• 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!

ZZ-CT Thread

Joined
Apr 18, 2009
Messages
462
Location
San Diego, California
WCA
2007ESPI01
YouTube
Visit Channel
So I've been thinking about this alot and the more I think about it, the more this seems like another version of MGLS. Now, now, hear me out. I know it's annoying af when someone says stuff like this, as it's along the lines of "...or you could just use _____" and it doesn't seem like constructive feedback at all. But listen:

ZZCT is essentially like MGLS but instead of requiring specific placement of both the final edge AND final corner for execution of the LS, it only worries about the permutation of the final edge, and works the permutation of the corner into the algs of the final algorithm, resulting in a larger alg pool than standard CFOP PLL.

It's basically forcing an OLL skip using mainly R U triggers into a larger PLL alg pool.

Why does that sound familiar?

Oh yeah, MGLS. I was thinking if there were any other methods that force an OLL skip, and MGLS kept coming to mind, but I kept batting it away as too simplistic of a comparison to be taken seriously.

But no, by all measures that we gauge the greatness of ZZ-CT, MGLS could be seen as equally as awesome if not better.

1) 2-Gen-ness of the CLS/TSLE steps and cool triggers good for OH/2H
2) 104 algs in CLS compared to 94? in TSLE, so comparable amount of algs for this step.
3) 72 algs to learn how to recognize, recall and react (execute quickly) for TTLL, compared to 21 PLL algs we already know that have proven to be all sub-1-able by elite cubers.
4) Greater chance for a LL skip in MGLS (1/72 or 1.38%) than a TTLL skip in ZZ-CT (1/360 or 0.22%).
5) Similar move counts for CLS and TSLE, combined with greater chance for skip allows for theoretically just as great of a potential for insane 30~ish move solves and insane times when done using ZZ.
6) The difference is one super trivial step of ELS which adds maybe 2-3 moves (translates to 0.25~ish seconds) to your solve. Sometimes ELS is skipped as the edge is already placed (20% chance) in which case ZZ-MGLS is definitely the superlative. If I told you that you could save maybe 0.25 seconds off of your average by learning 72+94 more algs, would you bother if speed is your main goal? Maybe, but most would see this as not worth the time.

I was thinking about how you kept describing TTLL as a bunch of conjugated PLLs, and how you are even working to force PLL during some TSLE cases by paying attention to the position of the bad corner and maybe learning more algs that pay attention to this corner placement to increase an LL skip by forcing PLL. But if your goal is to eventually learn all the algs required to force a PLL skip more often (which I imagine would be pretty dang high) why not just learn CLS and do the 2-3 extra moves it takes to insert the edge?

The entire method definitely sounds like an exciting prospect, but only because of the "LL skip" framing. The real thing that is being skipped is the ELS equivalent of this method and that's because of ZZ orienting all the edges. So ELS is reduced to a simple AUF to position the edge in the UF position. Another way to look at it is ELS is being combined with CLS (without permutation) in TSLE. But it seems like a bunch of alg learning when you can trivially insert an edge and save so much time that would be spent learning algs.

I feel like in fact this is so close to reinventing an alg dense MGLS that it comes across in the additional techniques notes of the OP. In fact to show this let me hypothetically propose a way to expand on the forcing PLL during TSLE idea in a way that would logically follow this progress: you mentioned the algs that have symmetry being easiest to force PLL because you can easily position the corner where it needs to be to force a PLL. An aspecf of those symmetrical algs is also that the edge is already in place, right? So instead of learning a bunch of different TSLE algs (500 more?) to force PLL, how about we just insert the edge and figure out all the algs necessary to force PLL for all the TSLE cases that have the edge in place? How many algs would that be? I think it would reduce it from 500ish to 104? Definitely doable, all it takes is one extra trivial step and we can force PLL during TSLE all the time. Imagine the skip potential!

In summary, I am 100% on team ZZ-CT. I think it's a great new method that provides a unique way to approach the LS+LL. But after thinking about this a lot, it seems like MGLS would make more sense to learn as you don't have to learn to recognize recall and react to 94 new algs and you would have a greater chance of skipping the last step.
 
Last edited:

JTWong71

Member
Joined
Jan 24, 2016
Messages
251
Location
East Coast of the US
So I've been thinking about this alot and the more I think about it, the more this seems like another version of MGLS. Now, now, hear me out. I know it's annoying af when someone says stuff like this, as it's along the lines of "...or you could just use _____" and it doesn't seem like constructive feedback at all. But listen:

ZZCT is essentially like MGLS but instead of requiring specific placement of both the final edge AND final corner for execution of the LS, it only worries about the permutation of the final edge, and works the permutation of the corner into the algs of the final algorithm, resulting in a larger alg pool than standard CFOP PLL.

