How many groups of methods are there to solve a 3x3?

[FJT97]

So what i meant is that there are plenty of methods which basically are layer by layer.

Then there is all the methods which do corners first (or edges first) like most of the bld methods.

Then there is all of the fmc approaches.

But then there is the question how Roux fits into this...



Thats my view on this. What do you think? How would you group the methods?

 

[bcube]

How would you group the methods?

Psi´s cube methods and states map might be one way.

 

[FJT97]

Psi´s cube methods and states map might be one way.

wow, that is heavy. i have a hard time wrapping my heay around that...

 

[Thom S.]

But then there is the question how Roux fits into this...

Roux is by definition a Columns first Method

 

[FJT97]

Roux is by definition a Columns first Method

oh, okay! so i found collums first and corners first.
i assume there is edges first too. what are the other groups? Layer first, eo first?

 

[sqAree]

oh, okay! so i found collums first and corners first.
i assume there is edges first too. what are the other groups? Layer first, eo first?

Orient first is probably a group, with methods like human thistlethwaite, SSC etc.

 

[Thom S.]

Belt Methods count as a group, Reduction Methods, direct Solving Methods,

A guy on the Forum made a couple of Permute first/orient last methods

 

[_zoux]

But then there is the question how Roux fits into this...

roux is Corners First

 

[FJT97]

roux is Corners First

hmm, thats weird. you dont even solve the corners first with roux...

 

[_zoux]

hmm, thats weird. you dont even solve the corners first with roux...

You do, but you also solve 6 edges along with first layer

 

[Logiqx]

oh, okay! so i found collums first and corners first.
i assume there is edges first too. what are the other groups? Layer first, eo first?

I created a diagram showing the better known 3x3x3 methods some years ago, grouped into basic categories.

I think you will find the diagram quite interesting along with the rationale for the groupings.

https://www.speedsolving.com/forum/threads/solving-methods-a-primer.41913/page-2#post-890380

 

[shadowslice e]

Personally, I don't really place methods into one and only one specific category but prefer to assign each a few labels.

The main ones I use are:
Blocks: eg Petrus, Roux etc
Orient First:eg ZZ, SSC etc
Permute first: this exists but there's not really anything in it tbh (keeping it open just in case)
Corners First: Self explanatory: eg Roux, Waterman, RICE, PCMS etc
Edges First: ZZ, EF
F2L: CFOP, ZB, ZZ
Columns first: PCMS, PEG, PORT, Roux etc
Edges last: Roux, PCMS, some types of SSC
Belt: uh Belt?, SSC
Conjugates: 42, CTLS

There are of course quite a few other you could argue but I think that this way of categorising methods is more useful than simply trying to run around in circles fitting in methods to categories

Spoiler: With that said...

With that said I would probably go with 5 based on what they do first: blocks, corners first, edges first, orient first and permute first

 

[alwin5b]

1. speedsolving methods (which gradually solve the cube, but mess up the unsolved part)
2. blind methods (which leave most of the cube untouched in each step)
3. Human Thistlethwaite (which doesn't solve anything until the very last step, instead the whole cube gets reduced into 'subgroups')

these 3 categories of methods are as different as it gets. And yes, Human Thistlethwaite gets its own category, that's how little it has in common with basically any other method.

A subcategory of speedsolving methods would be methods that pseudo-solve parts of the cube (e.g. Heise where 2x2x1 blocks get formed but not placed in their correct position)