D B L U' F ////EO+rzp
(L2 U2 R2 D F2 R) //2c4 in 11
(D R2 D' R2 D F2 D') //HTR in 18
(F2 L2 U2 F2 L2) //slice in 23
widen all qts = +0/23

L U F2 D' //EO+rzp 7c8e
(R' F L' B) //4c3 in 8
R D2 R2 B2 L' B2 L //HTR in 15
U2 B2 U2 L2 F2 //M+S in 20 but the S doesnt quite work. Inserting S moves to B2 and F2 gives VR but not solved. But we can alter that by widening the R2 and L2 on the M slice. So after DR.
R^ D2 L2 F2 L' F2 L^ D2 [F2 D2 R2 F2] //still M+S in 20 but now:
[]=B' F' L2 D2 F B //slice in +0/20
^=w //+0/20

Got SO lucky that I tried this BEFORE trying NISS HTRs since there is R (B2 L' F2 L) HTR in 12 that I wasted all of my remaining time on and would have probably missed my DR finish if I saw this HTR first... It just gives the cursed S/S post-HTR shape and so do all the extensions I tried... Such a bait.
Best possible result from that HTR seems to be 21 via +6 to double slice and +3/+0. I actually had that double slice written down but no point solving it since I already had a 20 and 20 is the best possible it can give.

F B' (L B) //EO
U , L2 D2 B2 R U' //2c4 in 10
L' U2 F2 L'^ F2 L^ D2 F2 L' //HTR in 19
U2 R2 U2 F2 B2 //slice in 24-1 = 23

Lots of 8 move HTRs via L -> NISS -> 1 move to sq+sq and lots of 9 move HTRs by doing L or L' -> 2 moves to sq+sq...

Only decentish DR I had...

U (D' L U)
L U2 F' R D2 L' //4c3 in 10. Only decent-ish DR I had...
D2 [B D2 F2 L2 B' D2 B] //HTR in 18
R2 F2 L2 //double slice in 21
L2 F2 D U' R2 F2 D' U //8-6 slice in 23
[]=Bw' R2 F2 U2 B D2 Bw //+0 blockyfish thing

(R F2 U' L) //EO+rzp
(R2 F2 D F) //4c4 in 8. Second solve in a row with DR in 8 :D
(R2 U L2 U) L2 F2 U' L2 U' //HTR in 17
L2 B2 U2 F2 //slice in 21
widen all QTs = +0

Horrible subset but luckily the optimal is the first thing you check. Shortest possible reduction to 4b2 and only one of them (from this side) that goes to 2e. 

D' F U F //EO+rzp 3c2e 
(F2 R U R) //4b2 2e in 8
R2 B2 U B2 U2^ B2 D' //HTR in 15
F2 B2 L2 U2^ //slice in 19
^=+0

Wow! What a scramble... EVERYTHING was rzp in 4. Didn't even have nearky enough time to NISS every rzp I wanted but luckily this one gave 4b2 2e in 8 at 35 minute mark so still plenty of time to finish it.

U R' U' //EO+rzp
R2 U2 L2 F' R2 F' //2c3 6e in 9
R (B2 L'^ D2 F2 R) //HTR in 15
(F2 R2^ D2 F2 L2 U2 F2) //slice in 22
^=M2 //+1/23

(B' L U) //EO
(F2 R) (B L' B' L') //2c3 2e in 9
(F' D2 L2 F'^ R2 F^) //HTR in 15
([U2 F2] R2 U2) //double slice in 19
[]= U' D' L2 D U' //slice in 22
^=w //+0
DR opt.

Also had
(B' L U) (L R') (U2 B' R' B' R) //4b2 2e in 10

(R) U' B L //EO
B' , U' L2 U2 F' U' B' //4b2 0e in 11 
U F2 R2 F2 R2 U' //HTR in 17
B2 R L F2 R' L' U2 //solved in 24

The other variation is 4b2 2e. There is s slice in 22 but it is +2...
Both DRs are 24 opt.

(L2 F2 D F B') //EO
F2 D B2 L' D2 R //2c4 in 11, but there is D F B on one side and jzp sequence on the other so more equivalent to like 2c4 in 9-10
(U) D'^ F2 L2 D' F2 D^ //HTR in 18
B2 U2 B2 R2 //slice in 22
^=w //+0/22

Literally the only (set of) DR(s) I had that was sub12 and not 5qt... So got quite lucky here

(U) U R' U //EO
B L' , B F L2 U2 L //2c3 4e in 11
U2 B' D2 B' R2 * F //HTR in 17
L2 U2 L2 F2 //2e+slice in 21
*=R2 B2 L2 B2 R2 F2 //slice in 6-4 -> 23
+1/24
Also had another slice in 23 from (B)->NISS but that was +1. There is a 23 from another slice in 23 from the same HTR but I missed that...

