"How to Solve a Rubik's cube for MIT graduates"

Hello all,

I'm thinking of putting together a "How to Solve a Rubik's cube for MIT graduates" video. Basically, it will be a fast paced video that will go over a completely algorithm-free method of solving the cube. Since people sometimes complain to me about how lame it is that you have to learn algorithms, and that there is no intelligence to solving it. Well, I can show them how to solve it completely intuitively, and make them feel dumb in the process, even as I try to make perfect sense.

I'm thinking about what this method should be. For now, I'm leaning toward this way of solving it, which I sometimes play around with. It's a Fridrich+Commutators hybrid as follows:

1. cross
2. f2l + orient edges (petrus style)
3. commutators to twist corners, and permute corners and edges.

I'm not super happy it. Can anyone think of ways to complete the last layer completely algorithm free? Or maybe position the edges as well as orienting then while finishing F2L? Or a whole different method altogether that could be suitable for this?

Suggestions welcome! Thanks!!!
-meph

You could try to look into the sexy move solution. It has a rather intuitive way of orienting and permuting the last 5 edges. You could try a BLD method hybrid maybe.

Freestyle F2L, comms for CLL and ELL.

How about block building petrus style the first 3x2x2. That's very intuitive once you get pushed in the right direction. Then orient edges and finish f2l minus on pair. Permute last edges. Permute corners using a keyhole like commutator with the RDF corner. Like: (L' D2 L, U2) I don't remember correct com notation, but you get the idea.

I love your idea, badmephisto, but honestly everytime I encounter somebody who says they want to learn it "intuitively" and not learn algorithms, I just respond with "If you want to figure it out on your own, why are you asking for someone to teach you how to do it?"

The Sexy Move solution though is probably one of the best. You only learn one "algorithm" and it's just a commutator anyway. Unfortunately you don't build the standard first two layers and do the last layer separately. If I remember right, do all the edges, then corners, then orient the corners.

May I suggest an edges first approach:
Cross
3 middle layer edges via keyhole
EO and Edge permutation a la Heise
Commutators to solve corners

In this method the playing with edges and EO is very intuitive.
You can try a lot of things when working on your last layer without too much risk of messing up the already solved F2L edges.

Commutators will be a little more work but hey, we're talking MIT MIR?

An easy introduction to commutators could be to place one corner at a time:
(Set up)
A place corner
B replace by random unsolved corner (reccomend half turn)
A'
B'
(Set up)'

Reason for only solving 1 at a time is to reduce the number of set up moves: less chance of messing up
Same reason for the D2 move (got this D2 from a commutator based beginner method described by Chris Hardwick)

After this you can show that by choosing the 2nd corner carefully you can actually solve 2 (when lucky 3...) at the same time.
For the last three corners you MUST choose carefully to ensure you solve all three at once (instead of ending up with twisted corners)
Since even the MIT guys will eventually run into twisted corners you could show how to solve twisted corners by two commutators.

1st comm
A Place 1st twisted corner in location of 2nd twisted corner
B replace by random solved corner (reccomend half turn)
A'
B'

2nd comm:
Solve the resulting 3 cycle

If you really want to challenge them you could offcourse use corner twisting commutators.

Corners first is nice and intuitive (although not fast). A basic intuitive solution would look like this:
- first layer of corners (intuitive inserts)
- second layer, CP and CO separately with commutators (for instance: R'DR U ... and R'D2RFD2F' U ...)
- insert 3 edges in left and 3 corresponding edges in right (intuitive inserts)
- roux style L6E (no algorithms, completely intuitive, pretty efficient)

Firstly, YAY for new videos. Secondly, how about a Heise finish. Do the F2L minus one CE pair (could do this with keyhole to make it easier), then use that open slot to orient and permute the LL edges, then use Commutators for the last 5 corners.

qqwref said:

Corners first is nice and intuitive (although not fast). A basic intuitive solution would look like this:
- first layer of corners (intuitive inserts)
- second layer, CP and CO separately with commutators (for instance: R'DR U ... and R'D2RFD2F' U ...)
- insert 3 edges in left and 3 corresponding edges in right (intuitive inserts)
- roux style L6E (no algorithms, completely intuitive, pretty efficient)

Click to expand...

