Very NOOB Idea I Came Up With

[Jilvin]

Okay, I was just doing some thinking. Feel free to RIP on this idea at any time. I know how dubious it sounds.

Here it is: You start the cube by completing the cross, then you do the first 2 pairs of f2l. On the third pair, you apply a ZBF2L type thing and do an orientation of all the edges. Then on the LAST pair you solve it simultaneously whilst orienting the corners of the last layer without disturbing the edges you have already completed.

I got a spark of this idea by a poster who had an idea which implemented the last slot of F2L and OLL at the same time, but this arises over 8000 algorithms

Splitting this into two steps and implementing them into 2 slots reduces the amount of cases and is would still be faster than conventional Fridrich.


Go ahead and tear my idea into pieces, but TRY to think of anything, ANYTHING that could be helpful!

 

[Lt-UnReaL]

I dunno how to calculate this, but I'm guessing that'd be around 500 algorithms. :/

 

[dRaGoN]

The thing is, when you have 500 algorithms, what's the chance of you recognizing and pulling it out of your brain within that one second that you see it? =/

 

[CraigBouchard]

You'd be surprised. Ask someone like Chris Hardwick. Or Doug Li. They both know a LOT of algorithms. Or Stefan Pochmann. Yet, they still manage to do it. But for the amount of time you'd save, not sure if it'd be worth it...

 

[Me]

You're not the first person to come up with this idea
*searches forum...*
ah-ha! i knew i read it somewhere.
http://www.speedsolving.com/showthread.php?t=479
the image the topic links is bad, but its pretty much what you're describing.
So i hope that answer suffices.

 

[joey]

This is basically very similar to MGLS. Except for each of your stages for the 1st and 2nd slot, is done with 1 slot.
http://cube.garron.us

 

[AvGalen]

Another "humanly impossible" idea:
Combine 4-look last layer with 4 slots during F2L
1st slot: EO
2nd slot: CO
3rd slot: CP
4th slot: EP
You will always get a last layer skip (unless you start with an X-Cross

Good luck finding the algorithms

 

[Dene]

Good luck finding the algorithms

LMAO. Yes indeed your "humanly possible" idea would be so slow, as after the cross, the 5 edges you are looking for could be anywhere around the cube, then you have to see which way around they are, and figure out which algorithm, out of the very many, is needed. I dunno, it sounds like fun though, someone with some uber program should do it and post the algorithms somewhere quickly, I wanna try it now:rolleyes:

 

[Speedy McFastfast]

I don't think that solving portions of the LL during the 3rd and 4th slots would help very much. There would be a seriously huge amount of algs for the third slot, seeing as you can have more positions than a conventional ZBF2L.

I also think that reco for orienting the corners would be a pain. It's easier just to do a ZBF2L during the last slot, and then orient the corners with 1/7 OLL algs. And furthermore, it's even easier just to do OLL

 

[Jilvin]

Well, another idea is to split up ZBF2L into two slots to make it easier.

Since there will always be at least 2 top layer edges on the top layer with 2 slots solved, you can orient two of them on the first slot and the remaining two using the ZBF2L cases that have 2 Edges already oriented.

Im not sure if this would reduce the amount of cases though....

 

[pjk]

That would be very hard to do. How could you preserve the EO while combining and inserting the last pair, and at the same time orienting all corners? That would be a ton of algs in itself, not counting all the ZBF2L algs from the third pair. I am not going to say it is impossible, but that would be an insane amount of work.

 

[Speedy McFastfast]

It would take a lot of algs to solve the corners, but what if you only concentrated on edges? I think breaking ZBF2L into two steps might work, if you are only worrying about edges.

Preserving edge orientation is easy once you get the hang of it. And with only 7 OLL cases, all algs should be sub 3 seconds or less.

 

[mizzle]

Instead of 500ish algs and all, think about this: orient edges at the beginning of the solve. Saves you a lot of cases for your third pair and gets you no rotations, reduced cases for your fourth pair, and an RUL F2L.