F' (B2 L' B') //EO
(L) L2 U R //4b2 4e in 8
(D' R2 [U L2 U F2 R2 L2 D']) //HTR in 17. HTR through 2c3 6e but it is 8 move DR so check every path...
(R2 B2 R2) //slice in 20-1=19
[]= Uw L2 U F2 R2 L2 Dw D2 //+1/20

Also looked at:
F' (U R B') (R) L2 U R //2c3 6e in 8. Also 20 opt with 9 move linear HTR just like the DR I used.

(F) F' D' F //EO+rzp
(L2 U2 F2 L2 U R') //4c3 in 10
(F2 U R2 U^ B2 U2^ R2 D^) //HTR in 18
L2 D2 //2e2e in 20
D2 L2 B2 R2 U2^ B2 L2 F2 //8-5 -> slice in 23
^=w //+0/23

Also looked at:
D' F U' L' , D' , B D2 F D //4c3 in 9, 23opt but unfindable
F U' D2 L , U (F' U B2 U' B2 U) //4b2 2e in 11 but with a nice jzp sequence
U' D' F L (R2 D' B R2 D2 F) //2c4 in 10, 24opt

F' R2 D //EO+rzp 2c2e
F' L D2 B2 L' B2 L' //4a1 6e in 10
B (U2 B' L2 B) //HTR in 15
(L2 D2 R2) //2e+slice in 18
(R2 D2 F2 U2 L2 F2 D2 B2) //solved 8-5 => 21

Holy overwhelming scramble...
Had a DR written down from the EO that gives 19:
F' L F' L , U2 B' L2 U2 B D
But didnt see the correct variation: R2 B R2 D2 F' U because this was quite a sprint scramble :D

(F L' F D') //EO+rzp 8c8e
(R2 B' L2 F R L) //2c4 in 10
F' L2 U2 F' L2 F D2 F' //HTR in 18
D2 [L2 B2 L2 B2] //M slice in 23
[]= R' L' U2 R2 D2 R' L //+3 but 3 also cancel with inverse so +0/23

There is this weird 21 from here but I dont know how one would find that:
nissy-# solve -o -t 64 R' U' F D2 F L' D' R' D2 L' U2 R' U2 L2 B2 D' B2 L2 R' D B2 F' R' U' F (F L' F D' R2 B' L2)
F B' R L F' L' F L2 B' L F R2 D2 B (14)

First you go to DR-2s 8fe (I tried that) but then weird stuff happens :D It is almost like a DR flip and then the 7 move weird HTR trigger but not quite.

Only thing I had was:
F' D' F' L (B2 D F2 U2 B) //4c4 in 9 but it was hopeless

(B2 D2 R') //EO
(B) U2 R2 U' L2 B //4c3 in 9
(U' F2 L2 U2 L2 U' L2 U') //HTR in 17
(F2 L2 U2 B2 U2 B2) //solved in 23-1=22

DR opt.

Also had:
(B2 D2 R) (D) (B' U2 L2 B' U) //2c4 in 9. 23opt

Overwhelming scramble with so many 3 move EOs

([U' B2 D]) //EO
B' R, B2 L2 F R' //2c3 2e in 9 with double slice potential.
B U2 {F2 U2 F'} U2 F' //HTR in 16
R2 U2 F2 L2 //double slice in 20
[]= Uw' B2 Dw' D2 //slice +1/21
{}= FW F U2 Fw' //+1/22

DR is 20 opt but all 20 and 21 movers need 5qt :DDD

U (B L D') //rzp 4c2e
(R B2 L2 B' U2 B') //4b2 2e in 10
(L' D2 R') //HTR in 13
(B2 R2 D2 F2 L2) //wrong slice in 18
(L2 D U' F2 D' U) //solved 24-4/20

B' R2 U R //jzp
B R2 L2 [F' U2 F //2c3 2e in 10
(D' F2 D2^ F2 D^ F2 D^) //HTR in 17
(R2 U2 L2 [F2 B2) //slice in 22-1=21 but kind of in 20 since there is sliceable B2 F
[] = (B F2 R2 B L2 R2 B') //slice in 21
^=w //+0/21

Somehow missed this 8 move DR. I have the EO written down and I thought I checked it. Must have undone some moves wrong and thus had a wrong cubestate at that point...
R B' U R'             // eolr       (4/4)
                      // rzpfb-eolr (0/4)
D2 B' F2 D'           // drfb-eolr  (4/8)
(D2 R2 B R2 F D2 F')  // htr-drfb   (7/15)
(R2 F2 D2 R2 U2)      // fin        (5/20)

D B' U //EO
R' F' , L B' L2 R2 F //2c4 in 10
B2 [R U2 R' F2 R] D2 R' //HTR in 18
B2 U2 L2 U2 //slice in 22
[]= Rw' R2 U2 R' F2 Rw //+1/23

I had so many 10 move DRs but all were 4-5qt... So annoying, probably missed something simple it felt like.

