(B2 L' U')
(L2 F' D2 B U2 L) // 9 4b2 4e
(B2 U2 B' U2 R2 F')
(R2 L2 F2 R2 D2) // 20 M, +1

Pretty much spend the whole second half of the attempt trying to prove there is a 20 from this dr and it's variation + 4 10move extensions, but apparently there isn't

Only other dr I worked on a bit before finding the 9+2:
F R L B'
L' D2 L' F2 L U // 10 2c3 2e

This is kind of reasonable-looking but switching rzp in 5 with 1 pair is tough on such a scramble:
(R2 F' L D)                   // eoud       (4/4)
(B)                           // rzplr-eoud (1/5)
R F2 R D2 F'                  // drlr-eoud  (5/10)
U2 F2 R U2 B2 R2 D2 U2 L' F2  // fin        (10/20)

D2 R' B'
U' D' L D2 L F2 D // 10 2c3 4e
R (R2 B2 L' B2 U2 B2 U2 L)
(U2 L2 U2) // 22 M, +0

Really happy to have found this one of 2 optimals on this dr, the other one is 3 half turns to trigger on inverse

Other:
R2 D2 R' B
L (D' L2 D L) // 9 2c4, unique 6q 22 opt lmao(had a backup 23 here)

R2 D2 R' B
L (L D L D2 L) // 10 4b3

D' B2 L B
D R' U2 R' U R // 10 4b2 4e, 23 opt

Completely missed this somehow, i have (U' R2 U...), (U' L2 U...) written down for the R F2 R variation which isn't correct. I think what happened is i did (D' R2 D B2 D2 L2 D) on the cube for htr but wrote it down as (U' R2 U...), and then undid (R2 U) instead of (R2 D) and instead of finding the 21 i tried (U' L2 U L2 D2 F2 D) which isn't a real htr:
R2 D2 R' B
L (L D R F2 R)
(U' F2 U R2 D2 B2 D)
(U2 B2 R2 B2) // 21

Could've also found a different 21 if i had switched the 4a2 4e reduction but the one above is about as free as they get

This is the only 20 mallard gives so the 21 opt from 4b3 would've been scramble findable optimal, sad:
(U D2 R F')                 // eofb      (4/4)
U' D                        // arud-eofb (2/6)
(U2 F2 B2 L2 D)             // drlr-eofb (5/11)
(B2 R B2 R' D2 L F2 L' B2)  // fin       (9/20)

(L' U2 D R)
R2 F2 U' B D' F // 10 4b2 2e
(L2 D2 F2 U' L2 D')
(L2 U2 R2 B2 R2 F2) // 22-1 E, +1

Mallard gives this 21 which is kind of findable, i guess? I saw 7c8e with ar-6c3e and switched immediately, didn't look at additional rzp, but looking at it now if you try (F) and notice it's 4c6e the rest is pretty reasonable
(L' D U2 R)                     // eolr       (4/4)
(F)                             // rzpud-eolr (1/5)
(B2 U D2 B' L2 F')              // drud-eolr  (6/11)
(B2 L2 D2 B2 U F2 R2 U' F2 U2)  // fin        (10/21)
Solution (21): U2 F2 U R2 F2 U' B2 D2 L2 F B2 L2 B U' D2 F' B2 R' U2 D' L

L F2 U D2 L'
B R2 F2 D' // 9 4a2 4e
L2 U2 F' D2 F2 R2 F'
D2 L2 D2 L2 B2 // 21 S, +1

Annoying scramble, a lot of stuff to check and short drs but all bad subsets:
(U' L' U)
(B D2 R2 B2 L F) // 9 4b5

D' B D'
B L' R2 B2 U2 L' B // 10 4c4

L F2 L2 F'
L2 U' F2 U' B2 R // 10 4c4

L B2 F' U' R
D B' R2 B' R2 B // 11 4a1 2e

D L' F' D
B L' F R' F2 L // 10 2c4

(F U' R2 F')
(R) U F2 U F2 L // 10 2cx 8e

(U' L' U)
R' L2 F' U2 B' D2 R // 10 2c3 6e

Should've found this dr(i saw jzp on inverse but didn't switch), but all the 20s are 5q, 21 is quite doable tho, and i don't see any other findable for 20s from mallard(edit: of course LMR found the one with 1 to 6 switch eo lmao):
L F' B2 U' R'              // eolr      (5/5)
(D')                       // arud-eolr (1/6)
L2 D2 F2 U                 // drfb-eolr (4/10)
B U2 F U2 B U2 F U2 L2 F'  // fin       (10/20)

