New method extension to Fridrich

Feryll

Veteran
Joined
Dec 31, 2009
Messages
557
After some thinking and research, I have come up with a new way to (hopefully) speedsolve with. Maybe 'new' isn't the best way to phrase it, as it is a big extension off of the very popular Fridrich method. Basically, you use algorithms as you insert f2l pairs to achieve two things at once, sort of similar to MGLS. This method should be an easy read and idea to intermediate and expert cubers.

1. Solve the cross + 2 adjacent f2l pairs - Basically just that. You build the cross and first f2l pair normally, and then solve one more the isn't diagonal from the first solved pair.

2. Build the 3rd f2l pair and place it above it's slot - Basically, have the 3rd pair U R U' R' away from it's solved position. One more thing is that, with the bottom, or single color, of the f2l pair facing you, the empty slot should be on that same face. Example: F2 R' F2 R would be a properly made one. R' F R U' F' Would not (The UF edge and URF corner being the pair). If you make an improper one, you would have to learn twice as many algorithms for the next step. It should also be on the right side, or you'll have twice as many algs for the next step (but at least they'd be mirrors;))

3. Perform an algorithm to orient all edges and place the f2l pair - If I'm correct, there should be 14 algs to learn, being around 7 moves (I haven't taken a real average yet) They are relatively fingertricky, as they use mostly R, U, F moves to do it near optimally. If you don't want to learn the algorithms, this can be done intuitively, but slower (unless you're extremely good at it).

4. Make the last f2l pair, plus an extra corner - Basically, have an extra corner on the f2l pair ready to be inserted. Be sure not to make any F, F', B or B' moves, as that disturbs edge orientation. Example: L' U2 L U2 is an extended f2l pair to be inserted.

5. Insert 1x1x3 block, and solve corners - This can be done with a one-look or two look. I suggest 2 look, as the one look has 81 algorithms (8-9 moves) and the 2 look only has 9. If you know about commutators, you can orient the corners, while permuting them, but we need not worry about permutation in the first look of the 2 look. You can research commutators if you don't already know about them. Then there are 9 permuting algorithms.

6. Permute the last four edges - Remember how we oriented edges in step 3, and solved the corners in step 5? Well, that only leaves a yummy edge permutation :D! This will be an average of 10 moves, depending on how long your 2-gen algorithms are, and a big chance to get a skip on this step.

This means that this method solves the cube in 27 algorithms (Or just 13 if you do edge orientation intuitively), or 99 if you choose to go 1-look on step 5 (So I can say my full optimal method requires less than 100 algorithms!), with an average of 30-35 algorithm moves. A variation would be to go Petrus and replace steps 2 and 3 with a block version of the algorithm, not sure how it would work out.

EXAMPLE: Scramble - D2 F' D' R2 B F D2 F' L2 D L' U D' R' U' F2 U2 L' D' R2 L D2 B2 L' U
Step 1 - L' D B' U B' R2 L U' L' U L U L2 U2 L B' U2 B2 U B2 U B
Step 2 - U' L F' L' U2 F
Step 3 - U R B U' B' R'
Step 4 - B U B' U B U2 B' (This is a generated optimal solution; it will be hard to perfect this step)
Step 5a - U R' D R U' R' D' R (Commutator)
Step 5b - U R' U L' U' R U' L U
Step 6 - M' U' M'2 U' M'2 U' M' U2 M'2 U

Please give me your ideas and comments on this :)
 