B U' R D2 F //EO+rzp
(F2 D2 R2 D' L) //4b2 2e in 10
(F2 U' F2 D') //HTR in 14
(F2 L2 [F2 D2 R2 D2 B2]) slice in 21
But there is a slice shortening:
[]= {B2 U2 L2 U2} F2 //-1/20
There is no "direct" +0 for the slice but we have blockyfish also:
{}= U2 L2 U2 B2 //different slice in 20
And now widen all U/D layer moves gives +0

There is still F2 B2 in the end but it isnt sliceable unfortunately.

This is also the best cubelib can find with my settings :)

L2 U2 F U //EO+rzp
(B2 D2 R2 B2 R F') //2c4 in 10
(L' B2 L' F2 L D2 F2 L) //HTR in 18. about 1 million linear 8 move HTRs to check...
([F2 L2 U2]) //double slice in 21
[]= U2 D2 B2 R2 D2 //slice in 23
widen all qts for +0
DR opt.

Such an annoying attempt. I wrote down 10 move DRs from 8 different EOs but all were 2c4 or worse subset :D

D2 L' U2 F //jzp 
D U R2 U D' R //4b2 2e in 10. LOTS of variations.
L2 B2 U L2 U //HTR in 15
R2 B2 U2 F2 R2 F2 //slice in 21
widen all for +0

One of the variations has a 20.

What a scramble. Several 3c2e rzps in 4. And even more 4c4e and 4c2e.


 

(F) R' L' F //EO
(R) L2 F2 U' D R' //2c3 4e in 10
B2 U F2 U'^ L2 B2 U'^ //HTR in 17
F B' U2 F B' //slice in 22
^=w //+0
 
DR opt. But variation of the DR is 21 opt.

L (F L') //EO
(U') (F2 R2 U2 F2 B U') //4b2 2e in 10 with variations
D2 B (B2 D2 R2 F) //HTR in 16
(U2 L2 D2 B2 R2 B2) //slice in 22-1=21
+1/22

DR variation is 21 opt...

Other DR I looked at:
L' (F L) B U2 B D R2 F2 U //2c3 2e in 10, 23 opt

(D R' U) //EO+rzp
L' B2 R L2 D2 B //2c3 6e in 9, +1 extension
R' B2 D2 F2 D2 [R U2 R //HTR in 17
B2 D2 F2] //slice in 20
[]= L F2 L F2 L2 B2 //+0/20

What a ridiculous set of scrambles! 20,18,20 = 19.33 PB mo3.
And I even missed an 18 on the A1 and there is a simple 19 on this one as well... 
So the set is 18, 18, 19 = 18.33 findable :D

L2 F2 U2 R' //EO
B' (U R2 F' D' B) //4b2 2e in 10
D' B2 [D //HTR in 13
B2 L2 D2] R2 //slice in 17
[]= Dw B2 L2 Dw' D' //+1/18

LOL what are these scrambles! :D

D B R' D //EO+rzp
L2 B U2 L //4b2 2e in 8
(L2 D2 [F U2 F]) //HTR in 13
(R2 B2 U2 R2 D2 F2 L2) slice in 20-1=19
[]=Fw U2 Fw' F2 //+1/20

This also has another slice in 19 but no +0:
L2 D2 B U2 D2 R2 F R L' B2 R L' //slice in 20-1 but doomed

There is also an 18 from:
D' R (D' L)               // eolr       (4/4)
                          // rzpud-eolr (0/4)
F2 R2 D B'                // drud-eolr  (4/8)
F2 U' B2 R2 D' R2 U (U')  // htr-drud   (8/16)
(L2 F2)                   // fin        (2/18)
But just no time for that at my level since that is 2c4 and the EO I used already gives 2x 4b2 in 8 and 2x 4b2 in 9...

(D' L R' F) //EO+rzp 2c2e
(U [L' F2 D2 L' D2 L) //2c3 2e in 11 with nice jzp sequence
D F2 R2 D' R2 U' //HTR in 17
[R2 B2 D2 F2 //double slice with 3qt on both sides in 21
[] = R L2 U2 F2 L' F2 R B2 U2 F2 L2 //slice +1/22
So after DR we have on normal:
D F2 R2 [D' R2 U' L2 F2 U2] B2
[]=Dw' R2 Uw U2 L2 F2 Uw2 //+1/23

(R2 U) //EO
F' (R' B2 L' R F) //4c3 in 8
(L) R' F2 U2 D2 L'^ //HTR in 14
U2 B2 U2 F2 R2 D2^^ //slice in 20 
^=M
^^=M' //+1/21

DR opt.

There is a 20 from 4 mover to 10 move DR or +3rzp on the 2 mover ->11 move DR.
But no time to check those. I already had 11 rzps in 2-4 moves that I needed to NISS just from the 2 move EO :D What a sprint scramble...