The New Method / Substep / Concept Idea Thread

[notfeliks]

I see what you mean; you'd use it differently to how I was thinking about it i.e. you would only use it occasionally. I was thinking more of using it all the time. If you only use it occasionally, it would probably be something you'd remember when solving.
Also wouldn't a normal cross be useful 50% of the time?

If you only solve on one colour, then yeah, probably. I'm CN though, and I only go into this if there's not an obvious 4-5 cross on one side and a 2-3 pseudo-cross on another.

 

[TDM]

If you only solve on one colour, then yeah, probably. I'm CN though, and I only go into this if there's not an obvious 4-5 cross on one side and a 2-3 pseudo-cross on another.

I still think if you had no preference it would be 50/50, but I don't know how to prove it.

 

[notfeliks]

I still think if you had no preference it would be 50/50, but I don't know how to prove it.

What you are missing is the move count proportion that makes forming a psuedo-cross advantageous. If you have a 5 move normal cross with decent finger tricks, and a 4 move psuedo cross with decent finger tricks, you are obviously going to go for the normal cross (unless there's some other advantage, but that's unrelated). The cross move count or resulting F2L needs to be greatly better to justify using a psuedo-cross over a normal cross, hence the disproportionate use of it in solves, and especially CN solves.

 

[TDM]

What you are missing is the move count proportion that makes forming a psuedo-cross advantageous. If you have a 5 move normal cross with decent finger tricks, and a 4 move psuedo cross with decent finger tricks, you are obviously going to go for the normal cross (unless there's some other advantage, but that's unrelated). The cross move count or resulting F2L needs to be greatly better to justify using a psuedo-cross over a normal cross, hence the disproportionate use of it in solves, and especially CN solves.

So you're saying this method is worse than normal CFOP, unless the pseudo-cross is much easier?

 

[notfeliks]

So you're saying this method is worse than normal CFOP, unless the pseudo-cross is much easier?

Yes, absolutely; the only real advantage (besides weird skips and slight F2L improvements if you notice them) that the pseudo-cross provides is that it will be much easier to make when you decide to use it.

 

[Logiqx]

Every now and then I enjoy solving with a variety of the methods which are already known to the community.

I was pondering alternatives to the first three stages of Petrus a few days ago (2x2x2, 2x2x3 and edge orientation). The goal was to come up with something finger-trick friendly and to avoid the move restrictions imposed after building a 2x2x2.

I came up with the idea below and did a dozen or so solves to see if it might have any merit. I found it to be quite enjoyable and I think it could be pretty quick in the right hands.

I've trawled this thread looking for anything similar and couldn't find anything. My apologies if it has already been proposed elsewhere on the forum.

1: Block build a 2x2x1 on left side (LDB) using whatever moves seem best, aiming for 4 moves
2: Block build a 2x2x1 on the opposite side (RDB) using rRMU moves, aiming for 8 moves or less
3: Orient the remaining 8 edges (typically you'll have 4 bad edges). Modified recognition due to a pending rotation!
i) Place 2 oriented edges in FL + FR using FU moves, occasionally using M' to flip edges
- 25% of the time FL and FR are already oriented so it's no moves... Yay!
- 25% of the time FU and FD are already oriented and it's just an F or F'... yay!
ii) Orient the remaining 6 edges using MU moves (L6E cases in Roux)
- The "Moo" case seemed to crop up a lot for me but I might have been lucky!
4: Place DB edge whilst solving M-slice centres (i.e. 2x2x3 block at the back) using MU moves, typically 3 or 4 moves

Rotate (y') so the 2x2x3 is on the left and finish the solve as per Petrus or ZZ...
5: Finish the F2L using RU moves (i.e. Block build 2x2x1 and insert final F2L pair)
6: Complete the Last Layer using whatever you know - edges already oriented (i.e. OCLL + PLL or COLL + EPLL)

Advantages:
Step 1: Planning the first 2x2x1 is really easy, even for block building newbies. Ok for "zero inspection" event?
Step 2: Ability to track during the first 2x2x1. Advanced solvers should be able to plan part of it during inspection.
Step 3: Orienting the 8 edges is easy. After FL and FR it's basically the Roux cases but slightly modified recognition.
Step 4: This can be merged into step 3 in the same way that Roux solvers can place UL an UR during orientation.
Step 5: Building a 2x2x1 block can be done efficiently (move count) and is finger-trick friendly (RU moves).
Step 6: You can use the well known algorithm set(s) such as PLL so there isn't anything new to learn.