R' F' B' R'
(D') D B2 R2 U // 9 4b2 4e
D2 B (L2 F)
D2 L R' F2 R L // 19

There was actually a 4q 18 from the exact variation i actually switched before having the 19 but i didn't check 4q fuuuuck:
R F' B' R'
(D') U R2 F2 D
(F' U2 F D2 B U2 B)
(D2 F2) // 18 S, +0

Returning to this after finding the 19 was probably not worth it but i absolutely should've found this when switching the dr before checking for niss htrs on the variations from the other eo, such a skill issue, idk why i didn't even check if there were +7 paths with 4q

(D' L B' L' F')
F2 B2 D' R U' // 10 4c3
(U2 L B2 L') D2 L2 D2 B2 L
F2 U2 D2 R2 D2 // 24, optimal from dr smh

Such a fucked scramble. This was my only sub11 dr, there is a single findable 23 from variations of this but you need to switch +2 4b2 2e reduction from the right variation on normal:
(D F' L B)
(D) R2 L D F2 R2 D // 11 2c3 4e + 11 4a3 var if you don't do jzp

Other:
(D L2 B L F')
(L' B2 L D L' U) // 11 4a2 4e but fucked, the other variation is 2c4

(R') F R U' L
(B2 R2 D B D F) // 11 4c4

(F' B' D' L' U)
L2 D2 R' B2 R F // 11 4b5 that looks like +3 to htr, sad

I don't see any other reasonable sub23s from mallard except for this but i missed the eo smh:
R' D' B' U' F'                    // eofb       (5/5)
D'                                // rzplr-eofb (1/6)
L' U2 R F2 D                      // drlr-eofb  (5/11)
L' U2 R' D2 L' D2 R2 D2 U2 F2 L'  // fin        (11/22)

B' L' D' R F' // double jzp lol
R F2 D2 L' U2 L // 11 2c3 4e
D' R2 B2 D' R2 D'
R2 L2 B2 // 20 E lol, +2 =(

Found this at like 15 minutes from checking like 2 htrs on the 2 dr variations on normal i wrote down

Also had this but boring:
(L' D)
L B (U2 L' D2 L B') // 9 2c4
(L' U2 L F2 L' F2 U2 R)
(B2 D2 F2 U2 L2) // 22

Other sub11 drs:
B R' L' U R'
D2 B2 U2 F U // 10 2c4

B' U2 F' D
F' L R2 F2 R B // 10 2c4

Apparently you can get a 21 from the double jzp eo, should've checked the ...F R' F2 D2 L B2 L variation on inverse:
B' L' D' R F            // eofb      (5/5)
                        // arlr-eofb (0/5)
L' U2 F2 R' F2 L2 B2 L  // drud-eofb (8/13)
B2 U F2 R2 D' F2 U' R2  // fin       (8/21)

There's also this but 0 pairs on normal moment:
(L' D')                           // eoud       (2/2)
B R'                              // rzpfb-eoud (2/4)
(B' R2 D2 R2 B R')                // drfb-eoud  (6/10)
(F' U2 B2 U2 R2 F' L2 B U2 B D2)  // fin        (11/21)

U' R' U' B
U R' B2 D2 L' U' // 10 2c3 4e
D2 R (F2 R D2 R')
(F2 L2 U2 B2 D2 R2) // 22-1 M, +1

Ugh, 20 22 22 and the 20 is on a doomed scramble and the other 2 had 3movers

Ah yes, how could i miss that(the dr is pretty reasonable to find considering L D rzp on the non-extended eo is good on inverse, but +9 on 2c4? come on):
U2 B2 D B                   // eofb       (4/4)
L D                         // rzplr-eofb (2/6)
(L' B2 L' U')               // drlr-eofb  (4/10)
(F2 L' B2 U2 L' F2 L B2 L)  // fin        (9/19)

