Petrus + MGLS = EJF2L

[Michelle137137K]

What cube are you using?

[CaptianCrash44]

@Michelle137137K I think I was using a type C in this video. I use a zanchi now.

[felixduhcat89]

What is the algorithm for that J perm at 2:57 ??

[CaptianCrash44]

x' R2 u' R' U R2 x y' R' U R' U' R2

[ottozing]

@CaptianCrash44 alg dosent work :/

[CaptianCrash44]

@ottozing x' R2 u' R' u R2 x y' R' U R' U' R2

[1234fivesixseven8]

Dude would be awesome if you could do CP whilst doing CO with this method but there would be a lot of cases

[Hyrtsi]

Yeah and could you explain the part at 0:35 -> again please?

[CaptianCrash44]

@Hyrtsi I was pretty unclear. hahaha. I was just explaining how to use MGLS the normal way. once you get to the last slot you do ELS: orienting all of the edges and placing the last F2L edge. CLS: placing the last F2L corner while orienting all of the last layer corners.

[Hyrtsi]

@CaptianCrash44 Yup thanks!

[Hyrtsi]

I'm using Petrus F2L + Fridrich OLL&PLL. My average is like 45 seconds. Does this make my F2L any faster or should I begin studying Fridrich?

[felixduhcat89]

@Hyrtsi Have you tried it?

[Hyrtsi]

@felixduhcat89 Yeah! This method is awesome. I've been doing it like 6 months now ;)

[blackzabbathfan]

this is a good idea and i might use it because i'm afraid to learn fridrich last layer because it sounds easier that fridrich last layer but really isn't. 78 fridrich LL against 103 CLS algs but i still might just learn CLS

[mynameisdarthtater]

The more I think about it the more it seems useful to learn COLL. It's good for MGLS, ZBLL, roux, 2x2, and forcing easy PLLs with friderich. Like I need to know MORE algs :-/

[TheChaosFenix]

@mynameisdarthtater yeah, i think that way 2, im gonna learn coll cause im gonna learn roux, and because the cases arent much slower than the OLL, so i can use them in fridrich to force E-Perms, I currently only use the Bruno and Double sune cases, but they sure do help a lot; if u end learning it, see if u can find good algs 4 the triple sune and superman case; i just cant seem to find them XP

[mynameisdarthtater]

@TheChaosFenix I know the double sune and superman cases, although I almost always switch two for the superman. Anyways, I've taken a COLL breather, because I've just gotten CN and I want to be sub20 for the competition in Washington DC

[alghazali90]

you do like this click click click clcik and then you could do click click click click easy right? :D

[mrgiggums0]

@faroeq33 yeah but then you might as well do R U R' and use OCLL which has 7 cases not 27.

[faroeq33]

and it has lesser mgls algs. and with F2LL it has only 27 algs

[faroeq33]

I think this EJf2L is very great but if you use petrus+f2LL( winter variation) its even better. Because you dont have to make a f2l pair with twisted corner

[poptheman1]

i lol'd at the tags. fridrichmethodisnotfun

[Tips4Cube]

can u make a vid of your plls because good cubers=good plls. because my plls suck

[bobthegiraffemonkey]

I'm experimenting with various methods just now and I think that you're a genius and an inspiration for more ideas. My latest idea was to use the first step of the human thistlethwaite to flip edges at the start. I really like your idea of leaving any corner unflipped and only learning the MGLS cases with the corner placed.

Thank you!

PS Wow you're fast!

[SYKWYDIT]

i dont no if i would use this just a personal thing for me. I can see my self using zbf2l wich i think is very use full but nice vid and the 2 gen is a good idea

[CaptianCrash44]

I personally can't see anyone seriously using ZBF2L for speedsolving. There are over 100 cases you have to learn just to have all edges oriented when you get to the last layer. It's not worth it.

[SYKWYDIT]

i might just learn a couple cases because knowing full oll and pll well get repetative and im going to want to learn a little more advanced stuff but i agree it dont really seem to have a big benifit

[grolich]

Actually ZBF2L may actually cause the opposite of having a benefit - since the number of cases is in the hundreds (a lot more than a 100), while you theoretically gain in number of turns(not that much, btw), the extremely high number of cases means recognition takes longer than with other systems.

[RookieN08]

i agree with you Ej :)

ZB method have more than 500cases but MgLS is a lot better

[IPwnAllAndLeaveNone]

@SYKWYDIT

F2L, OLL, PLL, Edge Control, And COLL is very advanced, but a goal for me atleast. If you have the time to look into them, then i suggest you use them.

[Hyzing]

what cube do u use

[CaptianCrash44]

type c

[dudemanpp]

This is very genius. Good job!

[CaptianCrash44]

if you fix bad edges before you get the the LL then it would be faster. I try to use this for every OH solve I do because the 2gen OLL is so fast.

[RookieN08]

you should make more speedsolve videos and give some tips while solving the cube with petrus :)

[comedyishere]

What is the algorithm for the J:a perm at around 2:56?

[CaptianCrash44]

x R2 u R u R2 x y R U R U R2

[comedyishere]

Uhh, I just tried it, I think it's wrong lol.