Last edited:
03:23:10 <+Kirjava> I'm getting sick of random method thread
03:23:11 <+Kirjava> because
03:23:13 <+Kirjava> you can make one
03:23:16 <+Kirjava> in like 30 seconds
03:23:25 <+Kirjava> and the people who make the threads
03:23:27 <+Kirjava> don't have algs
03:23:32 <+qqwref> new methods are generally stupid as hell
03:23:35 <+Kirjava> and havn't practised with the method or anything
03:23:37 <+qqwref> some noob just wanting name recognition
03:23:40 <%j`ey> (and most of them are not original)
03:23:42 < Jude1> http://www.speedsolving.com/forum/showthread.php?p=355896#post355896
03:23:44 <+qqwref> we should be like
03:23:47 <+qqwref> you have to get sub20 with this method
03:23:51 <+Kirjava> HERE IS MY NEW METHOD
03:23:51 <+qqwref> or at least a beginner version
03:23:53 <+qqwref> in order to post it
03:23:57 <+Kirjava> SOLVE THE CROSS (OR DONT)
03:23:58 <+qqwref> that'd be good
03:24:09 <+Kirjava> THEN SOLVE SOME CORNERS (OR EDGES)
03:24:13 <+DavidWoner> I DON'T KNOW HOW MANY CASES OR MOVES OR WHATEVER BUT I THINK IT IS GOOD
03:24:21 <+Kirjava> AND FINISH THE REST IN PROBABLY 2000 ALGS
03:24:26 <+Kirjava> DOES THIS SOUND LIKE A GOOD IDEA
03:24:40 <+qqwref> HOW ABOUT THIS
03:24:41 <+qqwref> DO FRIDRICH
03:24:46 <+qqwref> BUT FOR SOME STEP ONLY DO LIKE 2/3 OF IT? IDK
03:24:50 <+qqwref> THEN SOLVE THAT LATER WHILE DOING PLL
03:24:58 <+Kirjava> XD

Give up and learn L2L4 already if you want to be useful.

Basically, no one cares unless it's actually interesting.

03:28:00 <+Kirjava> qqwref: the thing is
03:28:09 <+Kirjava> we could literally reel off liek 20 method ideas
03:28:17 <+Kirjava> that are much better than all these threads
03:28:23 <+Kirjava> in like 2 mins
03:28:31 <+qqwref> of course
03:28:36 <+qqwref> or at least
03:28:36 <+qqwref> like
03:28:41 <+qqwref> slight shitty variations on fridrich/petrus/roux

03:29:58 < Athefre> Thought I would come here after seeing Thom's post
03:30:06 <+qqwref> heh
03:30:18 <+Kirjava> the funny part is
03:30:23 <+Kirjava> I didn't even read the guy's post
03:30:34 < Athefre> Is this how you feel about my idea?
03:30:35 <%j`ey> inb4 he actually has a good idea
03:30:42 <%j`ey> (he doesn't but it'd be funny)
03:30:45 <+qqwref> the idea isn't a very good idea
03:30:55 <+qqwref> it's like... oriented edges and then
03:31:00 <%j`ey> Athefre: say what?
03:31:00 <+qqwref> make the 4th f2l pair with an extra corner
03:31:02 <+qqwref> so a 1x1x3
03:31:05 <+Kirjava> Athefre: I'm complaining about how he's thinking about it
03:31:07 <+qqwref> and then insert it while doing LL edges
03:31:07 <+Kirjava> not his idea
03:31:09 <+qqwref> which is hueg difficult
03:31:11 <+Kirjava> he doesn't have algs
03:31:15 < Athefre> Ah
03:31:18 <+Kirjava> he has no intention of learnning anything
03:31:21 <+qqwref> yeah
03:31:26 < Athefre> Ohhh
03:31:28 <+qqwref> and no intention of seeing if this is better than, say, fridrich
03:31:29 <+Kirjava> he wasn't us to tell him that he made a great idea and we're learning it
03:31:31 <+qqwref> just wants a method named after him

03:33:08 <+Kirjava> gilles didn't post his method on speedsolving.com
03:33:10 <+qqwref> yeah corners
03:33:13 <+Kirjava> he LEARNT it
03:33:17 <+qqwref> it's seriously a ridiculous difficult step
03:33:18 < Athefre> That's right
03:33:20 <+qqwref> kir: and USED it
03:33:24 <+qqwref> and got SUB FIFTEEN with it
 
Last edited:
4. Make the last f2l pair, plus an extra corner - Basically, have an extra corner on the f2l pair ready to be inserted. Be sure not to make any F, F', B or B' moves, as that disturbs edge orientation. Example: L' U2 L U2 is an extended f2l pair to be inserted

And how would you pair up the LL corner with the F2L pair without disturbing the LL edge orientation?
 
