(L D' F2 L) //EO
(D' B) (D R2 F' D2 F') //DR 4b2 2e 11/24
(L2 U^ F2 R2 L2 D'^) //HTR 17/24
(R2 F2 D2^ R2 B2 D2 L2) //slice 24/24
^=w

WTF WAS THAT SCRAMBLE!!!

D2 B2 U //EO+jzp 4c6e 3/18
B2 R B2 L //DR 4a2 4e 7/18
 U2 B (D2 F2 D2 R2 F) //HTR 14/18
(R2 U2 B2 D2) //solved 18/18

7 moves to 4a2 4e. There was also 8 moves to 4c3 (L2 F) (U L U' R' D2 L)

R' L' (F L) //EO 4/23
(B) F2 D2 R2 D' B' //DR 4b2 2e 10/23
(L2 [U' F2 B2 R2 U') //HTR with extension 16/23
(L2 F2 L2 U2] B2 R2) //slice 22/23
[]= D' L2 R2 F2 U' B2 L2 B2 U D //+1 23/23

Literally the only sub5 EO on the scramble gave 10 move 4b2 2e :D and from 10 bad axis too.
DR is 22 optimal but all optimals seemed quite hard. But I missed a trivial DR variation that has an easy 22...

(U2 D2 F') //EO 3/23
(L') B2 L2 D' L2 D L' //DR 2c4 2e 10/23
(U^) F2 D R2 B2 D R2 * D'^ //HTR 18/23
B2 R2 //2e2e+slice 20/23
*= R2 D2 L2 R2 U2 L2 //slice 6-3 = 23/23
^=w //+0

Optimal from this DR but a variation has 22.

(R' [D F' B) //EO4/22
(R' L2 D L2 U2 R)] //DR 4c3 with at least 16 variations! 10/22

First I went to normal and found that just doing U there will reduce to 4b2 4e.
Then I got back to inverse and tried all variations with that U reduction.
2 variations gave a case that was directly setup to a trigger:
[]= (D F B' L' R2 U L2 U2 L') //new DR 4c3 in 10
U^ (U^ F2 R2 U) //HTR 15/22

First I saw this leave slice: invert the last move (F2 R2 U2 L2 F2 L2 F2 U2) //slice in 22 but +2
Then I decided to NISS the HTR since 15 to HTR is kind of short and it seems kind of blocky.
on normal this is easy to see: 
F2 R2 U2 L2 F2 L2 F2 //slice 22/22  
^=w //and this slice is much better and gives +0 for 22/22

I dont really know how to analyse this since there are so many variations :D But I feel like my approach was reasonable implementation of "whack the variations".

EDIT: as far as I can tell there are no findable sub22 from this set of DRs. There is a 19 if you just add a random U2 before the F B in EO but good luck finding that...

(F') U2 D' B //EO 4/24
F2 D' , L U2 D R' D //DR 2c3 4e 11/24
(B2 [L U2 L' F2 L']) //HTR 17/24
(D U' L2 F2 U D') //slice 23-1 = 22/24
[]= L2 R' F2 L R2 U2 R' //+2

Annoying solve. Had this slice in 22 and 3 different slices from 2 different DRs in 23 but still no better than 24. :/

(U' D' R F' B) //EO 5/21
(D) (L' F2 R' U) //DR 4b2 4e 10/21
R' U2 F2 R //HTR 14/21
B2 L2 B2 D2 //S-slice 18/21
D2 B F' R2 F B' //6-3 = 21/21

Nice! Especially on a scramble that had only 2 4 move EOs, I usually struggle with those but got lucky this time.

Also had this DR:
(U' D2 F' R' L') (D) L2 U F2 B2 D //4b2 2e in 11 last U can also be U'. The U' variations is 23 optimal and U variation is 25.

Nice mo3 as well 22, 21, 21 = 21.33

(B U) //EO 2/21
(R' F') //rzp 3c2e 4/21
D2 L B L B' //DR 2c3 6e 9/21
R' (U2 R2^ B2 R'^ U2 R'^) //HTR 16/21
U' D' L2 U' D //slice 21/21
^=w //+0

I have been getting a lot of good solves recently. Maybe I am actually improving!!

Only had time to check 2 mover and th3 movers on this scramble since completely checking a 2 mover takes a looong time for me.
And I found this DR in 9 pretty quickly so fidured that I will at least have some solution I can most probably find on that so no need to push till 35-40minutes just searching for DRs and risk DNFing.

