Phasing to always solve all 4 corners (ZZ)

[Adrian Currier]

So, I proposed a ZZ variation in 2014 (ZZ-reduction) to minimize the number of algs needed for last layer. It only requires 16 algs for OLL and PLL, the lowest for any 2LLL to my knowledge. I was intimidated by the idea of learning 78 algs for full CFOP. ZZ-reduction relies on phasing while placing the last block pair of F2L to place opposite edge colors in place (green/blue and orange/red), though they may be swapped (e.g. green >orange > blue > red rather than green > red > blue > orange). Then, OLLs that preserve 2 opposite or all 4 edges are used to reduce PLL cases.

However, I'd like to reduce PLL cases even further by phasing in a way that doesn't ever result in parity, where the opposite edges may be swapped, but instead always in solved position. This should limit PLL to only A, E, and H cases, thus reducing algs needed and speeding up recognition even more. Can anyone help me with figuring out how to phase so that all 4 edges are always solved when finishing F2L in ZZ? Thanks ahead of time

P.S. At this stage, simple is good for me. Even if phasing to permute all edges every time requires finishing F2L and then moving block pairs around to shuffle edges into place rather than finishing F2L in a way that results in edge permutation (which might be more complicated), I'm okay with that for now!

 

[Deleted member 36445]

I'm gonna follow this thread, I love concepts like this!

 

[PapaSmurf]

This (unsurprisingly) already exists. It has been proposed with various names - E, F, superphasing and other things I can't remember. The way you go about it is slow, especially intuitively. If you don't care about speed, you can probably figure it out/gen the algs (should be two more).

 

[ProStar]

I guess this works if your priority isn’t to get faster. It’s almost never advisable to try to find some workaround to learning OLL and PLL, becuase it always just ends up being slow and building bad habits in the cuber. I’d personally recommend starting with 3LLL, since that only has 13ish algs, and then slowly learning PLL from there

Although if you’re scared of learning algs, I’d recommend you switch away from ZZ; it’s a very alg-heavy method

 

[OreKehStrah]

(U) R U' R' - UL UF are correct relative to each other, FR is op to UF

(U2) R U2 R' - UL UF are adj incorrect order, FR op UL

f' R U R' U R U2 R' U f or (U) R U' R2' U2 R U R' U R - All 3 U edges are permuted

(U2) R U' R' U R U2 R' - UL UF are op , FR color comes goes CCW from UF

(U) L' U2 R U R' U2 L - UL UF Op, FR goes CW from UF

R' F R2 U' R' U' R U R' F' - UL belongs in UR, FR op UF

those are the algs I genned for the built pair subset a while back.

 

[ruffleduck]

You may be interested with one of my concepts, Zeroing LSLL (BT+): How to skip LL every time

 

[DuckubingCuber347]

You may be interested with one of my concepts, Zeroing LSLL (BT+): How to skip LL every time

It's literally just a consistent way to do zeroing. I don't know why you're claiming that's it's an original idea. You were, to my knowledge, the first to document but Feliks Zemdegs had already been using it in his solves for years.

 

[Jorian Meeuse]

It's literally just a consistent way to do zeroing. I don't know why you're claiming that's it's an original idea. You were, to my knowledge, the first to document but Feliks Zemdegs had already been using it in his solves for years.

Zeroing is more of a meme than an actual technique

 

[ruffleduck]

It's literally just a consistent way to do zeroing.

That is correct.

I don't know why you're claiming that's it's an original idea. You were, to my knowledge, the first to document but Feliks Zemdegs had already been using it in his solves for years.

My system for forcing it is completely novel and unique. Feliks never had developed any system for forcing zeroing, nor has anyone else. Sure, he does zeroing when it's painfully obvious to set up without the necessity of any system, but that's like saying a ZZ-A user is a ZZ-C user when they get an OCLL skip. Whoops! Looks like I did OLS! Definitely intended that!

 

[Deleted member 36445]

ZEROING. IS. A. MEME.

 

[Adrian Currier]

I'm gonna follow this thread, I love concepts like this!

Have you used ZZ much? You're welcome to check out my attempt at explaining the ZZ-r method on the wiki: https://www.speedsolving.com/wiki/index.php/ZZ-reduction

This (unsurprisingly) already exists. It has been proposed with various names - E, F, superphasing and other things I can't remember. The way you go about it is slow, especially intuitively. If you don't care about speed, you can probably figure it out/gen the algs (should be two more).

I hadn't seen this variant when I looked through the ZZ variants on the speedsolving wiki back in 2014. From what I can tell, ZZ-e and ZZ-EF are not the same as what I've proposed in ZZ-r. I'd love to know more about superphasing if you can point me to some info, At this point, I'm more interested in being able to solve every case without having to memorize so many algs (as someone who learned how to do CFOP starting out). I tried some other variants (like Roux) before proposing ZZ-r.

