Orient First method

[LarsN]

I've been messing around with the idea of an orient first method (3x3) for a few weeks. I don't know if the idea is original, but anyway here is a link for a guide I wrote on the idea.

http://www.speedcubing.dk/index.php?view=category&id=40:advanced&option=com_content&Itemid=58

I think the method could be used for speedsolving. The movecount is pretty good. I did an avg of 10 with a move count of 50.4 htm. If you count M moves as 1 move because in speedsolving M moves are not that slow, the avg is more like 42.

I've been using fridriech for a year and I'm almost a sub20 avg cuber. My PB with this method after a week of practise is 26.53. I'm positive that you can do sub20 avg with the orient first method.

Give it a try if you feel like it and please give me some feedback and ideas on how to improve this method.

 

[Lofty]

Oh wow!
At first I thought you has simply reinvented the ZZ method but that one only orients edges and not corners. Sounds like a very interesting method and pretty good move count too! I may look at it if only just for fun. I like new methods.

 

[LarsN]

Oh wow!
At first I thought you has simply reinvented the ZZ method but that one only orients edges and not corners.

I should have posted a rough sketch of the idea instead of just the link. So here it is:

Step 1: orient corners
Step 2: orient edges and place middle layer edges in middlelayer
Step 3: Build f2l (not fridriech style)
Step 4: Permute last layer

It was fun to invent I think it could still use some optimization. I've got something coming up for the advanced tricks section soon.

 

[mizzle]

This sounds a lot like the PB method Ron Van Bruchem and Richard Patterson were working on last year (EDIT: I just read through again and realized they were doing something different). As I recall, they started the same way as you, and they averaged very low moves with it. Instead of F2L, they tried to place E layer edges, then separate U edges and corners to U and D edges and corners to D, to make for easy permutations.

However, they gave it up because edge orienting and permuting sometimes is terrible, and lookahead wasn't very good. Also, lots and lots of half turns. This was actually one of the inspiring ideas for my method (orient edges first, finish most of F2L, finish F2L and influence corners, finish with truncated ZB).

Let us know how you progress!

 

[dChan]

I like this idea and I will probably try it out after this week. It looks very interesting and I like the restrictions for the F2L building because I know several F2L algorithms that preserve orientation so it should be pretty easy to do, at least for me but I think even for others it should be quite simple. I would love to see someone use this method to get sub-20. A 42 move average is pretty impressive to me so maybe it could give other methods a run for their money?

 

[badmephisto]

very interesting. i'll check this is a little more detail later

edit: ok i was looking at it for about an hour, and I can't really figure it out. the description is pretty bad, particularly for step3, but i think i get it. but only because i am a genius . Also in step 2 i don't know what I am looking at in those diagrams. More examples would be nice.

 

[LarsN]

Even though the method is pretty clear in my mind I had a hard time trying to describe step 2 and 3. I'll but in some more examples at each step and try to find some better ways to explain

 

[dChan]

Yes, doing the edges was a little bit frustrating. I tried some solves using the method yesterday and while I did not have time to learn the cases for orienting the corners I was able to understand the edge part a bit better because of its similarities to solving edges in corners-first solving although, in my opinion, the 3x1x1 blocks have a lot more restrictions. I can't wait to see the advanced techniques section as I would love to fully learn the method and its tricks and practice it alongside Fridrich.

Great work so far, Lars. Keep it up as I will be keeping my eye on on that method.

 

[Lt-UnReaL]

Even though the method is pretty clear in my mind I had a hard time trying to describe step 2 and 3. I'll but in some more examples at each step and try to find some better ways to explain

Yeah I don't get it either. :S Looks like a cool method, though. First step is similar to 2x2x2 Guimond method (actually exactly like it )

 

[Radical Seal]

Im starting to practice this method a little. So far im getting 50-60second avgs.

 

[LarsN]

As requested, step 2 and 3 are now revised. Step 2 now have a more thorough step by step explanation, and the diagrams that nobody understood have been taken away (and burned). Step 3 now has a very detailed example walkthrough.

I'm planning on doing a video tutorial for step 2 and 3, but I don't know when I'll get the time for it.

I would love to hear from people if they improve with this method

 

[dChan]

The diagrams for step 2? Those were easy to understand. Although I have to admit for the step 3 part it took me a little thinking before I got it. Although I have no idea how to do case 3 still. I'll reprint my copy a little later today and see what you have done.

 

[badmephisto]

As requested, step 2 and 3 are now revised. Step 2 now have a more thorough step by step explanation, and the diagrams that nobody understood have been taken away (and burned). Step 3 now has a very detailed example walkthrough.

I'm planning on doing a video tutorial for step 2 and 3, but I don't know when I'll get the time for it.

I would love to hear from people if they improve with this method

thats great Lars, it looks much better. I will try to learn this method in the coming days. Also, a video is a great idea too.

 

[niKo]

I oriented one side of corners fine, but then came to a corner-orientation case that I can't find on your table. It looks like the third one on the second row, except the red bar on the lower-left corner is on the bottom. What's the algorithm for that?, or did I screw something up..

-niKo

EDIT: Hold on - I misread. I might be able to figure it out.

EDIT 2: I found my problem. After getting the 3 corner pattern on D layer, I rotated the cube to the pattern on the U layer instead of turning the U layer. (lol)

 

[niKo]

I don't understand where the edges have to be located for the algs in the table of step 2 to orient them.

-niKo

 

[Radical Seal]

Actually one of the cases is wrong on the sheet. Its the one on the bottom row, 4th from the left and the alg reads: L2 U L'. If you do L U' L2 on your cube, it doesn't match the picture thats on there.

 

[LarsN]

I don't understand where the edges have to be located for the algs in the table of step 2 to orient them.

-niKo

First of you must rotate the cube to have the oriented corners on U and D, then count how many edges need to be flipped, if some edges in the middle layer are flipped then you've done something wrong in the beginning of step 2:
8 - since they are all flipped you can't really change their position. just do the alg
6 - I'm still looking for a good 6-flip alg, instead do a 4 flip and then a 2 flip.
4 or 2 needs flipping - the position is stated in the table. UF is the edge that sits both in the U and F face, etc...

I'll have to do that video soon...

Actually one of the cases is wrong on the sheet. Its the one on the bottom row, 4th from the left and the alg reads: L2 U L'. If you do L U' L2 on your cube, it doesn't match the picture thats on there.

Thanks, for pointing that out. The alg is correct but the picture is wrong. It has been corrected

 

[niKo]

Oh, I understand, the notation in the table is pointing out where the edges are. (I should've figured that out).

-niKo

 

[dChan]

I'm just wondering: is anyone thinking of fully adopting this method or at least completely learning it and having it as a secondary speedcubing method? I definitely want to try to adopt it as my primary method but I am more worried about getting my time below 20 so I'll probably try learning it but I won't fully adopt it until later on.

 

[pete]

I would love to hear from people if they improve with this method

i actually tried it first time you post it, i felt it was a promising method but like
others i struggled to understand the explanation & some diagrams.
then i found some errors in some algorithms (for example the one that orients
4 edges did not work).

i have then decided to run your example in a hope that it would give me some
clues but there was an error in the example too so i have eventually gave up completely.

are you saying you have now corrected those issues ? let me have another look...