The New Method / Substep / Concept Idea Thread

[napstablook22]

I have made a made a method with mah frends. I call it csr. That is the combination of our last name initials. It starts with eo tho


Sent from my iPad using Tapatalk

 

[PetrusQuber]

Okej, but i think Petrus don´t use the F2L method.

I will give you an pdf.
Please tell me if this has potential. I realy think this approach HALL with A34 can be a new member of how to solve LL.
It can be an alternative for lazy cubers
What i can feel is that this way can for sure bring cubers to reach better PB becaus it is so simple and straight forward.
I realy love this way...

But once you’re finished with your HALL, you need to do ELL. The edges haven‘t been oriented or permuted. Also this is pretty similar to Lars Petrus’s beginner LL system. https://lar5.com/cube/fas5.html

 

[brododragon]

Edges first Idea
1. 2x2x3 block
2. remaining edges
3. solve BDR and FDR corners
4. last 4 corners

How many algs would L6C be?

 

[Tjacke]

But once you’re finished with your HALL, you need to do ELL. The edges haven‘t been oriented or permuted. Also this is pretty similar to Lars Petrus’s beginner LL system. https://lar5.com/cube/fas5.html

I think many cubers are to much into orienting pieces and permuting edges and corners rather that solve the top layer. This idea is built to not perute, prepair, ELL. Maybe many cubers like that? but after HALL you need to pick one of 4 PLL then all edges are in place and then you have to rotate the edges in worst case. This is a straigt forward solution.

Try do F2L or any other F2L without thinking on bad edges, just solve F2L and then use HALL.

In many cases all edges will be permuted by it self and rotate the edges are in my opinion as fast as if you wast on preparation algorithm in Petrus approach.

With HALL you will never do anything for prepairation, in each step you can solve the cube.

This is a new approach and let some more try this. It is a great way of solving LL and i cant find this close to be same as Petrus, becaus with Petrus you need to think and prepair for the last layer so the edges get in place. If you do ONE step to move an edge while building up the F2L with Petrus you have wasted time and moves. With HALL you don´t need to worry by doing something wrong.

 

[PetrusQuber]

I wasn’t questioning how good your LL method was, just curious about the last stage after HALL. That makes sense now, it could be a good way to do LL easily, but I can’t judge alone. I don’t agree however that doing EO in F2L is wasting time though.

 

[Tjacke]

I wasn’t questioning how good your LL method was, just curious about the last stage after HALL. That makes sense now, it could be a good way to do LL easily, but I can’t judge alone. I don’t agree however that doing EO in F2L is wasting time though.

What can happen right after HALL is that all edges can acctualy be permuted correctly, then you need to do one of 4 PLL. and done.
Or all edges can be in right place but not facing right, then you need to rotate the edges.
You have to calculate how much time you spend to prepair for getting pieces correct, while doing F2L.

What i can see with OLL is that you need to recognize alot of colors and that can take time.
Another thing With HALL is after the first trigger, you might endup with a good pattern that you can solve easy, if it is not a good pattern you do the Sune and now another good pattern can show up.

If you dont see any good pattern that you can solve easy, just do PLL and then rotate edges.

If you absolutely want to permute edges and do OLL you can do after HALL becaus you have only 3 OLL pattern to choose from. You will reduce the OLL to just 3 from 57. But if you take that path maybe you need to learn all 21 PLL instead of only 4.

I agree it is the edge rotation that can take slightly more time, but put this in the whole process it can be a good alternative. In the pdf document you have plenty of ready algoritms that turning edges while rotating the corners.

I agree that more cubers need to try this out and give it a chance.

With all new things it seams to be odd at first, but after a while it can be something great. this is a totaly new concept to solve last layer without prepairing for the LL.

Thank you for trying HALL

 

[Tjacke]

I am not the one who can make this approach great, it is all of you enthusiast cubers. I have just found a new way that has potential, but this approach needs your intrest and your skills and better set of algorithms.

The idea is to orient the corners and turn them right with only 2 algorithms. That works everytime nomatter how the LL look like after F2L.
You apply a trigger algorithm right after F2L and you either have an X or no correct color on the edges. That means you only have 3 OLL pattern left if X and slightly more 12 OLL if no color.

