update on my progress with AHTA:

So I have been busy learning the corner cases and have decided to learn only a subset of them. normally using the parity+EPOCBL system there would be 63 (49 EPOCBL, 7 normal OCLL, same 7 but on D) cases after parity. I figure a cube rotation wouldn't hurt since it drops the alg count to 35. Counting mirrors as the same case drops the count down to 28. also if we sacrifice a couple moves the cases u/u, t/u, and t/t can all be set up as H and Pi cases which brings the count down to 25. other cases could potentially reduce to a single case as well but they aren't worth it imo because of extra moves and harder recognition.

subtracting the 6 ocll cases (antisune counted as mirror) that most cubers already know this results in 19 unique cases that need to be learned and this system still provides very nice CO for speed. I will list the algs I am going to personally use below.

(top/bottom)

alg) (mirror?) (if I have learned this case yet or not)

sune/sune: R' U' R' U R U2 D R' D' R (yes) (yes)

anti/sune: R' U' R U2 R (U' D) R2 U R' (yes) (yes)

H/sune: U' R (U' D) M2 U' r2 U R' (yes) (yes)

U/sune: R' U2 R' U D R2 U2 R' U' R (yes) (yes)

T/sune: L D' L' U' r2 U2 L' U' L' (yes) (yes)

L/sune: R D r2 R' U R U2 r2 R' (yes) (yes)

Pi/sune: L U D L' U D2 L' U2 D L' (yes) (yes)

H/H: R2 (U' D) L (U D') M2 (U D') L (no) (yes)

U/H: R D R' U D2 l2 U2 R' U' R' (no) (yes)

T/H: L D L' U' D2 r2 U2 L' U' L' (no) (yes)

L/H: L U' L2 R2 U' L' l2 D U2 L (no) (no)

Pi/H: R' U' L2 U R2 (U' D) r2 R' (U' D) R' (no) (yes)

L/U: D' L' U' L' U2 L U' L (no) (no)

Pi/U: L' U R U2 R' U L R U R' (no) (no)

L/T: U R U D2 R' U' D2 R' (no) (no)

Pi/T: L' U' L l2 U' l2 U' L' U2 L (no) (no)

L/L: L U' D' L U D L' (no) (yes)

Pi/L: L' U' L2 U' L' U2 D' L2 U' L' (no) (no)

Pi/Pi: L U D L' U2 D2 L U D L' (no) (yes)

also does anyone have some advice for the final stage? with belt solved there are only 8 edges needing to be permuted and all M edges are in M and S edges in S so I am guessing it would be very doable to generate algs for all the cases but idk. right now I can usually get this stage done in 1 or 2 algs but it isn't too rare for me to run into a situation where I need to use 3, I feel like this is the worst part of my solve right now besides corners.