It's basically forcing an OLL skip using mainly R U triggers into a larger PLL alg pool.

Why does that sound familiar?

Oh yeah, MGLS. I was thinking if there were any other methods that force an OLL skip, and MGLS kept coming to mind, but I kept batting it away as too simplistic of a comparison to be taken seriously.

But no, by all measures that we gauge the greatness of ZZ-CT, MGLS could be seen as equally as awesome if not better.

1) 2-Gen-ness of the CLS/TSLE steps and cool triggers good for OH/2H
2) 104 algs in CLS compared to 94? in TSLE, so comparable amount of algs for this step.
3) 72 algs to learn how to recognize, recall and react (execute quickly) for TTLL, compared to 21 PLL algs we already know that have proven to be all sub-1-able by elite cubers.
4) Greater chance for a PLL skip in MGLS (1/217?) than a TTLL skip in ZZ-CT (1/360).
5) Similar move counts for CLS and TSLE, combine with greater chance for skip allows for theoretically just as great of a potential for insane 30~ish move solves and insane times when done using ZZ.
6) The difference is one super trivial step of ELS which adds maybe 2-3 moves (translates to 0.25~ish seconds) to your solve. Sometimes ELS is skipped as the edge is already placed (20% chance) in which case ZZ-MGLS is definitely the superlative. If I told you that you could save maybe 0.25 seconds off of your average by learning 72+94 more algs, would you bother if speed is your main goal? Maybe, but most would see this as not worth the time.

I was thinking about how you kept describing TTLL as a bunch of conjugated PLLs, and how you are even working to force PLL during some TSLE cases by paying attention to the position of the bad corner and maybe learning more algs that pay attention to this corner placement to increase an LL skip by forcing PLL. But if your goal is to eventually learn all the algs required to force a PLL skip more often (which I imagine would be pretty dang high) why not just learn CLS and do the 2-3 extra moves it takes to insert the edge?

The entire method definitely sounds like an exciting prospect, but only because of the "LL skip" framing. The real thing that is being skipped is the ELS equivalent of this method and that's because of ZZ orienting all the edges. So ELS is reduced to a simple AUF to position the edge in the UF position. Another way to look at it is ELS is being combined with (O)CLS in TSLE. But it seems like a bunch of alg learning when you can trivially insert an edge and save so much time that would be spent learning algs.

I feel like in fact this is so close to reinventing an alg dense MGLS that it comes across in the additional techniques notes of the OP. In fact to show this let me hypothetically propose a way to expand on the forcing PLL during TSLE idea in a way that would logically follow this progress: you mentioned the algs that have symmetry being easiest to force PLL because you can easily position the corner where it needs to be to force a PLL. An aspecf of those symmetrical algs is also that the edge is already in place, right? So instead of learning a bunch of different TSLE algs (500 more?) to force PLL, how about we just insert the edge and figure out all the algs necessary to force PLL for all the TSLE cases that have the edge in place? How many algs would that be? I think it would reduce it from 500ish to 104? Definitely doable, all it takes is one extra trivial step and we can force PLL during TSLE all the time. Imagine the skip potential!

In summary, I am 100% on team ZZ-CT. I think it's a great new method that provides a unique way to approach the LS+LL. But after thinking about this a lot, it seems like MGLS would make more sense to learn as you don't have to learn to recognize recall and react to 94 new algs and you would have a greater chance of skipping the last step.
That is a pretty name at perspective you have. For the TSLE that forces PLL's every time, there is already a subset for that, ZZ-C, where they explain it as Winter Variation from any F2L case.

I was thinking that it is best to use ZZ-CT when you get a 1 or 2 Trigger TSLE Case, so the Movecount is worth doing so.

Same thing with MGLS, sometimes the CLS cases aren't very good, so you might use some other Subset to solve the rest of the cube.

So, it really comes down to things that many other people say... Don't limit yourself to one Subset, try learning some of all of them so then you have multiple choices to choose from.
 

sub20oneday

Member
Joined
Dec 18, 2015
Messages
60
Location
uk
ive strted learning these algs but i dont understand when people are saying you can adjust he d face i understand this works with ttll but how do i do this for tsle or do i have to learn 4x the algs
 

AlphaSheep

Member
Joined
Nov 11, 2014
Messages
1,083
Location
Gauteng, South Africa
WCA
2014GRAY03
ive strted learning these algs but i dont understand when people are saying you can adjust he d face i understand this works with ttll but how do i do this for tsle or do i have to learn 4x the algs
It only works with TTLL. With TSLE you have two choices. Either you learn to always leave the last slot open, or you can learn to mirror the algs just like you mirror F2L algs.
 