Other:
B2 D B
L D (B2 L' U2 R' U) // 10 4a2 2e

B2 D B
U L D' F2 L2 U' L2 U' // 11 4a1 4e

U B2 D B
L (D' F2 D' R' U2 R) // 11 2c3 2e

also this scam:
(U R D2 B D)
F' L2 B D2 F B2 L // looks like +3 to htr but actually 12 4b5

(U F' U)
(R') B R' // fucking jumpscare 6 2c5
(B' L2 F) R2 D2 B L2 B' U2 F'
L2 D2 L R' D2 L' R' // backup 22

I spent like 15 minutes on these 2 extensions i thought were 8movers, ugh:
(U F' U)
(R) R U2 R' B R' // 9 4b3

Very disappointing after a1, apparently the 4b3 in 9 is 21 opt but i missed the M2 insertion before second qt after switch, and the 2c5 is 20 opt with these, come on:
nissy-# solve -t 64 -M 16 R' U' F R U B2 F U F2 U B2 U' L2 D' U2 F2 L' U' F' D2 F2 L2 F R' U' F (U F' U) B
R D2 F U2 R2 B2 U2 B U2 L2 B' U2 F D2 F R (16)
R D2 B' U2 F' U2 L2 D2 L2 B' U2 B D2 F U2 R' (16)
R L2 F' D2 B R2 B' U2 B L2 D2 F' U2 F2 D2 R (16)
R L2 F' D2 B R2 F' L2 B D2 R2 B' U2 F2 D2 R (16)

The only reasonable 21 i see from nissy is this, it's a +1 extension of an htr i had on inverse:
R B2 D2 F2 U2 F D2 L2 F' R2 U2 F' U2 F D2 F R (17)

Would've been better off not finding the 6mover honestly, there's this other 20 but i straight up missed the eo:
R' D2 F B (F)                       // eofb       (5/5)
(R)                                 // rzpud-eofb (1/6)
U L'                                // drud-eofb  (2/8)
B2 L2 U D2 L2 F2 U' F2 U2 F2 L2 F2  // fin        (12/20)

Also had this:
F' R
(U) U2 F D2 F D // 8 2c5

F' R
(U) U2 D2 F D2 F D // extend: 9 4b5

F' R
(U) U2 F2 D F D F2 D // extend differently just for 10 4c4

(U F2 R F B) // 4c6e with 4 pairs(+ FB eo) so switch
D2 R2 D' R F2 L' // 11 4b3
(U2 F2 U) // fake htr

Flip:
(U F2 R F B) D2 R2 D (U2 F2 U) // 11 4b2 2e with 16 variations
R2 F2 R' D2 L
U2 D2 R2 B2 U2 // 21-1 M, miracle +0

you can do this with this solution which almost works but it unfortunately it doesn't:
R2 F2 L' F2 R [U2 D2 L2 B2 U F2 U2] F' B' R' &

[] = R2 F2 U' B2 D2 // 2e2e in 17
& = M2 // 4x on M in 18, basically slice in 18, but it's doomed cause the dr portion is all F2s: R2 F2 L' F2 R' F2

Also found another cooler 20 later(the one i submitted):
(U F2 R F' B)
D2 % R2 D (D2 B2 U')
(D2 B2 R D2 L)
(F2 L2) // 18 2E2E

% = D2 R2 L2 U2 R2 L2 // +2/20

Has to be one of my favorite solves in a while, optimal from the first 3 moves

These are the only 2 sub22s mallard finds so i'm very happy with a 20(i didn't see the rzp for the first cause i already had a 20 and wrote but didn't check the second eo):
(B D L B2 R)                     // eolr       (5/5)
(B)                              // rzpud-eolr (1/6)
U L2 B                           // drud-eolr  (3/9)
R2 F2 U2 B2 U L2 D' U2 L2 D' R2  // fin        (11/20)
Solution (20): U L2 B R2 F2 U2 B2 U L2 U2 D' L2 D' R2 B' R' B2 L' D' B'