The trigger works good with one Sune to hide right color on the corners, but you might need 2 Sune sometimes to show all colors.

You don't need to prepare for the top layer in F2L which mean you can focus on doing the F2L as fast and efficient as possible. For beginners this might be a great approach becaus you can be fast and solve the top layer easy with few algorithms. Not much can go wrong and you don´t need many PLL.

The process
If you are a skilled cuber, this approach can still be of interest.
When you are done with F2L you either recognize a good pattern that give aPLL skip. If not a good pattern, you apply the trigger (by applying the trigger you can solve the cube). You will now get a new patter that can give a PLL skip. If not, you do a Sune and turn all corners up or all away so they eithe match the center or not. If you like that new pattern you solve the cube with a PLL skip.

After you done the second step, the Sune Phase you have 2 choices.
Either you apply one of the 3 OLL (only 3 possible OLL if you have the corners in place and turned right way) and finnishing with an PLL.

The outcome after step 2 when you turned the corners can be great because all edges can be permuted by it self and you can finnishing by doing one of the 4 PLL

Or all edges are in their position and you only have to rotate them.

This is a great compliment for beginners and semi-professional cubers.

This method have potential. Maybe some one see an approach or a change of trigger etc that can make this method more interesting.

This is an approach for the creative one that want to laborate and find their own way and algorithms that can boost luck or skips.

Look at the pdf, i have tried to explain the steps as easy as possible, and you also get some more or less usefull bonus algorithms that works with this method.

Thank you...

 

[xyzzy]

I didn´t want to learn all of that so i have spend hours on YouTube to try find another way of solve the LL without so many algoritms. I did not find that so i started to laborate to find a way and i did.

It can be an alternative for lazy cubers

There are two conflicting goals here: do you want to keep the number of algs low, or do you want the method to be fast? You cannot simultaneously attain these goals, unless your method is of the "subgroup descent" type, where the cube state moves along a descending chain of subgroups generated by single moves until it's eventually solved, e.g. the Roux method if you ignore the CMLL step.

Last layer methods that solve corner permutation (CP) before corner orientation (CO), including your method and Lars Petrus's beginner method, suffer from the problem that you either need to know CLL recognition (i.e. you also recognise the CO case at the same time) or check many more stickers than needed to recognise CO alone. In a speedsolving setting, that rules out the latter option. But if you're going to spend the time to do CO+CP recognition instead of only CO recognition, you might as well use a method that exploits that (e.g. CLL or OLLCP), which would then bring you to learning a whole bunch of algs to do that.

Your method may have more "skips", but one thing to keep in mind is that skips mean less when you're… skipping less. A CP skip happens 1/6 of the time, but you still have three-ish more last layer substeps to go through—your originally-4-step last layer became a 3-step last layer on this one solve, but it's still worse (in terms of move count and speed) than the true 2-step last layer methods like OLL+PLL or CLL+ELL. As another example, an EO skip (1/8 chance) is nearly worthless for a CFOP user who knows full OLL but doesn't use COLL/ZBLL.

So blah blah blah, your method isn't good for speedsolving. It does have a low alg count, which is nice, but that's not a unique feature of your method; there are also variants of OLL and PLL that require much less algs (but again, this comes at the cost of efficiency). For example, EO using fruruf (1 alg), CO using Sunes (1 alg), CP using T perms (1 alg) and EP using U perms (1 alg) makes for 4 algs in total, albeit with the last layer potentially taking up to 10 or 11 little steps.

 

[ProStar]

There are two conflicting goals here: do you want to keep the number of algs low, or do you want the method to be fast? You cannot simultaneously attain these goals, unless your method is of the "subgroup descent" type, where the cube state moves along a descending chain of subgroups generated by single moves until it's eventually solved, e.g. the Roux method if you ignore the CMLL step.

