MGLS short cuts?

[fanwuq]

I'm learning MGLS right now, it's a fun method and seems fast, but I'm a bit discouraged because it's not really lower move counts than regular Fridrich. But looking at the method, there seems to be many shorts cuts you can take even at the beginning of F2L.
One thing I want to try is to just insert one edge rather than the CE pair for the 1st, 2nd or 3rd slot. However, there is one major problem: you can end up inserting the wrong corner that also belongs in the D layer, so that messes up the rest of your F2L pairs. The only way to do this is with EJF2L. But it's not always worth it, especially if you have to orient many edges. With a situation like that, I think it would be more move efficient to just do that pair and then CLLEF or OLL.
One thing with MGLS is that there is almost no lucky cases, while it's very common to get skips with COLL.
Perhaps using MGLS as a main method is not worth it? Then what would be the best way to combine different F2L/LL strategies to always end up with the easiest cases?
So far I propose:
F2L: Do normal Petrus or Fridrich if convenient, or insert a wrong pair like EJ
LS: If EJ option was chosen, orient before you get to the last slot I do not recommend EJ inserting the 3rd pair unless you when you do it, all the edges are already oriented.
If EJ was not chosen: If no edges need to be oriented, ELS by RUR', then CLS.
If all edges are wrong, insert, then CLLEF+EPLL
If some are right, some are wrong, do whatever you feel like, maybe just finish with regular Fridrich.
Conclusion: This is probably way to much to think about during a solve and too many algs to learn.

 

[Escher]

i think that you should find one system that you really like and get it as fast as you can possibly get it. i think that mixing and matching systems all the time is very hard to get as fast as you could if you just spent the same amount of time on one system...
to be honest, i think OLL and PLL is as good as it gets for the majority of people. recognition is very easy, and the vast majority of algs are very fast.

 

[Dene]

OLL skips are not that common, and chances are if you are going to have an OLL skip, it would be obvious thus you would not use MGLS in that solve.

 

[ErikJ]

it sounds good but if you are already inserting an incorrect pair in the F2L and you don't do bad edges petrus style then it would take a lot of time to insert the last pair and flip edges.

btw, sometimes when I'm tired and I don't pay attention I'll end up with 2 EJ pairs in the F2L.

 

[fanwuq]

it sounds good but if you are already inserting an incorrect pair in the F2L and you don't do bad edges petrus style then it would take a lot of time to insert the last pair and flip edges.

btw, sometimes when I'm tired and I don't pay attention I'll end up with 2 EJ pairs in the F2L.

Yeah, that's one thing I need to figure out. If I do ELS at last slot when there's already an EJ pair, then I'd end up with 2 opportunities for CLS. Perhaps ZBF2L? That's probably also bad. There's got to be a simple and elegant solution.

I guess my main problem is to find a way to do MGLS where I can always just use RUR' to do one edge and skip a slot without taking in a bad corner (one that belongs in a different slot). Then, I should work on edge control throughout the solve and not just at last slot.
ZZF2L looks appealing here, but the main thing that I don't like about it is that you have to search for your blocks after inspection after you oriented the edges. Also LUR does not really appeal to me, most of the time when I'm solving calmly, I rarely get more than rotations during F2L.

 

[ErikJ]

you could do EJF2L but instead of inserting the corner wrong you just put a random LL corner in it. so after F2L you just do CLS with the F2L corner in the LL. I'd try this but I don't plan to learn the cases and I'd get confused during step 4 seeing 3 corners that go in the F2L.

 

[fanwuq]

I'm thinking about doing ELS at the 3rd F2L slot instead of 4th, not sure how much better that would be...

 

[blah]

On the contrary, I think it's much easier to get a "worked-for" skip in MGLS, I think I mentioned this somewhere else in the forum.

To skip a fourth F2L slot (and I mean work for that skip), the only known method, I think, is multislotting, which only a handful of people in the world do.

To skip ELS, all you need is edge control, which is much easier than multislotting. Edge control involves tracking all edges, which are always there. Multislotting involves tracking specific corner-edge pairs, which can be a real pain.