Tips:
Step 1: Try to be colour neutral for U and D (e.g. white/yellow on top/bottom, neutral for sides)
Step 2: You still have a lot of freedom during the second 2x2x1 and you can even use an F move if it is desirable
Step 3: This is intuitive so you should be able to find your own algorithms and shortcuts
Step 4: Ditto
Step 5: Take advantage of the visibility prior to rotation. LS extensions (e.g. WV) can be used to influence the LL.
Step 6: Spam!

Personally, I find this to be quite a fun way to get to the latter stages of Petrus and ZZ, utilising Roux techniques. It focuses on the 2-dimensional block building of Roux + ZZ, rather than the 3-dimensional block building in Petrus. Edge orientation uses the Roux-style (MU moves) but is a little earlier than Petrus + Roux but nowhere near as extreme as ZZ.

I'm wondering what more experienced / knowledgable solvers make of it. If nothing else, I hope it is a fun variation!

 

[elrog]

I have done this before and it is quite fun for doing something different. The way I did it was basically the same thing, but slightly different.

I would not do F moves at all. If the LF and RF edges are oriented and the DB edge is not one of them, I'd place DB while orienting the remaining edges. If that is not the case, I would use M and U moves (and sometimes R moves) to orient UL and UR and make sure neither of them is the DB edge. I would then do an x' rotation and orient the remaining edges while solving the DB (now DF) edge. You would also then do a y rotation instead of a y' rotation (assuming you are right-handed).

For speed, I think it can be good, but I don't think it is groundbreaking. If someone were able to plan both blocks, plan to have LF and RF oriented, and plan to keep DB out of the LF and RF position all during inspection, I think this could be a great method. I think going that far is pushing it though.

EDIT: Oh, and don't forget to solve the centers while you orient the edges and place the DB edge.

 

[Logiqx]

I have done this before and it is quite fun for doing something different. The way I did it was basically the same thing, but slightly different.

I figured someone must have played around with this idea. I like your trick of rotating around the x-axis in some scenarios.

For speed, I think it can be good, but I don't think it is groundbreaking. If someone were able to plan both blocks, plan to have LF and RF oriented, and plan to keep DB out of the LF and RF position all during inspection, I think this could be a great method. I think going that far is pushing it though.

That was my general conclusion. It's a fun method and pretty efficient but it doesn't improve on the "big four".

EDIT: Oh, and don't forget to solve the centers while you orient the edges and place the DB edge.

I've got that in step 4, just hiding amongst the rest of the text.

Thanks for the feedback!

 

[elrog]

I've got that in step 4, just hiding amongst the rest of the text.

I wasn't meaning that you left it out. I accidently left it out of my post.

 

[RatJelly]

I want to preface this proposed synthesis of two methods by saying 1. I haven't been cubing long and 2. I primarily solve using Roux but SUCK at block building. I only have a PB of 30 seconds and average of 10 is generally around 40 seconds. In my 40 second solves a full 30 seconds is used for building the F2B which is quite frustrating for me since the rest of my solve is much more fluid and quick. Here is what I've come up with to help reduce my times/improve the F2B. So far I have reduced my F2B to 23 seconds but my overall times have gone up since I switched to COLL from CMLL and my execution isn't quite there.

Method: Roux and ZZ synthesis (RouxZZ if it needs a name

Step 1: Use inspection time to locate bad edges

Step 2: EO (edge orientation) + line except place the line pieces at DL and DR instead of DF and DB for ZZ.
-advantage: allows solving of F2B like ZZ F2L with no cube rotations and helps a bad block builder like me recognize pairs. Since all edges are oriented there are only specific cases/locations on the cube where edges can pair up. Additionally, in contrast to Roux this allows both the R and L blocks to be solved simultaneously which increases the chances you can find a good pair. For the second 1x2x3 block it allows you to easily solve U layer corner pieces with the D stick facing up which is harder to do in standard Roux.
-disadvantage: F, B, and M slice now have restricted movement unless its an F2, B2, or M2. But this can also be an advantage i.e. edge piece you want to pair is on D layer, since all edges are oriented you know without having to think that you have to pair with its corner piece diagonally up in the U layer w/ an M2

