Orient First method

[badmephisto]

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.

Lars claims he can get 26 second averages with this method for just weeks practice. That is a pretty large incentive to try and get into this method After all, it is a New method too... Fridrich has been around forever and we now start to see its limits. This method without doubt has many unexplored ways of getting faster as well.
Lars if you get around to making that video for this method that would be nice too. I do think it has potential, i just wish more people would look at it a little more and say what they think.

 

[Lt-UnReaL]

i just wish more people would look at it a little more and say what they think.

I read through it and tried, didn't really understand it, so I'm waiting for the video.

 

[niKo]

I am learning it - I just have the corner orientation algs left to memorize. There is one awkward F2L case though, for building a 1x1x3 block. If two pieces are connected correctly in the upper layer, but the remaining corner piece is in the bottom layer twisted in such a way that a R2 or L2 turn does not align them. I found a way to get around it, by positioning the 1x1x2 in such a way that it's not disturbed when I re-position the corner. But is there a way to do this other than bringing the misaligned corner to the U layer, back to the D layer, then back up again? o.O

After I get some practice (a lot of practice) with the method I'll put some times in my signature.

-niKo

EDIT: Is it just me or is your website down, Lars?

 

[dChan]

Heh, I was so excited about learning this method, after only like 50 practice solves I started to try to learn it. I spent the rest of the day trying to remember the corner orientation algorithms and I completely neglected BLD practice. I learned 8 cases so far so now I have 8 more to go. The algorithms are actually very easy to understand so that helps a lot with memorization. I hope that if I can get in enough practice time I might be able to use it during San Diego 2008. Of course the only problem is that M-layer turns aren't good for one-handed solving so I'll probably still keep Fridrich around. However, I'm still not fully dedicated to the method yet as I want to give it a test run of a few hundred speedsolves before I think of switching over.

The Nielsson method shall one day rule the world! Yes, I am already calling it the "Nielsson" method. Got a problem with that you Fridrich lovers?

EDIT: I don't have any idea about Niko's problem as I have not spent much time learning the second step yet. I've only done some sample run throughs of it. But I just wanted to say that sometimes during corner orientation you can see that an algorithm you would normally use for orienting LL corners is partially completed for you so it is much faster to simply finish it off. This also works from when you see that you can pair up two sets of oriented corners leaving you with an unoriented U-layer. It is much faster to do a T-OLL on the U-layer than to do two set-up moves + the orientation algorithm for Guimond. I don't know if I am making any sense but I'll probably write a little guide on this at the end of my own learning period.

 

[LarsN]

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...

I didn't find any problems with the examples and I've run them through twice. Are you sure that you are doing the right M moves. In my descriptions M follows L.
The alg that flips four edge (M' U M') may look wrong if you're a blindcuber because it moves the edges around, but that doesn't matter with this method as long as they are flipped. Try doing it on a solved cube. It will look messed up, but when you look closer four edges will be flipped but corners are still correctly oriented and the middlelayer edges are also safe.

 

[dChan]

Lars, is there not an easier way of inserting the final edge into the middle-layer when you have three inserted edges already? The three-edge cycle seems a bit long for such a simple task(well, relatively).

 

[LarsN]

Lars, is there not an easier way of inserting the final edge into the middle-layer when you have three inserted edges already? The three-edge cycle seems a bit long for such a simple task(well, relatively).

I try to avoid getting into that situation. By looking ahead during the first step, or even better during inspection, you should be able to see how many middle edges are already in place. If none insert two edges and then two again. If one edge is in place insert one edge next to it, then the two remaining. This is the fastest way that I have discovered so far

Then of course there is the cases with misoriented but in place middle edges. This I plan to put in the advanced tricks section soon.

 

[dChan]

I can't wait for the advanced tricks section. Do you use the method presented on the page or do you use the advanced tricks stuff?

 

[LarsN]

I can't wait for the advanced tricks section. Do you use the method presented on the page or do you use the advanced tricks stuff?

I'm glad that you're so interested in this method

I mainly use what's in the guide now, but with some shortcuts that I don't describe because it is confusing enough as it is. The planed stuff for the advanced tricks is how to deal with misoriented middle edge pieces (I do this now), a little trick for saving movecount between step 1 and 2 (I don't do this yet), and some tricks for step three for special cases (I do the cases that I've found so far). Also I am still optimizing the method, so when I find something new it will end up in the advanced section.

 

[dChan]

Wow, I really want to check that out. Is your average dropping below 26 seconds yet? Also, could you do a simple speedsolve video of 5 solves? Not a one-step thing or anything like that, just a normal speedsolve set using the method. I just want to see the method in action because I'm doing very slow solves at the moment(I need to get used to the corner orientation algorithms and step 2 - only my step 3 and 4 are fast)

I am probably going to end up using it as my main method soon if I learn it quick enough. I just need to finish up learning the corner orientation and I am sure the rest will be easy as the restricted F2L step is something I am used to doing when preserving the LL during normal speedsolving.

