Re: your PM, I don't actually have very many of Breandan's f2l algorithms anymore.
This single example should help though:
R' u' R u R
Play around
a lot with using 'R' and 'u' moves, to move the f2l around in interesting ways. Learn all of those tricks on speedcubing.com and play around with them too.
Something else that will help with reducing rotations but doesn't require any algorithms:
Scramble: U' L2 D2 L2 R' D2 U' B L D U' B' U2 L2 U' B2 U R D2 L F2 B2 U2 R2 B'
cross: x2 R' F D' R2
from here, look at the 4 f2l edges that belong in the E slice. UF and UB edges are not solvable with <R, U, L> and nor is the edge in BR (read the ZZ tutorial cride5 linked you to understand edge orientation). So first we solve the f2l pair with the oriented edge (the one in UR).
f2l 1: U L' U' L2 U2 L'
at this point we have UF and UL edges oriented, but if we rotate so we can solve those in <R, U, L> then the edge that is in BR is still incorrectly oriented; we have to take it out first.
However, we don't want to waste moves, so we can join up the edge with the corner in the BR slot
EO setup and join pair: U2 R' U' R
now we can rotate, so that all of our pairs are solvable with <R, U, D, L>. However after rotation you'll notice you can use keyhole to solve one pair and preserve the c/e pair we made.
f2l 2: y D' L' U L D
then the next pair...
f2l 3: U R U' R'
and finally the last...
f2l 4: U L U L' U L U' L'
The idea behind the above example is to orient edges using a cube rotation. There are a couple of rules you can follow (but I urge you to read the ZZ EO tutorial):
- any f2l edge that is in the E slice will keep it's orientation after a cube rotation.
- any f2l edge that is in the U layer will switch it's orientation after a cube rotation.
If you use the idea of placing bad edges in the U layer then you'll only ever require one cube rotation in a solve.
Also, check the f2l section of
www.opticubes.com to find algorithms that use <R, r, U> to solve pairs with bad edges.
EDIT: Just so you know, this advice is only workable for more advanced cubers, and actually using this technique in a solve isn't particularly fast. I only teach it to help you keep track of orientation during rotations, which helps a lot in minimising the amount of rotations you use.