Step 2 alternative: just EO without placing the line.
-advantage: this allows for solving the F2B more like Roux was intended to be solved, L or R 1x2x3 block first then the other. Corner edge pairs, an edge paired with the center sticker, and corner paired with DL/DR are all viable block building options. Focus for look ahead can be geared towards a smaller set of pieces.
-disadvantage: Reduced to doing one 1x2x3 block then the other. If necessary for block building restricted to F2, B2, M2, r/r'2, and l/l'2

Step 3: F2B/ZZ F2L whichever step 2 you pick

Step 4: U layer corners solved using COLL to maintain proper edge orientation
-advantage: forces initial step 5 skip in Roux (in regular Roux this is would be a step 4a skip) since your edges are already oriented

Step 5 (just finish solving Roux style): Solve UL and UR edges to complete L and R sides, permute the rest of the M slice aaaaaaand your done.

Let me know what you guys think, I'm interested to hear more seasoned cubers take on this. Still not fully used to all the notation so hopefully you can all follow this. I haven't done a move count to see how it actually compares to Roux, but to me the real advantage is easier F2B since all the edges are oriented, the step 4a skip is a nice addition. Even if its not viable as a speed solving method I enjoy it cause I can actually look ahead much better vs standard Roux.

PS- In case they are on here, just wanted to say thanks to Waffle and Donovan for their Roux tutorials and to Phil for his ZZ tutorial!

 

[DeeDubb]

Let me know what you guys think, I'm interested to hear more seasoned cubers take on this. Still not fully used to all the notation so hopefully you can all follow this. I haven't done a move count to see how it actually compares to Roux, but to me the real advantage is easier F2B since all the edges are oriented, the step 4a skip is a nice addition. Even if its not viable as a speed solving method I enjoy it cause I can actually look ahead much better vs standard Roux.

I'm not super seasoned, but I'll give some feedback. I think this is an interesting concept for solving, but just not great as a speed solving method. You're using your inspection time for orienting edges, which means you're doing your first block on the fly. Seasoned Rouxers can do the first block blindfolded and efficiently, so EO is definitely not a benefit there. The move restrictions to preserve oriented edges will definitely hurt your move count for first and second blocks. Using COLL instead of CMLL hurts your move count a bit too in some cases. Your skip of 4a isn't really a skip since you had to orient those edges earlier anyway.

This sounds like a fun way to practice EO, but not really efficient, as the EO step is very unnecessary once you are comfortable with block building, and one of the greatest advantages of Roux is the move freedom allowed during block building.

 

[supercavitation]

That sounds like this: http://www.speedsolving.com/forum/showthread.php?11536-Stachu-Korick-s-New-Method-)

It's a decent method, and the pros and cons have been explained pretty well there (I know he says to use WV, but he also says that COLL works as long as EO is maintained).

 

[RatJelly]

I'm not super seasoned, but I'll give some feedback. I think this is an interesting concept for solving, but just not great as a speed solving method. You're using your inspection time for orienting edges, which means you're doing your first block on the fly. Seasoned Rouxers can do the first block blindfolded and efficiently, so EO is definitely not a benefit there. The move restrictions to preserve oriented edges will definitely hurt your move count for first and second blocks. Using COLL instead of CMLL hurts your move count a bit too in some cases. Your skip of 4a isn't really a skip since you had to orient those edges earlier anyway.

This sounds like a fun way to practice EO, but not really efficient, as the EO step is very unnecessary once you are comfortable with block building, and one of the greatest advantages of Roux is the move freedom allowed during block building.

Thanks for your input. I see what you mean about it not really being a "skip" since you had to orient earlier. Your post brings up a point about perspective too. I wrote my post more as a variation of Roux. While the move count might be higher than Roux, I wonder how it compares to ZZ. If you look at it from the perspective that this is ZZ to start with the ability to utilize the M slice and finishing with Roux. The ability to use the M slice could be considered an advantage in ZZ rather than looking it as an M2 restriction in Roux. Maybe I'm just trying to hard to justify using it hah. anyway, thanks for analyzing it for me and thanks for your suggestions a couple months ago in the Roux thread, you really helped me out when I first started cubing.

 

[RatJelly]

That sounds like this: http://www.speedsolving.com/forum/showthread.php?11536-Stachu-Korick-s-New-Method-)

It's a decent method, and the pros and cons have been explained pretty well there (I know he says to use WV, but he also says that COLL works as long as EO is maintained).

