d,x,e)/
41 184 C(c)/EDX(d,x,e)/EDX(d,x,e)/AB(a,b)/EDX(d,x,e)/
Also I found some other notable ones:
23 184 C(c)/EDX(x,d,e)/AB(a,b)/EDX(x,e,a)/EDX(e,a,d)/
26 184 C(c)/EDX(x,e,d)/AB(a,b)/EDX(x,e,d)/EDX(x,e,d)/
28 184 C(c)/EDX(x,d,e)/AB(a,b)/EDX(x,d,e)/
Some of these may overlap or be redundant. But not all the AB(a,b) are worthless, because sometimes you can have same a,b with different other values.