• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 40,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

u Cube

Member
Joined
Apr 6, 2018
Messages
518
Location
your imagination
YouTube
Visit Channel
New 3x3 Method: SpeedSliver

Hello everyone! Recently I developed a new speedsolving method that I think may have some potential.

Step 1: FS(First Sliver)
This step is like roux first block

Step 2: SS(Second Sliver)
This is a bit different, here you solve the df and db edges while orienting all of the edges

Step 3: TS(Third Sliver)
Here you solve a 1x2x2 sliver unlike the other 1x3x3 slivers

Step 4: WVCP(Winter Variation Corner Permutation)
Here you do winter variation on the last slot WHILE also permuting the corners at the same time. Currently algs are not generated for this.

Step 5: EPLL
Pretty self explanatory

Please leave thoughts below

Walkthrough solve: https://alg.cubing.net/?setup=U_R_F2_L_B_F_D-_B-_R2_U_L_B_D2_R2_L_&alg=x-_D-_U-_R2_U-_M_r-_F_r2_U_M2_U2_r_B- M_U_M2_R_U-_r_R2_U-_r-_U-_M2_U_M_U2_M-_U_M R2_U-_R2_U-_R2_ U-__R-_U-_R_U-_R-_U-_R_U-_R2-_U2_R_U2_R-_F_R_U_R-_U-_R-_F-_R2 U_M2_U_M2__U_M-_U2_M2_U2_M-_U-
 
Last edited:

Thom S.

Member
Joined
Sep 26, 2017
Messages
1,292
Step 4: WCLL(Winter CLL)
Here you do winter variation on the last slot WHILE also permuting the corners at the same time. Currently algs are not generated for this.

What you think about already has a name.
WVCP
Winter Variation Corner Permutation
This is the Spreadsheet

I believe there is no known cuber who learnt it all(can someone back me up or prove me wrong on this?)

+This belongs in the new Method thread
 

Hazel

Premium Member
Joined
Apr 1, 2017
Messages
1,681
Location
in your walls :3
This may not belong in this thread, but I don't think people give Hexagonal Francisco enough credit.

1) Solve 'hexagon' – first layer minus one edge and one adjacent corner
2) Expand the hexagon – complete F2L minus those same two pieces
3/4) Last edge + edge orientation
3/4) CLS
5) PLL

This method seems okay from these steps, but the first two steps are just meh. The hexagon could be fast, but I can see how the expansion would be slow. And so, I'll make some revisions:

1) EH – Expanded Hexagon. Basically, the first 2 layers aside from either the DB edge or the DF edge, as well as any D-layer corner.
2) CSO – don't know what this stands for, found it on the wiki page for Triangular Francisco. It's basically inserting the last D-layer corner while orienting the other 4 corners without regard to U-layer edge orientation. You would do one D-layer setup move (D, D', or D2) to bring the unsolved corner to DFR, excecute the algorithm, then undo the setup move.
3) EODF/EODB – inserting either DF or DB, whichever is unsolved, while orienting the U-layer edges. This is 2-gen <M, U>, and extremely easy to learn.
4) PLL

There are a lot of options for the EH which makes it very accommodating to any scramble. These options include:
- Roux blocks + DF or DB, minus one corner
- First Roux block, solve 2x2x2 in back, insert last E-layer edge
- Cross minus DF or DB, F2L minus one corner
- Petrus block + one F2L pair + last E-layer edge (my personal favorite)
- Just plain blockbuilding
- etc.

I don't know how good CSO algs would be, but I've seen CLS algs and they are very very good, so CSO would either be that good or better.
EODF/DB is very very quick. The best case that isn't a skip is 3 moves (STM), and the worse is 12. Since they're all <M, U>, they're all regripless. Recognition is also pretty much instant. My average time for this step (minus recognition) is 1 second, and I have pretty bad TPS in any method.
PLL, of course, has already proven to be super quick.

When playing around with this method (Advanced Hexagonal Francisco maybe?) I average around 17 seconds. Keep in mind that I have very little experience with this method, and since I don't know CLS or CSO I'm just inserting the corner and doing OCLL, which is most definitely considerably slower. I'm also horrible at blockbuilding, so I was restricting myself in terms of the EH step, so that could too be improved upon a bunch.

Anyway, I just wanted to state my opinion about Hexagonal Francisco and show some evidence that this method may actually be very viable and shouldn't be tossed aside like so many other methods in this thread.
 