Last layer methods that solve corner permutation (CP) before corner orientation (CO), including your method and Lars Petrus's beginner method, suffer from the problem that you either need to know CLL recognition (i.e. you also recognise the CO case at the same time) or check many more stickers than needed to recognise CO alone. In a speedsolving setting, that rules out the latter option. But if you're going to spend the time to do CO+CP recognition instead of only CO recognition, you might as well use a method that exploits that (e.g. CLL or OLLCP), which would then bring you to learning a whole bunch of algs to do that.

Your method may have more "skips", but one thing to keep in mind is that skips mean less when you're… skipping less. A CP skip happens 1/6 of the time, but you still have three-ish more last layer substeps to go through—your originally-4-step last layer became a 3-step last layer on this one solve, but it's still worse (in terms of move count and speed) than the true 2-step last layer methods like OLL+PLL or CLL+ELL. As another example, an EO skip (1/8 chance) is nearly worthless for a CFOP user who knows full OLL but doesn't use COLL/ZBLL.

So blah blah blah, your method isn't good for speedsolving. It does have a low alg count, which is nice, but that's not a unique feature of your method; there are also variants of OLL and PLL that require much less algs (but again, this comes at the cost of efficiency). For example, EO using fruruf (1 alg), CO using Sunes (1 alg), CP using T perms (1 alg) and EP using U perms (1 alg) makes for 4 algs in total, albeit with the last layer potentially taking up to 10 or 11 little steps.

I agree. This is a low alg-count method, but it comes at the cost of bad for speedsolving.

Also I think the simplest LL method is EO with Fruruf, EP using anti-sune, CO using sexy and CP using a variant of sexy(just R U R' and R U' R'). Also since Fruruf is basically just a sexy and antisune can be explained as Take pair out(R U R') sexy, put pair back(R U' R'), it's basically one alg.

 

[ProStar]

Petrus-W for 4x4!

1. Solve 2 opposite centers
2. Solve edges for a 2x2x2
3. Solve centers while creating a 2x2x2
4. Solve a 2x2x3 in the back
5. Solve the remaining edges
6. Do EO as in Petrus-W. If there is an odd number of bad edges, solve all but one and then flip the remaining one with the OLL parity algorithm
7. Solve the rest of the cube as in Petrus-W. This can be done in a variety of ways, including L2P->COLL->L5EP, L2P-1->WV->HKPLL, L2P-1->CLP->L5EP, etc.

I'm gonna make an example solve

EDIT:

Scramble: D2 R2 B L2 F' L2 R2 B' R2 B U2 R2 D' F' R D B' F2 D' R' F Rw2 F' D2 Fw2 F' Uw2 L B' Rw2 B L2 B Rw2 Uw' B' U2 L2 Uw' D2 Rw B2 Rw' Uw2 B' U'

x2 // Inspection

U Rw2 U' Rw2 Lw' U' Lw z2 U' Rw U Rw' U' Lw U' Lw' // F2C

x y F U' Rw2 F' D2 U' L2 U' Rw2 L' U' B' Rw' B Rw // 2x2x2 Edges

L2 U 2L2 U' 2L2 x L2 F Lw' U2 2L U2 2L2 U2 2L2 z' y2 D L2 U2 F' U F y2 F' Uw F' Uw' // Centers + 2x2x2

R2 L' U L Uw' U2 F R' F' R Uw R Uw' R U R' Uw L F' L' U L' U L U' L' U' L // 2x2x3

y U' F R' F' R Uw' U' F R' F' R Uw R U2 R' Uw' R U R' F R' F' R Uw F Uw' R U R' F R' F' R Uw // Edges

R' F R // EO

U' m' U2 m U2 L' U' L U' R U' R' U R U R' // L2P

U2 R U R' U' R' F R2 U R' U' R U R' U' F' // COLL

m' U2 m U' m' U2 m // L5EP

Probably could've been way better; 158STM. For my Redux-Yau mix I normally do for 4x4 FMC, I get around 130-140STM(both assuming no parity). Some of the steps could probably be altered, like solving multiple edges before making 2x2x3 or something

 

[Tjacke]

There are two conflicting goals here: do you want to keep the number of algs low, or do you want the method to be fast? You cannot simultaneously attain these goals, unless your method is of the "subgroup descent" type, where the cube state moves along a descending chain of subgroups generated by single moves until it's eventually solved, e.g. the Roux method if you ignore the CMLL step.