4Chan

Premium Member
Joined
Jun 21, 2008
Messages
2,984
Location
Lumbridge
YouTube
Visit Channel
I don't think this is similar to MGLS for a few reasons:

1. MGLS is three look, even with skips, provides less benefit than ZZCT
2. CLS subsets have some rough cases, whereas TSLE is 100% 2-gen.

I feel that at the highest levels, the extra looks are what sets the two methods apart.
 

gyroninja

Member
Joined
Nov 19, 2015
Messages
139
1) Chris! Thanks for doing that seminar and thanks to the person who posted it, I unfortunately had to leave the venue early that day so I missed this seminar. I'll have to watch it later.

2) I've developed a beginners method for people who aren't beast alg learners like you. It involves just splitting the TTLL algs into COLL equivalents. So since this method is like a ZB substitute (like Splenda or, let's say Stevia), it makes sense for me to learn to recognize the COLL case first to train my eyes to TTLL more efficiently. That's how I recog a lot of ZBLL cases, in fortunately learned COLL first and it made a great stepping stone to learn some ZB algs. So there can be an intermediate step to TTLL where you just do an easy TTLL case to force an edge PLL. This can be referred to as COTTL (Corners Orient Tran Thomas Last Layer) and can be reduced to the 6 cases that you have to learn to recognize anyway. This will only add 1 to 2 seconds to your time and allow you to get a gauge on what to expect as you start learning to recog and execute the specific TTLL cases. To someone liker Mr. Tran who can learn this method in a day with a gun to his head, this may seem ridiculous or unnecessary. But to the masses that want to make this more doable, this could be a great stepping stone. I'll post my favorite TTLL algs for each case as a suggestion for COTTL algs (again, I'm learning for OH so these will have an OH bias in terms of executablity, except both opp).

Front Bar: y' *G PERM* (R U' R y R2 u R' U R U' R u' R2 F2)
Right Bar: U F2 *G PERM* (R2 u R' U R U' R u' R2 y L' U L')
All Bars: R *J2 PERM* U L' U2 R U' R' U2 L R U' R2
Both Opposite: U2 x (R' U R U')X4 x'
Front Opposite (2-gen): R2 U2 R2 U' R2 U' R2
Right Opposite (2-gen): y' R2 U2 R2 U R2 U R2

This makes it REALLY REALLY beginners friendly in terms of algs as it relies on PLL's most cubers already know.

Again, this is not meant to be a suggestion for a better method, in fact you will be slower with this obviously, but with every new method or technique learned (like F2L) you should expect to be slower before becoming faster.

Now that I'm done with the embarrassment that was US Nats 2016, I think I'm going to dedicate time to learning this method in full, because why the heck not?
The algs I recommend for this are

Front Bar: R U R' U' R U' R' U' R U R' U' R' D' R U R' D R
Right Bar: [R2 U: A perm]
All Bar: [R2 U: T perm]
Both Opposite: U2 l' U R U' (R' U R U')2
Front Opposite: U R U' R' U R U2 R' U' R U R'
Right Opposite: U' D R2 U2 R2 U R2 U R2
 

Teoidus

Member
Joined
Feb 11, 2016
Messages
573
Location
Char
No? CFOP average movecount is 55 with a rather huge standard deviation. Literally just open up cubesolv.es and look at the first 20 solves:

56 44 52 39 53 57 47 64 68 59 47 38 61 64 61 72 52 47 73 55 STM

Mean: 55.45 STM
STD: 10.03
 

Sajwo

Member
Joined
Sep 25, 2011
Messages
1,002
Location
Poland
WCA
2012SZEW01
YouTube
Visit Channel
No? CFOP average movecount is 55 with a rather huge standard deviation. Literally just open up cubesolv.es and look at the first 20 solves:

56 44 52 39 53 57 47 64 68 59 47 38 61 64 61 72 52 47 73 55 STM

Mean: 55.45 STM
STD: 10.03

Yes, but
> Higher chance of skips
> ZZ-CT is heavy based on intuition and block-building -> lower tps
> Easier recognition of LL in CFOP
> Hundreds of possible LL and LS extensions -> lower movecount
> Easier
> And faster
 

JTWong71

Member
Joined
Jan 24, 2016
Messages
251
Location
East Coast of the US
Yes, but
> Higher chance of skips
> ZZ-CT is heavy based on intuition and block-building -> lower tps
> Easier recognition of LL in CFOP
> Hundreds of possible LL and LS extensions -> lower movecount
> Easier
> And faster
What do you mean 'Higher chance of skips', and Last Layer Recognition?