Yep, that's exactly what I was trying to explain. Thanks for pointing me in the right direction! Never thought of doing M U2 M' type moves, that helps a lot.

 

[GuRoux]

Thanks for your input. I see what you mean about it not really being a "skip" since you had to orient earlier. Your post brings up a point about perspective too. I wrote my post more as a variation of Roux. While the move count might be higher than Roux, I wonder how it compares to ZZ. If you look at it from the perspective that this is ZZ to start with the ability to utilize the M slice and finishing with Roux. The ability to use the M slice could be considered an advantage in ZZ rather than looking it as an M2 restriction in Roux. Maybe I'm just trying to hard to justify using it hah. anyway, thanks for analyzing it for me and thanks for your suggestions a couple months ago in the Roux thread, you really helped me out when I first started cubing.

through the few solve i have tried, i think it has better movecount than zz, probably close to regular oux.. It's kind of weird because it restricts you from doing F moves which i almost always use in the first block. Also, i'm not sure if you can do this as color neutral as most roux solvers are. I think most zz user can only start on around 2 positions, which adds extra constrictions and limits. As dedub said, you'll need some stellar lookahead for the first block. Also, I'm not a fan of the inability to use Ms and r for the second block and using coll. i think for pros there is better recognition and a more move efficient eo usually. For speed, it's probably just better to pick zz or roux.

 

[elrog]

I am going to compare this method (with a slight) variation to ZZ.

When doing EO, you should solve the line by placing two opposite U-layer edges in place of DF and DB, but you should keep the D center color on D.

Through the solve, the method is the exact same as ZZ (with COLL/EPLL) until the last step.
1 - bad EO line / regular EO line
2 - Left block
3 - Right block
4 - COLL
5 - second half of L6E / EPLL

In the first step, you have 2 lines to choose from, but this will only help if you normally have extra time to look for another option. For the last step, L6E is shorter than EPLL, and more intuitive. Of course, with time/experience, intuition and algorithms become the same thing (to a certain extent). You would just learn EPLL as algorithms and L6E as intuition.

While this seems slightly better than ZZ as shown here, I think ZZ with OCLL/PLL is slightly better for 2H than the ZZ variant I compared it to here. I won't compare this directly to ZZ with OCLL/PLL because that comparison is not as easy to make. Also, I can't see L6E being better than EPLL for OH because EPLL can be solved using only R and U.

 

[obelisk477]

Just a quick question. I like Roux, but am just coming to it out of CFOP really just to do some practice with M, M' and M2 fingertricks. So I came up with a weird hybrid that I like to use when practicing which consists of these steps:

1) Solve two opposite cross pieces only, and leave the 'line' (EOline without the EO) on bottom, with the solved edges at LD and RD extending to the centers on the L and R faces.

2) Do f2l with no cube rotations, by doing EO with M or M' turns. So if the edge requires a cube rotation to place it, move the edge to UF or UB, and use M or M' to place it at the opposite position on U, and then do the f2l insertion with the newly oriented edge.

3) Roux from here on out with just M and U turns.

Not sure that this is any faster than Roux, but could be a useful way for people too accustomed to CFOP to get used to the system.

And so the question is, is this a thing? Like a method some people use? Or just a bad idea

 

[guysensei1]

Just a quick question. I like Roux, but am just coming to it out of CFOP really just to do some practice with M, M' and M2 fingertricks. So I came up with a weird hybrid that I like to use when practicing which consists of these steps:

1) Solve two opposite cross pieces only, and leave the 'line' (EOline without the EO) on bottom, with the solved edges at LD and RD extending to the centers on the L and R faces.

2) Do f2l with no cube rotations, by doing EO with M or M' turns. So if the edge requires a cube rotation to place it, move the edge to UF or UB, and use M or M' to place it at the opposite position on U, and then do the f2l insertion with the newly oriented edge.

3) Roux from here on out with just M and U turns.

Not sure that this is any faster than Roux, but could be a useful way for people too accustomed to CFOP to get used to the system.

And so the question is, is this a thing? Like a method some people use? Or just a bad idea

Cyoubx made a rotationless F2l video once and he did this, and called this a 'stupid version of Roux'.

 

[obelisk477]

Cyoubx made a rotationless F2l video once and he did this, and called this a 'stupid version of Roux'.

link? i tried finding it but to no avail

 

[guysensei1]

link? i tried finding it but to no avail

http://www.youtube.com/watch?v=Js1LZscj18A