Hazel

Premium Member
Joined
Apr 1, 2017
Messages
1,681
Location
in your walls :3
Also on the matter of methods that few use, I think more people should experiment with PCMS :)
Not only is it super fun imo but it's not a bad method – I just got a 15.4 Ao12 with it on my first try today, and I'm being very inefficient at the 4-pairs step. I also definitely think there's a good way of doing L8E, I just haven't found one yet. Right now I'm just solving 3 of the D-layer edges and using EODF/DB on the last one and finishing with EPLL, but that's probably a pretty bad way of doing it.
Some may argue that PCMS is just a worse Roux, but PCMS doesn't use blockbuilding which many beginners would see as a big plus :p
 

PapaSmurf

Member
Joined
Jan 4, 2018
Messages
1,103
WCA
2016TUDO02
YouTube
Visit Channel
A way to do 1llsll for ZZ/Petrus/any method that has eo solved. This is just a concept, but it's almost certainly less algs than 1lll. It should be high 2000, instead of ~3900 for 1lll. Normal F2L, then in right block you solve the FR square. You then conjugate with an orientated corner, just like in 42. Then you recognise the corner case like in 42, and the ep case that goes with it, do an alg, undo the conjugation, AUF then solved. Even though the recog is certainly worse than 1lll, you have 1000 less algs to learn and therefore more time to perfect it. This is only a concept though, but is definitely a good answer to 1lll with CFOP. Doing it should achieve sub 45 movecount a if done right. 20 for eoline and left block, then 7 for the sqaure. You now have 18 moves to conjugate, alg, undo conjugate and auf, with cancellations if you're good. I know this might never be used, but it's a possibility for anyone willing to take it on.
 

Skewbed

Member
Joined
Jan 16, 2017
Messages
114
Location
California
WCA
2015LYON01
A way to do 1llsll for ZZ/Petrus/any method that has eo solved. This is just a concept, but it's almost certainly less algs than 1lll. It should be high 2000, instead of ~3900 for 1lll. Normal F2L, then in right block you solve the FR square. You then conjugate with an orientated corner, just like in 42. Then you recognise the corner case like in 42, and the ep case that goes with it, do an alg, undo the conjugation, AUF then solved. Even though the recog is certainly worse than 1lll, you have 1000 less algs to learn and therefore more time to perfect it. This is only a concept though, but is definitely a good answer to 1lll with CFOP. Doing it should achieve sub 45 movecount a if done right. 20 for eoline and left block, then 7 for the sqaure. You now have 18 moves to conjugate, alg, undo conjugate and auf, with cancellations if you're good. I know this might never be used, but it's a possibility for anyone willing to take it on.

So is it conjugated ZBLL instead of CxLL like 42?
 

shadowslice e

Member
Joined
Jun 16, 2015
Messages
2,923
Location
192.168. 0.1
YouTube
Visit Channel
A way to do 1llsll for ZZ/Petrus/any method that has eo solved. This is just a concept, but it's almost certainly less algs than 1lll. It should be high 2000, instead of ~3900 for 1lll. Normal F2L, then in right block you solve the FR square. You then conjugate with an orientated corner, just like in 42. Then you recognise the corner case like in 42, and the ep case that goes with it, do an alg, undo the conjugation, AUF then solved. Even though the recog is certainly worse than 1lll, you have 1000 less algs to learn and therefore more time to perfect it. This is only a concept though, but is definitely a good answer to 1lll with CFOP. Doing it should achieve sub 45 movecount a if done right. 20 for eoline and left block, then 7 for the sqaure. You now have 18 moves to conjugate, alg, undo conjugate and auf, with cancellations if you're good. I know this might never be used, but it's a possibility for anyone willing to take it on.
I believe you've just recreated CTLS
 

PapaSmurf

Member
Joined
Jan 4, 2018
Messages
1,103
WCA
2016TUDO02
YouTube
Visit Channel
I almost recreated CTLS. I created something better though. It permutes 5 edges and 4 corners relative to a conjugated one, not 4 edges relative to a conjugated one and 4 corners related to a conjugated one
 

WoowyBaby

Member
Joined
May 27, 2018
Messages
765
Location
Neptune
WCA
2018ISOM02
YouTube
Visit Channel
EDIT- I HAVE BASICALLY CONFIRMED THIS IDEA IS GARBAGE.

Alright so I have been thinking of an intermediate LS+LL method that solves the last 10 pieces of your solve in 4 steps: CO, CP, EO, EP.
I will call this ILSLL (Isom Last Slot and Last Layer) what else would I call it lol. I have done some exploring and I believe that this is the best order of steps for a 4 step method. For example, EO-CO-CP-EP or ELS-CO-CP-EPLL is definitely not as good as these method steps.
There are 44 algorithms, 29 excluding CO. CO by far the shortest step, because some are just 3 move inserts.

Anyway, why is this better than the normal way? Well, compared to solving F2L + 2-look OLL + PLL, it has less algorithms and *more moves. (38.3 vs. around 34-36) and recognition is equal on the last step. (So maybe it’s not better lol.)
I want to make an intermediate method that doesn't have a crazy number of cases like VLS, ZBLL, MGLS, etc.

Also, EP algs are totally 2-gen so they are really nice algs. Most CO's are 2-gen aswell.