Last layer methods that solve corner permutation (CP) before corner orientation (CO), including your method and Lars Petrus's beginner method, suffer from the problem that you either need to know CLL recognition (i.e. you also recognise the CO case at the same time) or check many more stickers than needed to recognise CO alone. In a speedsolving setting, that rules out the latter option. But if you're going to spend the time to do CO+CP recognition instead of only CO recognition, you might as well use a method that exploits that (e.g. CLL or OLLCP), which would then bring you to learning a whole bunch of algs to do that.

Your method may have more "skips", but one thing to keep in mind is that skips mean less when you're… skipping less. A CP skip happens 1/6 of the time, but you still have three-ish more last layer substeps to go through—your originally-4-step last layer became a 3-step last layer on this one solve, but it's still worse (in terms of move count and speed) than the true 2-step last layer methods like OLL+PLL or CLL+ELL. As another example, an EO skip (1/8 chance) is nearly worthless for a CFOP user who knows full OLL but doesn't use COLL/ZBLL.

So blah blah blah, your method isn't good for speedsolving. It does have a low alg count, which is nice, but that's not a unique feature of your method; there are also variants of OLL and PLL that require much less algs (but again, this comes at the cost of efficiency). For example, EO using fruruf (1 alg), CO using Sunes (1 alg), CP using T perms (1 alg) and EP using U perms (1 alg) makes for 4 algs in total, albeit with the last layer potentially taking up to 10 or 11 little steps.

I agree that the best is to know an algorithm for every outcome, but not everyone is there yet and many will never get there. I admire all who have learned 57 OLL and 21 PLL plus hundreds of more algorithms, amazing.

Today many records are very difficult to beat unless you have skips and luck. If you never solve the top layer in the way that can give you skips and luck after every set of algorithms you will have difficulties to beat a record, you will have hard time to get that jackpot because you don´t want to gamble.

This way might be slightly slower compared to someone who know full OLL and PLL, but this might also give a boost for those who don´t know full OLL and PLL.

Every methode to solve the LL have pros and cons. For HALL you need to locate the corners and do either single swap or double swap, thats it. It is pretty obvious after that what to do, you need to rotate the corners. You don´t need to learn a bunch of more algorithms, only learn more and more OLL till you feel comfortible.

HALL is not nessessary that bad for speed solving if you are on your way of learning all OLL and PLL.

This approach has few alg, the chance of making errors are few, you don´t need much planning except to know the corners after F2L. You have probabilities of skips and luck in every moment.
Question, can you solve the F2L faster and more efficently if only do that as easy as possible?
If yes, you save time. The LL can give you a nice setup without preparation.
If you don´t get that a good pattern, swap corners, when you do that you can solve the LL or get another OLL pattern that you can solve at once.
If you don´t get a good pattern now, use a Sune and you might solve the cube or get a new OLL pattern you can solve at once.

HALL is a gambling approach that is kind of error free and straight forward. I don´t think you will perform that bad with HALL if you don´t get a skip or luck with a pattern you can solve at once. But when you get a skips and luck with a pattern you can get a great result instead.

Maybe there are Cubers out there who like this kind of gambeling to reach better PB. It is up to everyone to find a trigger algorithm that suits different corner constalations so you by apply right trigger get all corner right with all colors right just by doing the trigger.

I think you need to gambeling to break new records today, you need to find ways that raise you luck to get skips and ways that reduces pattern you need for solving the cube at once.

Thank you for your professional view

 

[ProStar]

I agree that the best is to know an algorithm for every outcome, but not everyone is there yet and many will never get there. I admire all who have learned 57 OLL and 21 PLL plus hundreds of more algorithms, amazing.

Today many records are very difficult to beat unless you have skips and luck. If you never solve the top layer in the way that can give you skips and luck after every set of algorithms you will have difficulties to beat a record, you will have hard time to get that jackpot because you don´t want to gamble.