4. Make the last f2l pair, plus an extra corner - Basically, have an extra corner on the f2l pair ready to be inserted. Be sure not to make any F, F', B or B' moves, as that disturbs edge orientation. Example: L' U2 L U2 is an extended f2l pair to be inserted

And how would you pair up the LL corner with the F2L pair without disturbing the LL edge orientation?
I was putting up an example right after I posted, as it almost slipped my mind. I think this step is a big bullet hole in my method, but with practice, I believe it's possible.
03:23:10 <+Kirjava> I'm getting sick of random method thread
03:23:11 <+Kirjava> because
03:23:13 <+Kirjava> you can make one
03:23:16 <+Kirjava> in like 30 seconds
03:23:25 <+Kirjava> and the people who make the threads
03:23:27 <+Kirjava> don't have algs
03:23:32 <+qqwref> new methods are generally stupid as hell
03:23:35 <+Kirjava> and havn't practised with the method or anything
03:23:37 <+qqwref> some noob just wanting name recognition
03:23:40 <%j`ey> (and most of them are not original)
03:23:42 < Jude1> http://www.speedsolving.com/forum/showthread.php?p=355896#post355896
03:23:44 <+qqwref> we should be like
03:23:47 <+qqwref> you have to get sub20 with this method
03:23:51 <+Kirjava> HERE IS MY NEW METHOD
03:23:51 <+qqwref> or at least a beginner version
03:23:53 <+qqwref> in order to post it
03:23:57 <+Kirjava> SOLVE THE CROSS (OR DONT)
03:23:58 <+qqwref> that'd be good
03:24:09 <+Kirjava> THEN SOLVE SOME CORNERS (OR EDGES)
03:24:13 <+DavidWoner> I DON'T KNOW HOW MANY CASES OR MOVES OR WHATEVER BUT I THINK IT IS GOOD
03:24:21 <+Kirjava> AND FINISH THE REST IN PROBABLY 2000 ALGS
03:24:26 <+Kirjava> DOES THIS SOUND LIKE A GOOD IDEA
03:24:40 <+qqwref> HOW ABOUT THIS
03:24:41 <+qqwref> DO FRIDRICH
03:24:46 <+qqwref> BUT FOR SOME STEP ONLY DO LIKE 2/3 OF IT? IDK
03:24:50 <+qqwref> THEN SOLVE THAT LATER WHILE DOING PLL
03:24:58 <+Kirjava> XD

Give up and learn L2L4 already if you want to be useful.

Basically, no one cares unless it's actually interesting.

03:28:00 <+Kirjava> qqwref: the thing is
03:28:09 <+Kirjava> we could literally reel off liek 20 method ideas
03:28:17 <+Kirjava> that are much better than all these threads
03:28:23 <+Kirjava> in like 2 mins
03:28:31 <+qqwref> of course
03:28:36 <+qqwref> or at least
03:28:36 <+qqwref> like
03:28:41 <+qqwref> slight shitty variations on fridrich/petrus/roux

03:29:58 < Athefre> Thought I would come here after seeing Thom's post
03:30:06 <+qqwref> heh
03:30:18 <+Kirjava> the funny part is
03:30:23 <+Kirjava> I didn't even read the guy's post
03:30:34 < Athefre> Is this how you feel about my idea?
03:30:35 <%j`ey> inb4 he actually has a good idea
03:30:42 <%j`ey> (he doesn't but it'd be funny)
03:30:45 <+qqwref> the idea isn't a very good idea
03:30:55 <+qqwref> it's like... oriented edges and then
03:31:00 <%j`ey> Athefre: say what?
03:31:00 <+qqwref> make the 4th f2l pair with an extra corner
03:31:02 <+qqwref> so a 1x1x3
03:31:05 <+Kirjava> Athefre: I'm complaining about how he's thinking about it
03:31:07 <+qqwref> and then insert it while doing LL edges
03:31:07 <+Kirjava> not his idea
03:31:09 <+qqwref> which is hueg difficult
03:31:11 <+Kirjava> he doesn't have algs
03:31:15 < Athefre> Ah
03:31:18 <+Kirjava> he has no intention of learnning anything
03:31:21 <+qqwref> yeah
03:31:26 < Athefre> Ohhh
03:31:28 <+qqwref> and no intention of seeing if this is better than, say, fridrich
03:31:29 <+Kirjava> he wasn't us to tell him that he made a great idea and we're learning it
03:31:31 <+qqwref> just wants a method named after him

