The New Method / Substep / Concept Idea Thread

[DeeDubb]

Good point. But technically you don't need M for second block, you could do a more cfop style block
Also, as an example, if I were solving my blocks with white on bottom, the UL and UR pieces would be connected to their respective center, which would be yellow. Then the second block would basically be 2 F2L cases and 1 "cross" piece. It's truly not much harder than a typical second block, just a little less efficient.

Then you might as well do CFOP. The point of using M slices is that you don't need rotations, or F/B moves.

Also good Roux solvers can influence their UL and UR edges during EO, meaning preserving them after Step 1 is pretty irrelevant. Also, some CMLL algs take advantage pf the free M Slice, so you'd have to use less efficient CMLL algs.

 

[TorbinRoux]

Yeah, I myself just stick with original roux. This was just a proposal and adds many cfop aspects to roux. Also for EO to preserve the pieces on the bottom you could use olls like M' U M' U M' U2 M U M U M U2

 

[DeeDubb]

Yeah, I myself just stick with original roux. This was just a proposal and adds many cfop aspects to roux. Also for EO to preserve the pieces on the bottom you could use olls like M' U M' U M' U2 M U M U M U2

I appreciate the thought. I think Roux still has growing room. I don't think your suggestion is particularly an improvement as you claimed it to be in your original proposal, but I definitely appreciate people looking for ways to improve the method that I love.

 

[Rainbow Flash]

Two things:

1.
In the last few days I've pretty much become colour neutral because I've been learning Petrus and I want better crosses anyway. But the thing is, I've started unconciously combining Petrus and CFOP into one 'method.' I hate the cross, and I hate the expansion of the 2x2x2 block into a 3x3x3 block, so after finishing the 2x2x2 block I choose a colour of one of the faces of the block to complete the last two cross pieces, finish regular F2L (not the weird edge orientation thing in Petrus) and complete last layer. Now, I need someone to help me figure out whether this is good to do. The only pause in my Petrofop solves is when I choose the colour to complete the x-cross, and I'm not sure if this 'method' will be worth it when I get faster. I think this is a fairly good combination of methods, and I'm looking to improve it further...

2.
I'm trying to invent a method for the 4x4x4, which is going to be drastically different from Redux, Yau, and most other method out there for 4x4x4. Now, I need an algorithm finder, because I'm not going to spend time trying to figure them out myself. The program needs to...

1. ...find algs for certain cases.
2. ...calculate how many cases there would be for a certain step of my method.

Thanks.

 

[Christopher Mowla]

Now, I need an algorithm finder, because I'm not going to spend time trying to figure them out myself.

I understand you're probably busy, you got an idea, and you want to see it through before you forget it due to your busy life, but this just sounds really bad. You basically want someone to write a program that find algorithms (which is very unlikely, if you're talking about move-optimal, much less speed optimal) AND

2. ...calculate how many cases there would be for a certain step of my method.

you want people to count the cases for you too.

That's basically 99% of the work for creating a new method, especially for the 4x4x4 and larger cubes since such software has not been written.

If we don't know what the sub-steps of your method are, there is no way we can write a program which happens to do the exact calculations you will actually need without having to write a program which counts every possible scenario. (Which is impossible, because if we could write such a program, we would already be aware of all possible sub-steps/discovered all possible sub-step methods.)

Therefore, you must either state (and illustrate, if necessary) the sub-steps of your method so that one of us (I'm not going to personally because I, like you, don't have time) can write a program which calculates the number of cases in each sub-step of your specific sub-step method, or you must do all of this work on your own.

 

[Hypocrism]

Two things:

1.
In the last few days I've pretty much become colour neutral because I've been learning Petrus and I want better crosses anyway. But the thing is, I've started unconciously combining Petrus and CFOP into one 'method.' I hate the cross, and I hate the expansion of the 2x2x2 block into a 3x3x3 block, so after finishing the 2x2x2 block I choose a colour of one of the faces of the block to complete the last two cross pieces, finish regular F2L (not the weird edge orientation thing in Petrus) and complete last layer. Now, I need someone to help me figure out whether this is good to do. The only pause in my Petrofop solves is when I choose the colour to complete the x-cross, and I'm not sure if this 'method' will be worth it when I get faster. I think this is a fairly good combination of methods, and I'm looking to improve it further...

2.
I'm trying to invent a method for the 4x4x4, which is going to be drastically different from Redux, Yau, and most other method out there for 4x4x4. Now, I need an algorithm finder, because I'm not going to spend time trying to figure them out myself. The program needs to...

1. ...find algs for certain cases.
2. ...calculate how many cases there would be for a certain step of my method.

Thanks.

I don't think your petrofop is a good idea - the advantage of CFOP is that you do your setup at the beginning (cross planned and first F2L pair found in inspection) so that there are no real analytic steps for the rest of the solve and it can be done with simple fast TPS and look-ahead/know-ahead. The advantage of petrus is that it is move efficient, and if you get good at the analytic block building steps, you can get fast times because of the lower move count despite requiring generally lower TPS and has more propensity for pausing. By combining the two in this way, you lose the simplicity/execution advantage of pure CFOP (because you have to stop and think about your cross after your first step), and you lose the block-building/move count advantage of petrus (because you build a cross and this is less move-efficient).

I would use your petrus experience to practise seeing when an x-cross can be easily built instead of trying this.

 

[TorbinRoux]

I saw a video awhile back of someone doing a blindfold solve by memorizing where each of the piece were, then put on the blindfold and took the cube apart, organizing the pieces into groups, then put them back into the cube solved. This is sort of a new method, and I'd like to know if it'd be allowed in competition. Just curious, because it's very creative.