B' (B U R F)                   // eofb       (5/5)
(U)                            // rzplr-eofb (1/6)
F2 L' R2 U'                    // drlr-eofb  (4/10)
F2 U2 L2 U2 R' U2 B2 R' U2 R2  // fin        (10/20)
Solution (20): F2 B' L' R2 U' F2 U2 L2 U2 R' U2 B2 R' U2 R2 U' F' R' U' B'

L' D L U' R'
D F B L2 F' U' // 11 2c3 4e
(F L2 U2 F D2 F)
(B2 R2 D2 R2 L2) // S "21", +1

All mallard's sub22s are premoves on 5movers so nah

Cool dr but of course it's 2c5:
(F' L B' U R')
(U2 F2 U B F2 D) // 11 2c5

Also:
(U2 D' F)
U D' L' U2 R' B2 U // 10 4c4

(U2 B2 D' F')
(D) R' U2 F2 D R' U // 11 4c4

[R L] (U2 F2 R')
(U2 B2 U2 L2 F' D) // 11 2c3 4e
(F2 U2 B U2 F D2 F')

[] = D2 R L D2 B2 D2 // 22 S, +0

Also:
R B' R B L
(U2 B2 U L2 B')
D R2 D L2 U R2 U'
F2 B2 L2 // 2E2E in 20, move F2 B2 in dr line for +3 on the spot

Other:
R D2 L' B' R
F2 U R2 L2 F // 10 4b5

R B L' B R
(D2 B2 D' R2 D2 F) // 11 4a2 2e

R' F' D'
(U2 F') L B2 D2 L' B // 10 2c4
(U2 L' D2 R B2 R' U2 L2 U2 D2 R)
(D2) // 22 M, +0

Other:
(L U2 D' F R)
R F2 F D' // 9 4c4

D (D B' L2 D)
L2 F2 R' U2 L' B // 11 2c3 4e

Didn't write this eo, other than that 22 seems the best you can do somehow:
F (D' L R' B)             // eofb       (5/5)
                          // rzplr-eofb (0/5)
B2 L' R2 U' L2 U          // drlr-eofb  (6/11)
F2 U2 L' F2 L B2 U2 L D2  // fin        (9/20)

L' (B' D' B L)
(B' D2 B L2 D) // 10 2c3 2e
B' R2 B' R2 U2 B'
U2 L2 B2 D2 L2 // 21 S, +1

L U2 D B
L2 D L2 U2 L' D // 10 2c3 4e
(R F2 R F2 U2 R')
(F2 B2 U2 R2 F2 D2) // 21 M, +1

There's surely a 21 from one of my drs, but there are just too many decent 10movers. Edit: there isn't, only from a +1 extension of the dr from final

I am actually so stupid, how did i find a +1 to rzp shift 5mover but not the direct +2 to trigger, lmao:
L D U2 B'                       // eofb       (4/4)
                                // rzplr-eofb (0/4)
R F2 D B2 D'                    // drlr-eofb  (5/9)
R2 F2 L U2 F2 U2 L R2 B2 L' U2  // fin        (11/20)

There's also a 19 if you extend by 1(flipped from the output):
L D U2 B                // eofb      (4/4)
F2 L                    // arlr-eofb (2/6)
(R' U2 D2 R B2 L')      // drud-eofb (6/12)
(B2 D2 L2 D' L2 D' F2)  // fin       (7/19)

Other:
U2 L D B
(L D' R D' R' D) // 10 4b2 4e

U2 L D B
(R D2 L D' R' D) // 10 2c3 6e

(D) D2 R2 D'
F' L B2 U2 L F // 10 2c3 4e

(D) D2 R2 D
F B2 L' F R' F // 10 2c4 + ext to 11 4b2 2e

U2 L D B'
R' U' D' R2 L D L // 11 4a2 4e

(B) L' U' D' B
R (R2 D' L2 U2 D' R) // 12 4b2 2e
(D2 R2 B2 U' R2 D)
(B2 D2 F2 L2) // 22 E, +0

Lol, found this at 13 minutes. Also a less silly 23:
(U' B D' L)
(R2 F R2 F' D' B' U')
(B' U2 B' U2 F U2 B2 U2 F')
(L2 U2) //22 S, but +1 with both ways of solving 2e

Missed rzp on (F2 R F U) for a 21, ugh