This looks the best to me so far. The only really tricky part is the CLL. It seems really hard to make a completely intuitive corners method; the only I've seen is the sexy move trick for LL.

You only need to come up with one three corner cycle to solve the corners.

Kynit said:

This looks the best to me so far. The only really tricky part is the CLL. It seems really hard to make a completely intuitive corners method; the only I've seen is the sexy move trick for LL.

Click to expand...

[R' D' R D]2 U/U'/U2 [D' R' D' R]2 U'/U/U2

R' D' R U/U'/U2 R' D R etc.

badmephisto said:

Since people sometimes complain to me about how lame it is that you have to learn algorithms, and that there is no intelligence to solving it.

Click to expand...

1st of all kudos for thinking this!!
2nd, i guess some people who do not know cubing at all (read smart a$$es) will never appreciate the way it is to be solved. Many non cubers think it is to be solved "face by face" using some alien intelligence. So algs do turn them off The suggestions posted here by many people do involve some amount of algs (even if basic ones). So it wont really "answer" the critics.
I do agree on your point about "and make them feel dumb in the process" :tu
But do we really need to do this just to answer some smart $$$.??
I guess (personal guess) the evolution of cubing is such that it might have started out by intuitive solving and later people wrote algs for what they 'intuited'. New age cubers learn from people who struggled and made it easy for the next generation (read badmephisto as one of them ).
anyway i think im being neuteral (or skeptical) idk
Cheers!

Why not just teach them Heise?

You can teach them without explaining the Algorithm, An example

(For this case we will use R' D' blah blah blah)

Instead show them the theory in which the pieces need to go. For example tell them that the red white and blue corner should match up with the blue and white edge piece by moving it a certain way. But show them how it should move but without stating the Algorithm. Show them how they should connect and meet up with each other

(It's kind of like teaching the petrus method, But showing them fridrich.)

I have had much success with showing my friends how the colors connect with each other and how they should move. I "Show" them the Algorithm, i don't say it to them. It makes more sense to a beginner to see how they connect and their brain can process it better. If they want to speed cube i show them the Algorithm theory and the language of a cuber

Good luck

Thanks these are great answers and great ideas, I'll check them out.

I should comment that I was only half joking in topic description. I just basically want a video on a method to solve the cube that requires no algorithms whatsoever. The argument then is that if you learn to solve it this way, you will never forget, because it's a deeper understanding of how the cube works. Also, some people may prefer it, or see it as a challenge to learn it that way. The video would be more of a hint video, would not explain everything in too much detail (there's probably no time), but I want to at least lay out a method, show what needs to be done when, how things should look like at all stages, etc, and maybe give an example or two that show it.


Thank you!

I used a very intuitive corner first method, what i try perfecting now , and i used only for FMC.
first time i try a solution for all corners , usually with Ortega method, sometimes with Guimond or other algs,
then i changed this corner-algo, so that a few edges to be solved on two opposite sides, but not much more moves is needed,
then i solved more edges, to be Roux 6E4C position.
then i solved the last 6 edges.
For example: if the corner algo was FDU2R’BF’D2, but this not solve edges, some of the possible variant: Fdu2R’bF’D2 or fDU2R’bf'd2 and more endless variant…I make changes to it,which must be solved edges. This method the corners position still, but a few opposite edges effectívely solvable at one time with corners solve.
Two nice solution with this method: http://fmc.mustcube.net/ , round 328, 329, classic.

Roux has very little algorithms, the only non-intuitive ones being CMLL. You could use commutators for that stage.

Heise

Corners first.

Remember to constantly use intimidating words like, "Permutation" and "Algorithmic".