[CaptianCrash44]

ha ha ha. for some reason when I pasted it none of the ' showed up. x' R2 u' R' u R2 x' y' R' U R' U' R2. try that

[RookieN08]

EJF2L stands for Erik Johnson's First-two layer?

[CaptianCrash44]

indeed it does

[RookieN08]

oh ok lol :) what do you average now? 12 or 13sec?

[CaptianCrash44]

13s now

[RookieN08]

btw what diy are you using? type a? it looks so smooth :) averagin 13sec?0_0 wow I have never seen a petrus user averagin 13-15sec

[CaptianCrash44]

the cube it the video is a type a. I think the cube I use now is a type c but I'm not sure.

[RookieN08]

thanks

[boiswatching]

Great idea Erik, you will get down to 11 second averages when you have memorized all the cases :O

[bamilan]

I have already known the method you described(the original one).

The idea at the 2nd pair is good(less number of moves). But why do you orient all edges? What a bad finish it is? :O????

I think it is better not to orient edges, just solve the remaining part of F2L, and learn the cases for corners(altogether easier, than learn CLL cases), and then finish with ELL. (so no orientation, and ELL instead of PLL => less number of moves, less time.

Milan

[CaptianCrash44]

I use petrus method so I always orient edges. And 2 gen F2L finish is very nice. If you are talking about ELL where edges are not oriented then you are killed by recognition time.

[bamilan]

If you don't orient, you save some seconds(recognition+solving), and recognize ELL is not that slow(faster than recognition+orient edges), so this way you can save some time :)

[CaptianCrash44]

Not really. I'm doing F2L(edges oriented), CLS, PLL. I think your doing F2L(no edge orientation), CLS, CLL, ELL. your doing an extra step on the LL while I do an extra step with F2L. I'm pretty sure recognition + execution for bad edges is much faster than ANY set of LL algorithms. I'm <1.25 seconds for bad edges (recognition included). And flipping edges takes an AVERAGE of 5.1 moves so it's more efficient than ELL.

[bamilan]

CLS? What is that?

CLL for 5 corners, but 1 of them is placed incorrectly(if correctly, then a singly CLL) =>not many algs.

So I save 1 step, but this method requires more algs to learn.

[CaptianCrash44]

I don't think you know what you are talking about. CLS is the second step of MGLS. It orients the twisted corner in the F2L and all of the corners in the LL, saves EO and ignores CP. CLL is solving orientation and permutation the 4 corners in the LL while ignoring EO. COLL is the same as CLL but it preserves EO. Rowe had the same idea as you when I told him what I had come up with. he used the term "butt load of cases", 1440 I think. Recognition for 1440 cases would kill you.

[bamilan]

I know what I am talking about. I told you a better idea, using the shortcut you showed me in the video. 2x3x3 block(1 oriented or unoriented corner), then last part of F2L, and then CLL(if corner is oriented it is CLL, if not, it is something like CLL, but orients the unoriented placed corner), and then ELL.

CLL(here) is 3*CLL algs. But I(maybe you too) already know the normal CLLs, you have to learn CLLs, which orient 1 corner in the bottom layer.

I hope you can understand what I try to say.

[CaptianCrash44]

ok I'm gonna make this quick. CLL that orients the twisted corner has 112 including mirrors, this is not including regular CLL. I'm not gonna learn that many cases for one step. And ELL algorithms are terrible. just reading the algorithms make my hands hurt. look them up and you'll see.

[bamilan]

I know ELL and CLL, and CLL+ELL is as fast as OLL+PLL. Recognition time is lower for ELL.

If you learn this method, you can have really good times in 2x2 :)

[boiswatching]

Uh, I think you're sayin' not to orient the edges right? well, uh, when you do that you can simply do a 2-gen F2L, which is pretty quick (correct me if I'm misunderstanding this?)

[xLordOblivionx]

Wow, this is really great!

[changjessica12]

:0 woah! thats so awesome!

[gamemeister27]

Wow, that is genius! It solidifies my transistion to petrus.

[KConny]

Wow, this was really interesting, nice one.

[Genie1048]

what kind of cube is that?

[CaptianCrash44]

type A

[gyc6001]

nice method. but its quite complicated.

ur fingers is like machine gun. lol.

[Genie1048]

lol then what's nakaji's fingers like lol :)

[CaptianCrash44]

several machine guns

[esquimalt1]

you know EVERYthing about the petrus method lol

[lee0sensei]

But, what if you do F2L and not Petrus .

you could do this on F2L but it seem not very comfortable at speedsolving.

Any tips?

Oh, and this is amazing but I don't use Petrus :(

[CaptianCrash44]

practice.

[lee0sensei]

I know that -.-

I know if there is any tips another than practise if you do F2L

[fallofshadows]

practice.

[ScottKidder11]

If you use Fridirch F2L, it's just MGLS on Lucas's page which places the edge and orients edges, and then place the corner/flip the corner and orient all LL corners. Awesome amazing idea Crash

[somerandomkidmike]

Great idea. Thanks for the video.

[lemonycheese]

Would you recommend this over learning COLL?

[CaptianCrash44]

yes