Here is the sheet I've been working on with all of the algorithms: https://thisIsTheLinkToTheSheet.com (lelz)
These algorithms are not the best they can be yet so I will keep imrpoving them! (sidenote: ergonomics more important than movecount)

I have looked at the wiki and forums and made sure nothing like this has been posted, although this kind of has a similar idea to MGLS or ZZ-CT, but definitely different.

Idk what else to say honestly so here's some examples I guess:
LS+LL scr: U2 B L U L' D2 B' U B2 R2 B' U2 F D2 B' L2 F'
U R U' R2 F R F' // CO (Blob)
l' U R' D2 R U' R' D2 R2 x' // CP (Normal Adj.)
U M' U M U2 M' U M // EO (2 Adj. Flip)
U2 R' U' R' U' R' U R U R U2 // EP (Adj. Parity #1)
LS+LL scr: U F2 D U2 R2 D' F2 R2 B U B' U2 F2 U R2
U R U' R' // CO (Gun)
R U' R' U2 F R' F2 r U' r' F2 R // CP (Bar Left)
U r R U R' U' M U R U' R2 // EO (1 slot + 1 top)
U2 R U R' U2 R U R' U' R' U' R2 U' R2 U2 R // EP (Wd Perm)

Once I learn all the cases with good algs I will put a speed comparison of ILSLL vs. F2L->2lookOLL>PLL, and if ILSLL is measurably worse, I’ll ditch the idea but for now I’m not sure (right now when I get easy ILSLL cases I am faster than normal but of course slower when I don’t know that W perm or whatever)

If there is something obviously horrible about this then tell me below!

AGAIN, THIS IS TRASH
 
Last edited:

xyzzy

Member
Joined
Dec 24, 2015
Messages
2,877
Alright so I have been thinking of an intermediate LS+LL method that solves the last 10 pieces of your solve in 4 steps: CO, CP, EO, EP.
I will call this ILSLL (Isom Last Slot and Last Layer). I have done some exploring and I believe that this is the best order of steps for a 4 step method. For example, EO-CO-CP-EP or ELS-CO-CP-EPLL is definitely not as good.
There are 44 algorithms, 29 excluding EO. EO is by far the shortest step, because some are just 3 move inserts.

Anyway, why is this better than the normal way? Well, compared to solving F2L + 2-look OLL + PLL, it has less algorithms and fewer moves.
I want to make an intermediate method that doesn't have a crazy number of cases like VLS, ZBLL, MGLS, etc.

Also, EP algs are totally 2-gen so they are pretty ok algs. Most CO's are 2-gen aswell.

Here is the sheet I've been working on with all of the algorithms: https://thisIsTheLinkToTheSheet.com (lelz)
These algorithms are not the best they can be yet so I will keep imrpoving them! (sidenote: ergonomics > movecount)

I have looked at the wiki and forums and made sure nothing like this has been posted, although this is kind of a little like MGLS.

Idk what else to say honestly so here's some examples I guess:
LS+LL scr: U2 B L U L' D2 B' U B2 R2 B' U2 F D2 B' L2 F'
U R U' R2 F R F' // CO (Blob)
l' U R' D2 R U' R' D2 R2 x' // CP (Normal Adj.)
U M' U M U2 M' U M // EO (2 Adj. Flip)
U2 R' U' R' U' R' U R U R U2 // EP (Adj. Parity #1)
LS+LL scr: U F2 D U2 R2 D' F2 R2 B U B' U2 F2 U R2
U R U' R' // CO (Gun)
R U' R' U2 F R' F2 r U' r' F2 R // CP (Bar Left)
U r R U R' U' M U R U' R2 // EO (1 slot + 1 top)
U2 R U R' U2 R U R' U' R' U' R2 U' R2 U2 R // EP (Wd Perm)

If there is something obviously horrible about this then tell me below!
I was going to mention you'd need twice the algs for L5E (for either FL slot or BR slot) but I thought about it more and realised I was wrong, because you can just do EO in the FR slot orientation.

Seems like a reasonable method overall for 4-look LSLL, but there's no good progression to a 3-look method. Combining CP and EO seems to take 100+ algs if I'm counting it right.

Also, if your move counts are right, that adds up to 38.3 moves for LSLL; I just did a bunch of LSLL solves with basic edge control and 3LLL (with the usual algs I use, not move-optimal ones) and that came out to ~34 moves.
 

Skewbed

Member
Joined
Jan 16, 2017
Messages
114
Location
California
WCA
2015LYON01
I saw a video with a comparison between ZZ and Roux that I thought could be applied to other methods.

Basically, although the methods may seem very different, ZZ could be thought of as a worse version of Roux.


Roux

ZZ

First Block

Left Block

Second Block

Right Block

CMLL + L6E

EOline + ZBLL

If you compare how they solve certain groups of pieces, F2B seems a little harder than ZZF2L, but CMLL and L6E is way better than EOline and ZBLL.

I was wondering if there are any other good 3x3 methods that this could be applied to like this?
 
Top