Do you have any tips for doing a solve with one-hand? I have yet to try it but it seems it might be problematic due to the number of slice moves used during the second step.

 

[LarsN]

Sorry, I don't do onehanded at all. I can hardly get sub1 min with friedrich method. And I'm afraid I wasn't clear enough in my first post, 26s was a PB single at the time. Now I'm getting close to 30s avg. This is because I don't have much time to practise. I'll try to do some speedsolves tonight, for video and for practise

I've finally finish a video tutorial for step 2. It's the same stuff as in the guide, but it may be easier to understand for some. This is the first time I've done any kind of video and somehow the sound got very low, so you'll have to turn up the volume

http://www.youtube.com/watch?v=iXu137yrX9k

I hope it helps...

 

[dChan]

Sorry, the video is really too low. My speakers don't work so I use headphones, maybe that is the problem?

 

[Radical Seal]

Lars, is there not an easier way of inserting the final edge into the middle-layer when you have three inserted edges already? The three-edge cycle seems a bit long for such a simple task(well, relatively).

What I do is place the last edge in the UR position and its target in the UF position and do: F' L' F B' U B
Also, if the edge is in place but oriented wrong, I place it in the UF position and do: U' R' U2 R F R U R2 U' F'
Both of those algs preserve corner and edge orientation.

 

[dChan]

@RadicalSeal: Thanks! I like those algorithms. I always knew how to do those but whenever I came to the case I always forget. Now that I have the actual algorithm it should be easier. Thanks, again!

 

[watermelon]

I've thought a lot about these orient -> permute methods before, and the permutation part of this method is very similar to the one I developed . However, I will describe my alternative (which is optimized for HTM) to the orientation steps here.

Step 1: Orient the edges - very easy to do with a little practice, and you can average around 4-6 moves

Step 2: Put the E edges in E using RULD moves - again, very little skill needed, and a low move count as well

Step 3: Orient the corners - this step requires a few more algorithms and moves than the first step of your method, but they are still fingertrick friendly and fairly short - uses 1 algorithm if 3 or more corners are oriented, 2 otherwise

This is obviously more moves than your orientation method when counting in STM, but I think it is comparable in HTM (and since it doesn't use slice turns, it may be advantageous for OH solving).

I'd be interested to hear your comments Lars, since you seem to be the master of this type of method .

 

[LarsN]

Sorry, the video is really too low. My speakers don't work so I use headphones, maybe that is the problem?

I have removed the low sound video and added an amplified video:
http://www.youtube.com/watch?v=Y3kcTY12T8U

I'd be interested to hear your comments Lars, since you seem to be the master of this type of method .

Master is a strong word, but I'll accept it

I chose to orient corners first, because it can be done alg based. Most people will say that intuition is better, but with those very short algs you can learn to predict the situation for step 2 (orienting edges) during inspection and maybe even improve the situation by adding some moves to the CO algs. Further more I just like M-moves

I'm curious though...the algs you use in your method for orienting corners, are they similar to CO algs in blindfolded methods or shorter?

As a side note I would like to add that I finally found a 5 move 6-flip and 2-flip for step 2 U and D-layer edge orient it's in the guide.

 

[watermelon]

I'm curious though...the algs you use in your method for orienting corners, are they similar to CO algs in blindfolded methods or shorter?

I definitely prefer your orientation method over mine, but I just thought it would be interesting to explore some other alternatives .

The only algorithms for CO that I've used for this method so far have been generated mid-solve via ACube (keeping all edges oriented, and not disturbing E edges). However, if we were going for efficiency, it would probably be best to create a special KSolve file that orients U/D corners while keeping E edges on E (but not necessarily in the same place). If anyone is interested, I'd be glad make one this weekend and generate the necessary algorithms.

 

[niKo]

I'm interested : ) - sounds like it would definitely help with the lookahead.

-niKo

EDIT: @dChan- I found a solution. By doing a double turn to put the 1x1x2 block in the bottom layer I can position the corner so that I can bring the 1x1x2 block to the U layer & connect with it. So far that seems about the most thinking involved in the method, optimizing the F2L.

 

[badmephisto]

good vid lars. Ironically you skipped quickly over the part that i couldnt figure out for a while, that is when you insert edges into middle, but both of the the 2 remaining edges are on left or they are both on right. But I managed to figure it out

 

[Henrik]

Lars has now added a 6 edge flip alg and a shorter 2 flip edge alg. (which I found Im just taking some credit here. Lars the optimized them last night).

Henrik