 

[TDM]

I saw a video awhile back of someone doing a blindfold solve by memorizing where each of the piece were, then put on the blindfold and took the cube apart, organizing the pieces into groups, then put them back into the cube solved. This is sort of a new method, and I'd like to know if it'd be allowed in competition. Just curious, because it's very creative.

I think that was kinch...
It wouldn't be allowed. I'm not sure which specific regulation it is (maybe I'm being blind and not seeing something obvious?) but this one could be relevant:

5b1) If a competitor chooses to repair the puzzle, he must repair only the defective parts. [...]

Doing this is also not under the list of permitted repairs.

 

[lerenard]

Two things:

1.
In the last few days I've pretty much become colour neutral because I've been learning Petrus and I want better crosses anyway. But the thing is, I've started unconciously combining Petrus and CFOP into one 'method.' I hate the cross, and I hate the expansion of the 2x2x2 block into a 3x3x3 block, so after finishing the 2x2x2 block I choose a colour of one of the faces of the block to complete the last two cross pieces, finish regular F2L (not the weird edge orientation thing in Petrus) and complete last layer. Now, I need someone to help me figure out whether this is good to do. The only pause in my Petrofop solves is when I choose the colour to complete the x-cross, and I'm not sure if this 'method' will be worth it when I get faster. I think this is a fairly good combination of methods, and I'm looking to improve it further...

This isn't a new idea, I and (last time I heard) Jlarsen solve this way at least some of the time. The idea is that expansion to the 2x2x3 block requires seeing three pieces and how to put them together: you can't do anything until you have all three or at least one corner and one edge. This leads to pauses more often than not which, of course, are the speedsolver's arch nemesis. Tracking two cross pieces while you build your 2x2x2, on the other hand, is easier because there are only two pieces and also because each one can be placed individually. It is also very easy to place them both which facilitates lookahead into the first (second?) pair. I actually feel like it combines the advantages of the two methods instead of failing to make use of either. It's the most common incarnation of FreeFOP, which means you just build the F2L whichever way seems easiest. I will still go full Petrus if I see two pairs in inspection that can be part of the 2x2x3, but I always wait to do EO until the last slot (see: VHLS). I will also just use full CFOP if the 2x2x2 is annoying or I don't see one or there's an easy cross.

 

[2180161]

So this is a concept idea thread, so I was wondering if anyone had thought of this for look-ahead.
have your cube on really really loose tensions, so you are forced to turn slow, other wise it will pop. This gives you an opportunity to get better look-ahead, and then slowly tighten your cube.You then can end up with your normal tensions, and good look ahead.

 

[TDM]

So this is a concept idea thread, so I was wondering if anyone had thought of this for look-ahead.
have your cube on really really loose tensions, so you are forced to turn slow, other wise it will pop. This gives you an opportunity to get better look-ahead, and then slowly tighten your cube.You then can end up with your normal tensions, and good look ahead.

You could just practise your 2H method OH... that forces you to turn slower too.

 

[lerenard]

So this is a concept idea thread, so I was wondering if anyone had thought of this for look-ahead.
have your cube on really really loose tensions, so you are forced to turn slow, other wise it will pop. This gives you an opportunity to get better look-ahead, and then slowly tighten your cube.You then can end up with your normal tensions, and good look ahead.

Unless you have a weilong, which is unpoppable.

 

[2180161]

I have a weilong, and it does pop a crap ton. It pops about 1 in every 5 solves, even on tight tensions.

 

[GuRoux]

I have a weilong, and it does pop a crap ton. It pops about 1 in every 5 solves, even on tight tensions.

weilong doesn't really pop easily unless it's very, very loose or you took out the torpedoes.

 

[PenguinsDontFly]

weilong doesn't really pop easily unless it's very, very loose or you took out the torpedoes.

Wait what you can do that!

 

[lerenard]

I have a weilong, and it does pop a crap ton. It pops about 1 in every 5 solves, even on tight tensions.

Then you're doing it wrong.

Mine has literally never popped on me. The center caps, on the other hand, seem to come off every other z-perm unless I put tissue paper under them.

 

[lerenard]

Came up with a new 4x4 method today. I looked it up and apparently it's called Meyer, except mine is slightly different. After you build your first block you build an additional 1x2x2 on the other side. This leaves even less to be solved after edge pairing (only 9 edges left, 8 if you insert one that happens to build itself into the empty slot sometime between block building and the end of center building (this happens to me about half the time). It's different from stadler, though, because after centers you do normal edge pairing. Having the E slice, F face and the U face minus a 1x1x2 block gives you just enough breathing room to easily match up edges. I suck at roux so I haven't even bother timing myself with this, but I think it could be very fast, at least in cases where the 1x2x2 block is obvious.

 

[qqwref]

Do you mean 1x3x3 and 1x1x3? It's an interesting idea, I'll have to try a solve on it sometime.

 

[lerenard]

Complete steps:
1: opposite centers
2: 1x3x4 on L, 1x3x3 on R
3: Remaining centers
4: x' z' solve remaining 9 edges using E slice, F, and unsolved portion of U
5: z x solve remaining pair, making two 1x3x4 blocks
6: CMLL
6: LSE, possible parity as in normal redux solves (I'm assuming those who already use roux know how best to deal with these .)

 

[penguinz7]

Sounds like a lot of moves..
Also sounds like 3x3 ortega, but I don't actually know how that works so I can't say.