ProStar:

I guess this works if your priority isn’t to get faster. It’s almost never advisable to try to find some workaround to learning OLL and PLL, becuase it always just ends up being slow and building bad habits in the cuber. I’d personally recommend starting with 3LLL, since that only has 13ish algs, and then slowly learning PLL from there

Although if you’re scared of learning algs, I’d recommend you switch away from ZZ; it’s a very alg-heavy method

I appreciate your suggestion, I'll look into 3LLL. As I mentioned in my post, the ZZ-r variant I proposed only has 16 cases to memorize algs for. That's why I proposed it. I initially learned to solve using CFOP (at least 2LLL) and looked into Roux (and maybe Petrus, can't remember). I don't know that I'll ever commit to memorizing all 78 algs in full CFOP. But I can appreciate those who do, it certainly seems to be the fastest method

EDIT: After looking at the 2-look OLL, looks like there's about 30 algs (though many of the OLL algs are small). BLL is another 3LLL and it's about that many algs, too.

OreKehStrah said:
(U) R U' R' - UL UF are correct relative to each other, FR is op to UF

(U2) R U2 R' - UL UF are adj incorrect order, FR op UL

f' R U R' U R U2 R' U f or (U) R U' R2' U2 R U R' U R - All 3 U edges are permuted

(U2) R U' R' U R U2 R' - UL UF are op , FR color comes goes CCW from UF

(U) L' U2 R U R' U2 L - UL UF Op, FR goes CW from UF

R' F R2 U' R' U' R U R' F' - UL belongs in UR, FR op UF

those are the algs I genned for the built pair subset a while back.

Wow, thank you! I'll definitely give those a whirl

EDIT: So, double checking that UL = upper left edge, UF = upper front. What is FR? And these algs are employed once F2L if finished or before final block pair is inserted or something else?

ruffleduck:

You may be interested with one of my concepts, Zeroing LSLL (BT+): How to skip LL every time

I'll check it out!

 

[PapaSmurf]

I hadn't seen this variant when I looked through the ZZ variants on the speedsolving wiki back in 2014. From what I can tell, ZZ-e and ZZ-EF are not the same as what I've proposed in ZZ-r. I'd love to know more about superphasing if you can point me to some info, At this point, I'm more interested in being able to solve every case without having to memorize so many algs (as someone who learned how to do CFOP starting out). I tried some other variants (like Roux) before proposing ZZ-r.

That's because it's just not a good variant. ZZ-E and F are all common names given to many (bad) variants, so it's more that it's not surprising that I've heard them be called those names. The Wiki is (unfortunately) neither exhaustive, useful nor accurate when it comes to many things cubing (especially ZZ), so don't trust it. Superphasing is just another name for what you've described and tbh, I might be misremembering the name. Also, what OreKehStrah put are the algs for it and will be the best shot you have at it.

 

[Deleted member 36445]

Have you used ZZ much? You're welcome to check out my attempt at explaining the ZZ-r method on the wiki: https://www.speedsolving.com/wiki/index.php/ZZ-reduction

No but I'm aware of ZZ-r. I find it brilliant!
I have a thread about it

https://www.speedsolving.com/threads/zz-r-ol4c-incorrect-algs-on-wiki.74566/#post-1323820

 

[ProStar]

I appreciate your suggestion, I'll look into 3LLL. As I mentioned in my post, the ZZ-r variant I proposed only has 16 cases to memorize algs for. That's why I proposed it. I initially learned to solve using CFOP (at least 2LLL) and looked into Roux (and maybe Petrus, can't remember). I don't know that I'll ever commit to memorizing all 78 algs in full CFOP. But I can appreciate those who do, it certainly seems to be the fastest method

EDIT: After looking at the 2-look OLL, looks like there's about 30 algs (though many of the OLL algs are small). BLL is another 3LLL and it's about that many algs, too.

ZZ 3LLL only requires 13 algs I believe. 7 for CO, 2 for CP, and 4 for EP. 2LLL only requires 28, with 7 for CO and 21 for PLL.

ZZ, generally, gets significantly better the more algs you throw at it. It’s biggest advantage is that you have EO during LL (RUL F2L isn’t that much better, if at all better, than efficient RULFy), and it requires algs to take advantage of that and make up for having to do EO at the beginning of the solve.

If you’re alg averse but still want decent times, learn either CFOP with 4LLL (15ish algs) or learn ZZ with 3LLL (13ish algs). You can easily be sub-20 with either. Alternatively you could try Roux,
which has very few algs and is still very good. I think you only need 9 or so for beginner CMLL and maybe a couple more for the rest of the solve.

Your method is neither particularly new or good at all, and it doesn’t even require less algs than standard beginning speedsolving methods

 

[Adrian Currier]

No but I'm aware of ZZ-r. I find it brilliant!
I have a thread about it

https://www.speedsolving.com/threads/zz-r-ol4c-incorrect-algs-on-wiki.74566/#post-1323820

Oh, hey Clearly I have some algs to fix. I just ZZ-r seemed like a simple way to reduce cases and algs for ZZ. I had help with generating a lot of the algs, but I guess I put up some bad ones? I'm gonna go through and label the OLCs that permute all edges for what PapaSmurf is calling superphasing. This reduces PLL cases to A, E, and H. I've also included a link to a video demonstrating how to superphase in the external links for ZZ-r in the wiki:

. It's nothing super detailed, but better than nothing for the moment

ZZ 3LLL only requires 13 algs I believe. 7 for CO, 2 for CP, and 4 for EP. 2LLL only requires 28, with 7 for CO and 21 for PLL.