This way might be slightly slower compared to someone who know full OLL and PLL, but this might also give a boost for those who don´t know full OLL and PLL.

Every methode to solve the LL have pros and cons. For HALL you need to locate the corners and do either single swap or double swap, thats it. It is pretty obvious after that what to do, you need to rotate the corners. You don´t need to learn a bunch of more algorithms, only learn more and more OLL till you feel comfortible.

HALL is not nessessary that bad for speed solving if you are on your way of learning all OLL and PLL.

This approach has few alg, the chance of making errors are few, you don´t need much planning except to know the corners after F2L. You have probabilities of skips and luck in every moment.
Question, can you solve the F2L faster and more efficently if only do that as easy as possible?
If yes, you save time. The LL can give you a nice setup without preparation.
If you don´t get that a good pattern, swap corners, when you do that you can solve the LL or get another OLL pattern that you can solve at once.
If you don´t get a good pattern now, use a Sune and you might solve the cube or get a new OLL pattern you can solve at once.

HALL is a gambling approach that is kind of error free and straight forward. I don´t think you will perform that bad with HALL if you don´t get a skip or luck with a pattern you can solve at once. But when you get a skips and luck with a pattern you can get a great result instead.

Maybe there are Cubers out there who like this kind of gambeling to reach better PB. It is up to everyone to find a trigger algorithm that suits different corner constalations so you by apply right trigger get all corner right with all colors right just by doing the trigger.

I think you need to gambeling to break new records today, you need to find ways that raise you luck to get skips and ways that reduces pattern you need for solving the cube at once.

Thank you for your professional view

Learn 4lll, it doesn't have that many algorithms. And you can get skips. Also the best way to get skips is to learn extra algs:

wanna get OLL skips? WV!
wanna get PLL skips? COLL!
wanna get LL skips? Luck!

Knowing 4lll give way better solves on average, and a skip doesn't make it an amazing method. If I use LBL and get an LL skip, I'm gonna use more moves than a fullstep CFOP solve.

 

[Cubingcubecuber]

Hawaiian Kociemba variant:

Arrow(EO Arrow without EO, empty spot in DF)
F2L with M slice pairing so rotationless
HKOLL
HKPLL

I will call it Hawaiian CFOP

 

[PapaSmurf]

The problem with HKPLL is that the algs are almost certainly gonna be not that great. WLL (imagine TTLL with an edge instead of a corner) has awful algs. This is the same thing, but with the edge in DF instead of FR. I mean prove me wrong, but I wouldn't immediately say it's a fast idea.

 

[brododragon]

Easy 1LLL:
1. LB
2. Solve UF, U center, and UB on B layer. (Just solve UF and UB, then do a M2)
3. EO (Pretend like the B center is on U)
4. RB
5. Solve UF, UB, DF, and DB with an M2
6. COLL

I really like this idea, similar to LEOR but more flexible M slice.

 

[Skewbed]

Easy 1LLL:
1. LB
2. Solve UF, U center, and UB on B layer. (Just solve UF and UB, then do a M2)
3. EO (Pretend like the B center is on U)
4. RB
5. Solve UF, UB, DF, and DB with an M2
6. COLL

I really like this idea, similar to LEOR but more flexible M slice.

I don't think that COLL would always give you a 1LLL with this method, but if you solve UF and UB like you described, you would be able to use ZZLL for 1LLL.

 

[brododragon]

I don't think that COLL would always give you a 1LLL with this method, but if you solve UF and UB like you described, you would be able to use ZZLL for 1LLL.

In all of my testing UL and UR are automatically solved. Can you provide an example solve were ZZLL is needed?

 

[Etotheipi]

In all of my testing UL and UR are automatically solved. Can you provide an example solve were ZZLL is needed?

Try doing a T perm after solving UF and UB

 

[brododragon]

Nvm I found a case.

 

[Skewbed]

In all of my testing UL and UR are automatically solved. Can you provide an example solve were ZZLL is needed?

Nvm I found a case.

Reminds me of another post where someone thought they never got parity with their solving order for 3BLD.
If you can't figure out if it will be a skip or not intuitively, then please do some more testing before posting.