03:33:08 <+Kirjava> gilles didn't post his method on speedsolving.com
03:33:10 <+qqwref> yeah corners
03:33:13 <+Kirjava> he LEARNT it
03:33:17 <+qqwref> it's seriously a ridiculous difficult step
03:33:18 < Athefre> That's right
03:33:20 <+qqwref> kir: and USED it
03:33:24 <+qqwref> and got SUB FIFTEEN with it

I am trying to be different from the three sentence stupid new methods by actually putting effort into it. I listed the number of cases, and generally an average move count. And where did I ever even mention naming my method? Here, just for you, I'll name it 'crappy method'. And I've never heard of L2L4 when I searched it...EDIT: I found Chris Hardwick mention it, but the link is broken, and I can't find any real tutorial on it. Suggestions?
Basically, have the 3rd pair U R U' R' away from 'solvation'.

WWWOOOOOOOOOOOOOWWW
Fine, I'll take it down.
 
Last edited:
Solvation is a funny word.

But if you are going to do this why not just do ZZ-VH?
Edge Orientation + Line
3-gen F2L that can have blockbuilding
COLL (40 algs)
EPLL (4 algs) 1/12 chance to skip

Fewer algs and 2LLL

Orientation later in a solve leads to more complex situations. Usually requiring quite a few more algorithms.
 
The example you edited in was a good start. Now your goal is to prove to people that it is as fast or faster than the original. By getting fast.
 
The example is broken.

Fixed.

I always liked ZZ-VH, and I considered switching to it as my main method. I got the idea of putting edge orientation in this crappy method from it. But I think that the COLL algorithms are a bit lengthy and hard to remember, at least from the sites I've seen.

I will be generating the algorithms now, if anyone cares. Hopefully, this idea doesn't bust, as I never meant for my posts to make this thread look mediocre.

Also, I remembered how to apply commutators to corner orientation, so that reduces the number of algorithms to 27. Or, if you choose to do the edge orientation part intuitively, that reduces the algorithms to 13, hardly more than the beginners method.
 
Last edited:
It seems like the move count for this method is about the same as normal fridrich. But, fridrich should have much better/easier recognition. So in what ways is this method an improvement?
I think its an interesting idea, but I just don't know if it actually works better than fridrich.
 
But I think that the COLL algorithms are a bit lengthy and hard to remember, at least from the sites I've seen.

My COLL algorithms average less than 10 moves, and are optimized for right handed speed.
 
My COLL algorithms average less than 10 moves, and are optimized for right handed speed.

impossible. are you sure you aren't talking about CLL?

...well, kind of.

-This is not factoring the probabilities of each case.
-This is not factoring the AUF's for recognition.
-This is not factoring the AUF's after recognition.
-This is measured in HTM.
-This is not accounting for when the corners are oriented.

So, I guess I shouldn't post such an audacious remark. I just wanted to point out that COLL algorithms aren't horrible.
 
no wai

A good portion of my algs are optimal, and they're not 10 moves.
(Also HTM without AUF)

Yes way. I actually edited that page into my CMLL page, but many algs are the same (or have r switched with R in a few places). I'm going to see if I still have them.

EDIT: Some of my CMLL cases are actually longer than the COLL cases, but for the better. COLL: R U2 R2' U' R2 U' R2' U2 R (9/14) CMLL: F R U R' U' R U R' U' F' (10/10). So although my COLL and CMLL algs have almost the same HTM movecount, the QTM is extremely reduced in the CMLL (under 11 moves).
 
Last edited:
Back
Top