For ZZ Last Layer if you use OCLL/PLL, you don't have to worry about Edge Orientation, so there are only 7 cases instead of 57 to recognize.

Along with the Last Slot/Last Layer Extensions, what most of them do is they skip OLL by decreasing the overall movecount by 1-3 moves. ZZ also has lots of Extensions, Winter Variation is a common one used in both ZZ and CFOP, only thing is, it requires presolved EO.

Even though ZZ uses lots of Blockbuilding, it is mostly 3-Gen, so almost no F/B moves are going to be used in the solve excluding the first step. It just takes practice to get everything by muscle memory and quick recognition.

And... The two last things you said are mainly just bias...
But I do agree with you saying CFOP is easier. CFOP just requires memorization, so everything is automatic.

The reason I like ZZ is because you don't have to regrip/rotate nearly as much as you would in CFOP.
 

Teoidus

Member
Joined
Feb 11, 2016
Messages
573
Location
Char
Yes, but
> Higher chance of skips
This is literally false. CT is based on maximizing skip chances.

> ZZ-CT is heavy based on intuition and block-building -> lower tps
Er, not quite. ZZF2L usually ends up being CFOP recog + learning to build pairs with "cross pieces." The only methods I can think of that are "heavily based on intuition and blockbuilding" are Petrus and Heise. Even roux SB becomes pretty mechanical after a while, and that's far more complex than ZZF2L.

> Easier recognition of LL in CFOP
I don't see where you're getting this. Recog for ZZ-CT seems identical to OLL-PLL. Look at shapes, then look at blocks/color patterns.

> Hundreds of possible LL and LS extensions -> lower movecount
Also false? ZZ has the most LL+LS extensions of any method.

> Easier
> And faster
Bias ftw.

Look man if you don't have anything to contribute to the method you're kind of wasting your time here.
 
Joined
Apr 18, 2009
Messages
462
Location
San Diego, California
WCA
2007ESPI01
YouTube
Visit Channel
I don't think this is similar to MGLS for a few reasons:

1. MGLS is three look, even with skips, provides less benefit than ZZCT
2. CLS subsets have some rough cases, whereas TSLE is 100% 2-gen.

I feel that at the highest levels, the extra looks are what sets the two methods apart.
Yeah but what is that extra look? An R U' R'? Or maybe a U' R' in some cases. It could be considered an extension of the F2L-1 portion it's so trivial. If the lure of this method is the skip probability and the chances for great single times, then it seems like MGLS would give ZZ-CT a run for its money, at least on paper, with a PLL skip probability having a higher chance than a TTLL skip, am I right?

2-gen is awesome but it might be as overrated as low move count as clearly we could get 10 second averages OH using non-2-gen PLL.

All this being said, I got my first low-16 second time using ZZ-CT for OH today. As I continue learning this, I really feel like the recognition for TSLE can become as intuitive and as quick as regular F2L.

visualcube.php

Found a new one for this:
Old alg:
R U' R' U' R U R' U2 R U R'
My alg:
U2 R U2 R' U2 R U R'
 

4Chan

Premium Member
Joined
Jun 21, 2008
Messages
2,984
Location
Lumbridge
YouTube
Visit Channel
How does this method compare to ZZ-b?

ZZ-b requires you to learn lots of new skills, such as ZZLL recognition, memory tricks, and the ability to phase quickly. ZZ-b also requires one more look than ZZCT, and I think it's best as a stopgap to ZZ-A



@PhillipEspinoza : You do have a point there. The skip percentage is alluring indeed.
I think maybe it just needs someone to take it to it's fullest potential.
 
Joined
Apr 18, 2009
Messages
462
Location
San Diego, California
WCA
2007ESPI01
YouTube
Visit Channel
I think if you look at ZZ-CT as an alternative to ZZ-b or ZZ-a etc, it's so much better as it is the equivalent of forcing one specific cross case but instead of phasing for it you do maybe a less intuitive but arguably more speed adaptive LS step because of the recognizability of TSLE. It's just my opinion but yeah. I also think TTLL is very learnable.

Does anyone know the possibility of a TSLE skip? I had one recently and I think anecdotally it might be less likely than a LL skip.

Also, would it be worth it to investigate a CFOP-friendly version of CT? I guess with ELS...? Or maybe another way to make it more CFOP-friendly?
 
Top