At this point, you have more options with what to do on your final pair. The three things left to do: Orient corners, permute corners, and permute edges. Of these three, orient corners would leave you with PLL, permuting them would get you the 60 something remaining easy ZB cases, and permuting edges would leave you with corners algs that largely commutators but quite a few that suck.

Corner orienting really wouldn't save you much time or many moves over conventional Fridrich, and complicates things a lot. I vote against.

Permuting them would require some serious ACube work, and really, would probably be hard as crap to recognize. Of course, it would leave you with under 70 relatively easy to recognize and perform ZB cases. This is worth exploring.

The final option is what I'm currently working on. Use the final pair to influence your corner orientation to always achieve U, T, L, or skip. Then, you have a smaller subset of ZB (216 and 21 PLLs) to learn. Not bad, compared to 500ish, especially because almost all of these are easy to perform, and the average movecount is something like 12. Short solves are great.

 

[Lucas Garron]

At this point, you have more options with what to do on your final pair. The three things left to do: Orient corners, permute corners, and permute edges.

There is one nice thing: Pre-orient with MGLS.

LS edge is often placed, often 2 extra moves (sometimes 3 or 4, like U'RUR')
This leaves you with CLS (2-gen) and PLL.
It may just about be the ultimate OH method (3-gen all the way after orientation!); I havent focused on it, though.

 

[Lofty]

How fast are you guys able to orient all edges at the beginning of the solve? With Petrus you have to only look at the edges for two faces of the cube making the edges easily visible and obvious, but to do them all at the very start at least to me seams a bit challenging to do all at once especially if you also have to focus on permutating two edges in order to make it 3-gen. I suppose with sufficient practice it won't be hard but I think I am just going to stick with full ZB (or VH for now).
I have practiced much OH and have found 3-gen is not always best.

 

[mizzle]

Well, orienting edges at the beginning averages something like five turns, and you can usually orient and get there or four of your cross edges done easily in eight to ten turns.

When I was using this full time, it took me about a second or two longer than standard cross, but because it made everything else easier, I was able to average in the 18s with it. My average with F2L -> COLL -> EPLL was only ever in the 19s, and I worked on it quite a bit longer than this.

I'm still very curious about permuting corners with the final pair. Has anyone really looked into this?

 

[AvGalen]

I'm still very curious about permuting corners with the final pair. Has anyone really looked into this?

Yes, but that is all I can tell you. I am sworn to secrecy.

 

[Speedy McFastfast]

Yeah, I don't think orienting all the edges at the beginning would be easy, or efficient. If you are going to worry about edges, Petrus seems to be the easiest way to do things.

Although here is what you Fridrich users can do you if you feel you need to stick with c/e pair solving: Solve cross, do two adjacent F2L pairs, fix bad edges, and then finish the solve with whatever crazy ZBcorners ideas your little heart desires

 

[mizzle]

Not easy or efficient? You mean...tacking three or so moves onto your normal cross is too difficult? Man, and all this time, I thought I was on to something...

Also, I never really liked Petrus orientation. I still prefer to do the orientation with a different definition of good and bad, so that the 2x2x3 is held at DL or DR, and I can fix edges using R or L turns (and R U R' type moves). Then, orientation is still 2 gen and even faster, and you can carry on as normal after. Always seemed so much better to me that way. Never understood why no one else did it like that.

 

[Speedy McFastfast]

Also, I never really liked Petrus orientation. I still prefer to do the orientation with a different definition of good and bad, so that the 2x2x3 is held at DL or DR, and I can fix edges using R or L turns (and R U R' type moves). Then, orientation is still 2 gen and even faster, and you can carry on as normal after. Always seemed so much better to me that way. Never understood why no one else did it like that.

Every fast Petrus user that I've met holds the 2x2x3 at DB. I'm thinking this is what you meant when you said you use R U R' type moves. I don't know why Petrus shows his orientation like he does on his site, but every faster Petrus user I know holds it like you seem to be explaining.

So much for nobody using it, eh?