The New Method / Substep / Concept Idea Thread

[CodingCuber]

How do i make it speedsolvable

Tbh the method would have to be completely changed into a different method for speedsolving. There are too many rotations and a high movecount.

 

[Devagio]

How do i make it speedsolvable

When you make a method for speedsolving, you need to do quite a bit of work. CFOP for instance took months worth of time, and development by many people across years to make it what it is; it’s not something one can beat overnight.
Recognising that, here is a bare-minimum of what you should do:
1. Make a method
2. Estimate movecount of each step
3. Estimate alg count and see recognition of each alg step
4. See recognition, look ahead, and ease of intuitive block building steps and even other kinds of steps like EO, CO, etc.
5. Identify chokes in your method and see whether you can improve them.

These can be done in any order of course, but this is by no means exhaustive. Not even close.

For your new method, here is a little move analysis;
EO-psuedocross would be 12 moves
3 belt pieces would be 12 moves
Keyhole 4 corners would be 20 moves
Last edge would be 8 moves
OCLL on top would be 8 moves
2 PLLs would be 30 moves
Parity would be 3 moves

This is 90+ moves; and it seems unlikely to have double TPS of Roux (~45 moves) so this is objectively worse. Maybe you could pary this by combining step, etc.

 

[Hazel]

How do i make it speedsolvable

Here is some advice:

- To invent a good method, you have to understand what makes a method good or bad. You said you're about sub-50, so I don't think you're at the point where you have a solid grasp on that. It just takes time and experience, especially with trying out various other methods.
- In general, solving something incorrectly and then fixing it later is slower than solving it correctly in the first place. In your ZZ-Bridge method, it would simply be faster to solve the cross correctly the first time, same with the corners. But at that point, you're just doing a much worse version of CFOP, which ties into my next point:
- If a method is just a worse version of an existing method, it isn't worth pursuing. If I proposed a method where you solve two F2L pairs, then insert 3 cross edges around it, then do EO followed by OCLL/PLL, that's just a worse version of Petrus, and so there's no point.
- A huge amount of methods have been proposed and developed for 3x3. Making a method that's unique and hasn't been thought of before is very difficult. Making a method that's fast is very difficult. Making a method that hasn't been proposed before and has very good potential is incredibly hard. This goes back to my first point—you need to have a very thorough understanding of the cube and of many different methods to even have a chance at making one that's worth people's attention.

This isn't to say that you should stop trying to make methods. We as a community need people who are inventive and care about pushing forward the boundaries, or we'll never get anywhere. However, you should be careful not to have baseless confidence in your creations. If you think a method you create is good, you need to be able to explain what makes it unique and what makes it fast, and you need to be able to identify its strengths and weaknesses before you propose it for the first time. You need to have reasons why your method is worth our time.

Good luck with your future method creation

 

[ribbon method]

How about 3 Cross edge pieces(mock cross), f2l, co, cp, l5e

I will call it 3cfccl/mcfccl

 

[Hazel]

How about 3 Cross edge pieces, f2l, co, cp, l8e

First of all, why would you do CO and CP separately? Doing that in one alg total would be faster.
Secondly, that already exists: https://www.speedsolving.com/wiki/index.php/PCMS

 

[ribbon method]

First of all, why would you do CO and CP separately? Doing that in one alg total would be faster.
Secondly, that already exists: https://www.speedsolving.com/wiki/index.php/PCMS

Thats not the same

 

[Hazel]

Thats not the same

From the wiki page:

Steps:

1. Columns:
   1. Solve 4 CE pairs using intuition.
   2. CMSLL (42 algs), something between 2x2 CLL and CMLL (see CxLL).
2. Slices:
   1. Solve centres and the RD, LD and BD edges using intuition (3x centre/edge pairs actually).
   2. L5E, last five edges, solve the FD edge and the LLedges in two short alg steps.
      1. L5EO, orientation (5 algs minumum but here you can add more to force easier permutations).
      2. L5EP, permutation (16 algs).

Your method and PCMS are exactly the same, it's just layed out in more detail on the wiki page.

 

[ribbon method]

Mine is not the same I do co and cp if I did co cp together that just be roux with an extra cross edge

 

[Hazel]

Mine is not the same I do co and cp if I did co cp together that just be roux with an extra cross edge

Ah, you edited the post after I first replied.
Solving CO and CP at the same time via CxLL is objectively better than solving them separately (meaning if you're going for speed, there is no good reason not to do them at the same time).
So let's replace your CO+CP with CxLL—now your method is just CFOP with a slightly worse F2L and a slower last layer (CxLL is about as fast as OLL, and L5E is definitely worse than PLL).

So, if the method is just a worse version of CFOP, why would anybody want to use it?
Again, I don't mean to be rude and only filled with criticism—these are just the kinds of things you should be thinking about before proposing new methods.

 

[ribbon method]

Or u could do
Solve LEFT BLOCK

Then a 2x2x2 in RIGHTBACK and then expand to a 2x2x3

Solving f2l then do whatever LL

 

[Username: Username:]

Or u could do
Solve LEFT BLOCK

Then a 2x2x2 in RIGHTBACK and then expand to a 2x2x3

Solving f2l then do whatever LL

worse Freefop.

 

[ribbon method]

goddam it

 

[Username: Username:]

This is getting annoying, like @OreKehStrah has said once, your ideas screams "I want to make a method that is obviously bad so that I have a wiki page and the knowledge of having created a method."

 

[ribbon method]

How is it bad I dont get how yo make a good or bad method how do u tell

 

[WoowyBaby]

How is it bad I dont get how yo make a good or bad method how do u tell

If it is fast, it is good.

 

[ribbon method]

Anything else cause zb is fast but way too many algs

 

[Hazel]

Anything else cause zb is fast but way too many algs

It has to be fast, and there should be reasons for people to want to use that method over other, more popular methods.

 

[RedstoneTim]

Or u could do
Solve LEFT BLOCK

Then a 2x2x2 in RIGHTBACK and then expand to a 2x2x3

Solving f2l then do whatever LL

Actually it even already exists as a method separate from FreeFOP.

 

[PetrusQuber]

Or you could just not invent a method.....
99% of methods in this thread get ripped up since they can’t compare to normal methods.

 

[BenChristman1]

Or you could just not invent a method.....
99% of methods in this thread get ripped up since they can’t compare to normal methods.

Yes, sometimes (99% of the time) you need to just give up.