D R F //EO+rzp 7c8e 3/22
U2 [L D2 U' L D] //DR 2c4 9/22 but hopeless on both sides. All short reductions go to 2c3 6e or 8e.
[]=L' R2 U2 D' L U //extended DR 2c4 10/22
(L'^ U2 B2 R'^ D2 R^ D2 R^) //HTR 18/22
(U2 R2 F2 B2) //slice 22/22
^=w //+0

Good solve. Managed to discard the hopeless 9 move DR early enough so I still had time to find the finish for the extension

U2 D2 B D' //EO + rzp 3c2e 5/21
(B2 L2 B L B) //DR 4a1 4e 9/21 wow
B2 D2 [F2 R2 U2 R //HTR 15/21
D2] F B' R2 F' B //slice 21-2 = 19/21
First I only found this:
[]= F2 R L F2 L U2 R2 L2 //+3 for 22 final

But much later I noticed that there is actually a blockyfish rewrite there!
[]= R2 U2 R2 ^ F2 ^^ R' D2 //new slice 20/21
^= M'
^^=M //+1 solved 21/21

U L F' R //EO 4/25
F' , D L2 U' B' L2 D2 B //DR 2c3 4e 12/25
R2 U' F2 U' L2 * U //HTR 18/25
^ L2 F2 ^^ D2 F2 //2e+slice 22/25
*= L2 U2 L2 U2 L2 U2 //slice 6-4 = 24/25
^= E
^^= E' //+1 25/25
Tough scramble. Only (set of) DR(s) I had. Optimal from all variations. 
Mallard gives 22 if you randomly decide to NISS a 4 move EO or 23 from 2c5 in 9. 

EDIT: Oh, nevermind I had R U2 L trigger turned of. Mallard gives 20 from 4c6e rzp but I still suck at those unless they are jzp or like 1 move away from jzp

(F' L2 U' L) //EO 4/23
(F) (U D2 L2 D R2 F') //DR 2c3 2e 11/23
(R2 D' L2 D'^ F2 L2 U^) //HTR 18/23
(L2 D2^ F2 R2 F2) //slice 23/23
^=w //+0

DR is 22 optimal but needs either reduction to 4a1 4e and then switch or 5qt. Reduction+switch is findable but not easy unless you use bounding.

B' D R' //EO+rzp 4c4e
(F B R2 F B2 U) //DR 2c4 9/23
F (F2 U2 B' L2 B L2 [B) //HTR 17/23
(D2 R2 D2 F2] R2) //slice 22/23
[]= F L2 D2 L2 F B //+1 23/23
Optimal from this DR but the obvious variation (F' B' L2 F B2 L) has a findable 22...

(R D F' L) //EO+rzp 4c4e 4/25
(B2 R2 U2 L2 B2 U' F) //DR 4c3 11/25 
Apparently I'm blind this week. There was a basic 10 move DR from that rzp but I didnt see it somehow...
(U^ R2 U' R2 U2 L2 B2 D^) //HTR 19/25
(L2 F2 R2 F2 D2 F2) //slice 25/25
^=w +0 // optimal from DR 

And of course the DR I missed is 22 optimal :/ And optimal is easy to find.
23.66 mo3 this week with insanely stupid misses on 2 and 3. Should have been 22.00 without anything magical...

L' B' //EO 2/22
D' (R L D' F2 L2 U) //DR 4b2 4e 9/22
(L2 D2 F2 B2 L^ F2 U2 R'^) //HTR 17/22
(D2 L2^ D2 B2 L2) //slice 22/22
^=w //+0 22/22

Variation of this DR is 21 optimal but missed that even though it is perfectly findable... 
Oh and there also seems to be a 20 with double slice stuff. Why am i so bad :D

F' (L2 R' F') //EO 4/24
(L) (U R2 D L2 U2 D R) //DR 4b2 2e 12/24
R2 L2 U'^ B2 U'^ //HTR 17/24
R2 U2^ L2 F2 U2 F2 U2 //slice 24/24
^=w //+0

-

(U' B' R F) //EO+rzp 4c4e 4/20
F2 R2 F2 U' L //DR 2c3 4e 9/20
D' B2 R2 D ^ B2 U //HTR 15/20
R2 L2 D2 ^ L2 //slice 19/20
^=E2 //+1 20/20

NICE!