With Jerry's help I have found a process for one of the <U, R>
antipodes! Note the number of turns required in the h turn metric.
I applied the turn R1 to the position so the program only needed to
search 24 q turns deep.
First Antipodal Process (25 q, 20 q) UR13 = U2 R1 U3 R2 U3 R2 U3 R1 U2 R3 U1 R3 U3 R1 U1 R2 U1 R3 U1 R3 Also.... UR13 ^ 2 = I
Using current techniques, this required a run of about 11 hours.
It remains to be seen how the Kociemba algorithm resolves this
position, and I will try this next.
Although somewhat unrelated, I found a square's group process
for 6 X order 2 (the pons asinorum) by hand, which does
not move the centres:
U2 B2 L2 U2 D2 L2 F2 T2 F2 U2 L2 F2 B2 L2 D2 F2 (16 h, 32 q)
-> Mark <-