ZZ, generally, gets significantly better the more algs you throw at it. It’s biggest advantage is that you have EO during LL (RUL F2L isn’t that much better, if at all better, than efficient RULFy), and it requires algs to take advantage of that and make up for having to do EO at the beginning of the solve.

If you’re alg averse but still want decent times, learn either CFOP with 4LLL (15ish algs) or learn ZZ with 3LLL (13ish algs). You can easily be sub-20 with either. Alternatively you could try Roux,
which has very few algs and is still very good. I think you only need 9 or so for beginner CMLL and maybe a couple more for the rest of the solve.

Your method is neither particularly new or good at all, and it doesn’t even require less algs than standard beginning speedsolving methods

Oh, ofc you were referring to 3LLL ZZ, not CFOP. I was referencing algs needed for 3LLL CFOP (30ish). I'm interested in learning more about ZZ 3LLL and I can take a look at 4LLL CFOP. I've already looked into Roux in the past, before I looked into ZZ and then proposed ZZ-r. I've actually figured out how to do EP when finishing F2L in ZZ-r now (the reason I made this thread), which limits PLL cases to A, E, and H. This brings algs to 7 OLLs + 4 PLLLs = 11 algs. If other people have already thought of ZZ-r before I did in 2014, that's awesome, and not too surprising, given how simple the idea is. But I still think it was helpful to document the variant on the wiki since I couldn't find it there already.

 

[ProStar]

Oh, ofc you were referring to 3LLL ZZ, not CFOP. I was referencing algs needed for 3LLL CFOP (30ish). I'm interested in learning more about ZZ 3LLL and I can take a look at 4LLL CFOP. I've already looked into Roux in the past, before I looked into ZZ and then proposed ZZ-r. I've actually figured out how to do EP when finishing F2L in ZZ-r now (the reason I made this thread), which limits PLL cases to A, E, and H. This brings algs to 7 OLLs + 4 PLLLs = 11 algs. If other people have already thought of ZZ-r before I did in 2014, that's awesome, and not too surprising, given how simple the idea is. But I still think it was helpful to document the variant on the wiki since I couldn't find it there already.

All this does is make a bad version of beginner ZZ with the same amount of algs as a good version of beginner ZZ/CFOP

 

[Adrian Currier]

I've updated the ZZ-r wiki to reflect the information about superphasing (the reason I made this post to begin with). For those who care

I still think it's a worthwhile method. It adds an extra alg (the superphase) for edge permutation right after F2L, but in exchange it reduces PLL cases to just A, E, and H with the added benefit of increased PLL skip probability of 1/12. I think this is a nice low alg option because you're still using OLL algs for the 7 ZZ cases (except these ZZ-r algs also preserve EP, which many OLL algs already do anyway so it's not a big change for someone using this to transition into full ZZ or CFOP) and using PLL cases like you would in regular ZZ or CFOP (even if only 4 cases).

I'm definitely still interested in learning about 4LLL CFOP, BLL (I'm assuming this is the 3LLL ZZ method ProStar was referring to), and Zeroing LSLL that ruffleduck recommended. I also plan to take more time to study the algs you shared for superphasing, OreKehStrah. Oh, and I removed the inaccurate algs for ZZ-r, Filipe Teixeira. Thanks for helping with that and bringing it to my attention, too!

Thanks for the input everyone

 

[Deleted member 36445]

can you check OLC 2? (antisune)

first alg of "preserves parity only" doesn't preserve parity at all (if fact, it doesn't even solve the case with that auf)

 

[Deleted member 36445]

I've learnt the method (actually I've had learnt it in the past prior to the superphasing proposal) with exception of the Pi case which is pretty long alg

my thoughs: phasing is an excellent idea (in theory) to reduce the number of cases on the last layer but having a bad phasing case and having to do super phasing afterwards is pretty exaustive and a weak point of the method.

having only 3 (mirrors excluded) plls is pleasant but the previously mentioned hassle must be taken in account.

Pi case ocll is the worst of all. I suggest this alg instead:

 Twizzle link 

z D' R U R' D R U2 R D R' U R D'

I have a different proposal:

eo cross + f2l without caring with last layer edge permutation
solve epll with sunes (this link provides my method for solving the epll with sunes easily... 6 different cases solve with sunes)
l4c (84 algs)

90 "algs" total

BUT this method I'm proposing may be just worst than oll + pll or coll+epll after all

I talked about a beginner version but no one seemed to care back then

let's not mention zbll or zzll

 

[Adrian Currier]

can you check OLC 2? (antisune)

first alg of "preserves parity only" doesn't preserve parity at all (if fact, it doesn't even solve the case with that auf)

View attachment 21444

Oh snap, thank you for pointing that out. I've removed that alg. I hope it was one fluke in the OLC algs at this point, I tried to make sure everything was categorized and bad algs were removed