If I'm not wrong, I think Lucas should be getting loads of ELS skips since I believe he uses a very heavy amount of edge control from the first F2L pair onwards.

Is there something I missed that makes this whole theory wrong?

And I think I just got a new idea after typing all that, I'll do some research on it before I post it.

 

[fanwuq]

How can you use edge control from the first pair onward? That's exactly what I'm looking for, but I'm not really sure how. Sometimes you have no yellow pieces in U to reference and when you rotate to solve the 2nd pair, it messes up everything.

This is what I'm doing today:
1. 2x2x2 block (easy, usually high move count (5-9), but sub-2 seconds)
2. Insert 2 edges for pseudo 2x2x3, make sure that there isn't a bad corner inside. (difficult to recognize, but total moves is usually 8-16)
3. Orient edges (I suck at recognition and does not do it optimally. 8-23 moves)
4. Finish F2L (3-5 seconds for this step. 15-35 for total move count.)
5. CLS (I don't know many algs, but assuming that I do, it would be 22-50 moves)
6 PLL (Easy, finished with 31-67 moves.)
I'm generally getting solves around 40-50 if I looked up the CLS or just assume it is an average of 10 moves.

I'm liking this method. I just need to work on step 2 and 3 and learn all the CLS.

 

[blah]

How can you use edge control from the first pair onward? That's exactly what I'm looking for, but I'm not really sure how.

Not too sure myself, because I take forever to recognize the yellow/white edges (I'm opposite color neutral), but I guess it's doable with a trained eye. What I do, which may very well be totally wrong, is just to maximize the number of good edges on the U layer from time to time, that's all I can think of.

 

[fanwuq]

How can you use edge control from the first pair onward? That's exactly what I'm looking for, but I'm not really sure how.

Not too sure myself, because I take forever to recognize the yellow/white edges (I'm opposite color neutral), but I guess it's doable with a trained eye. What I do, which may very well be totally wrong, is just to maximize the number of good edges on the U layer from time to time, that's all I can think of.

Oh , I do a little of that when I can, but it doesn't work on every solve. If only I was good at Petrus...
Any hints from Justin, Lucas, and Dene?
I wonder how would VH F2L work on the 3rd slot. If that is easy to use, almost everything would be solved.

 

[brunson]

How can you use edge control from the first pair onward? That's exactly what I'm looking for, but I'm not really sure how.

Not too sure myself, because I take forever to recognize the yellow/white edges (I'm opposite color neutral), but I guess it's doable with a trained eye. What I do, which may very well be totally wrong, is just to maximize the number of good edges on the U layer from time to time, that's all I can think of.

I've been playing with it a bit while learning ZBLS. I only try to do it when I have a joined edge/corner in U, then I look at the slot it's going into. Using the same rules for BLD EO, but especially if I have a U colored cubie in or adjacent to the slot, I choose to insert it with U F' U' F or F R' F' R.

 

[fanwuq]

I finally figured out what I'm trying to do and how to do it! I need to insert 3rd slot while solving ELS.
What do I need to do? VH3S (VH 3rd slot)
What is VH3S?
VH3S is a method where you do VHF2L for the 3rd slot so that the fourth slot and OLL is 2gen. (My goal is actually trigger for ELS and then CLS.)
When the last slot is oriented, it's VH3S is the same as VHF2L.
When it is not oriented, there are 3 more sets of VHF2L like cases. They are BR is flipped, LF is flipped, and LB is flipped.
Here is the downside:
1. If the LS edge is in U layer, recognition can be difficult.
2. There are a lot of algs.
I'm not so afraid of the number of algs, but that different edge would be annoying.
I guess I'll have to generate the algs myself.

After figuring out all this, I shall just learn regular VHF2L and actually use it for LS and forget about this idea for several months until I get nice at VHF2L. Guess learning MGLS is not that easy?! Apparently you have to mess with some type of mini-ZB before you could learn MGLS.